graham: new cluster is open for testing use
SIGN-IN

Research Profiles

Research Areas

Profiles

Acoustics

Complexity in cultural systems

Suárez Juan Luis, University of Western Ontario

Dr. Suarez approaches problems around complexity in cultural systems from several different perspectives: Humanism, Golden Age Drama, the Literary Act, the Baroque. In most cases he uses the Early Modern period of transatlantic relations in the Spanish world as the lab to trace the emergence of complexity in our world. He also works on present forms of cultural adaptation to complexity through the literary act. Methodologically, his work follows the principles of team-work, intensive research (both in the elaboration and creation of knowledge), use of information technology, and a clear concern to make the research a tool to better our lives today. If the lab is, mainly, in the past, the aim is clearly on the present. A substantial part of Dr. Suarez's current effort is focused on the project, "The Hispanic Baroque: Complexity in the first Atlantic culture" (2006-2013). This project involves the efforts of a group of 35 researchers from universities in different countries (Canada, Spain, Mexico, Australia, England, Bolivia and the United States) from different disciplines (Literary Studies, History, Sociology, Fine Art, Music & Musicology, Anthropology, Geography, Computer Science, Architecture & Mathematics). The team studies the origin, evolution, transmission and effectiveness of the baroque patterns of behaviour and representation in the Hispanic world. The project is financed by the Social Sciences and Humanities Research Council of Canada, by way of a Major Collaborative Research Initiatives grant. More info

Acoustics

Design and development of text analysis tools

Stéfan Sinclair, McMaster University

Dr. Sinclair's current research is focused on the design and development of text analysis tools for the digital humanities, and especially literary analysis where conventional text mining techniques are not appropriate. The emphasis is on interfaces that are accessible to scholars with less experience in text analysis but powerful and flexible enough for scholars with more experience. Sinclair has been involved in several dynamic visualization interfaces that are innovative in their crossover between large datasets on the one hand, and conceptual models of domain experts on the other. Several of the projects in which Sinclair has been involved have had a significant impact on practices in both the digital humanities (e.g. features of HyperPo) and collaborative literary criticism (e.g. the Satorbase database). More info

Applied Econometrics

Exploring how health and wealth factor impact retirement decisions

Todd Stinebrickner, University of Western Ontario

Professor Todd Stinebrickner of The University of Western Ontario is studying the link between health and wealth, and how this relationship may impact retirement decisions. Stinebrickner uses SHARCNET to run complex financial models that study and simulate the relationship between health and financial resources, and how these factors affect retirement decisions. This kind of research could potentially influence social and economic policies related to Canadian social security. More info

Applied Mathematics

Numerical "computer go"

Thomas Wolf, Brock University

Together with SHARCNET Research Chair, Alexander Odesskii, as well as students and people helping to maintain and develop the underlying system, Wolf is doing Computer Algebra which includes the development of algorithms, software and running the computations. The computations are concerned with the integra ability of differential equations or systems of equations. This leads either to overdetermined systems of partial differential equations or also polynomial (i.e. non-linear) systems of algebraic equations. Wolf also develops algorithms and computer programs for “computer go” which is an exciting area of Artificial Intelligence which is growing in popularity. More info

Artificial intelligence

Combating outbreaks of foot and mouth disease

Deborah Stacey, University of Guelph

Deborah Stacey, Professor of Computing and Information Science at the University of Guelph, is using SHARCNET to develop high performance software to model the spread of diseases such as Foot and Mouth Disease (FMD). Stacey is being funded by the Canadian Food Inspection Agency (CFIA), the Ontario Ministry of Agriculture and Food (OMAF) and the CBRN Research and Technology Initiative (CRTI). The development of the software to model disease spread on the SHARCNET computing environment will greatly augment strategies in disease control and disease spread prevention. More info

Artificial intelligence

Early detection of disease outbreaks

David Calvert, University of Guelph

The research of David Calvert, a Professor at the Department of Computing and Information Science at the University of Guelph, focuses on Artificial Neural Networks (ANNs), mainly for the analysis and modeling of biological data. An ANN can be used to detect changes in a particular data set and signal that something unusual is occurring. Professor Calvert uses ANNs on SHARCNET for applications like the early detection of lameness in horses and the early detection of disease outbreaks in a population. The early detection of disease outbreaks is achieved, using an ANN, by monitoring the sales of certain over-the-counter drugs. Changes in sales patterns can be used to detect the early signs of an outbreak and used to alert public health officials. The application of HPC to ANNs themselves is also very important. Once a computational profile of a given set of components is determined, it may be possible to predict the behaviour of a wide range of real Neural Networks. SHARCNET allows for a large number of ANNs to be trained simultaneously, to optimize performance and lead to more accurate prediction models. Applications of this research are wide ranging, including positive impacts to public health and agricultural industries. Dr. Calvert's work on syndrome surveillance is being done in conjunction with the Public Health Agency of Canada (PHA). More info

Astrophysics

Simulating the collapse and formation of planets

James Wadsley, McMaster University

Professor of Physics and Astronomy at McMaster University and SHARCNET Chair in HPC Tools, James Wadsley simulates complex gas systems that are too large to be simulated with regular computers. His research sheds light on the nature of the universe and, more specifically, the origins of planets, stars, and galaxies. More info

Astrophysics

Stellar Feedback in Dwarf Galaxy Formation

Sergey Mashchenko, McMaster University

As was published in Science magazine, by using supercomputer simulations researchers Hugh Couchman, James Wadsley and Sergey Mashchenko at McMaster University have exposed a very violent and critical relationship between interstellar gas and dark matter when galaxies are born – one that has been largely ignored by the current model of how the universe evolved. The findings solve a longstanding problem of the widely accepted model – Cold Dark Matter cosmology – which suggests there is much more dark matter in the central regions of galaxies than actual scientific observations suggest. Supercomputer cosmological simulations prove that indeed, this problem can be resolved. The researchers modeled the formation of a dwarf galaxy to illustrate the very violent processes galaxies suffer at their births, a process in which dense gas clouds in the galaxy form massive stars, which, at the ends of their lives, blow up as supernovae. “These huge explosions push the interstellar gas clouds back and forth in the centre of the galaxy,” says Mashchenko, the lead author of the study. “Our high-resolution model did extremely accurate simulations, showing that this ‘sloshing’ effect – similar to water in a bathtub— kicks most of the dark matter out of the centre of the galaxy.” Cosmologists have largely discounted the role interstellar gas has played in the formation of galaxies and this new research, says Mashchenko, will force scientists to think in new terms and could lead to a better understanding of dark matter. More info

Atmopheric Science

Land-Atmosphere Interface

John Lin, University of Waterloo

Dr. Lin's research interests lie at the land-atmosphere interface, where greenhouse gases, pollutants, water, energy, and momentum are exchanged between the land and the atmosphere. Extensive human activities are altering the concentrations of different gases in the atmosphere and the fluxes of energy and momentum, with potentially profound climate and environmental consequences. The work seeks to understand the impact of such human-induced changes under the following three broad themes: greenhouse gases; air quality; and atmospheric modeling. This group's research approach is to combine data analyses with models of the atmosphere and the land. These models are strongly anchored to available observations. More info

Bioinformatics

Fragment-HMM: A new approach to three dimensional protein modeling

Ming Li, University of Waterloo

Proteins play a number of important roles in the human system - from carrying oxygen through the body to fighting off harmful infections, they are essential to all living biological organisms. Following the protein forming process, or "protein synthesis", the resultant protein ribbons fold into three dimensional structures that vary depending on function. However, when proteins "misfold" a number of serious and potentially fatal diseases can occur, such as Parkinson's disease, Alzheimer's, and even cancer. As the structure of a protein is directly linked to its function, modeling protein structures has become one of the major focuses of pharmaceutical research in terms of medicine design and delivery. Although traditional protein modeling methods are slow and expensive, Dr. Ming Li, from the University of Waterloo, has made considerable strides in developing an effective, unified theory to model the protein folding process facilitated by SHARCNET's powerful processing resources. More info

Bioinformatics

Isolated mutations play significant role in human genome studies

Peter Rogan, University of Western Ontario

Human genetic diseases are most commonly caused by mutations that impact specific parts of the genome. These point mutations can alter the way DNA is translated and expressed, leading to the onset of diseases. Prof. Peter Rogan, Canada Research Chair in Genome Bioinformatics at The University of Western Ontario, says specific, individual mutations are linked to the onset of genetic conditions such as lupus, asthma and diabetes. Being able to identify these genetic disorders can play a huge role in selecting plants and animals with the most desirable traits for breeding or to select those with natural resistance to pests and infections. In order to disseminate single copy technology for genomic analysis worldwide, Dr. Rogan has founded Cytognomix Inc. to market products to diagnostic and research laboratories. More info

Biophysics

Molecular mechanisms of light-catalyzed photochemical reactions in photosynthetic organisms

Sergey Vassiliev, Brock University

Dr. Serguei Vassiliev at Brock University is studying physical mechanisms determining function and dynamics of proteins, macromolecular assemblies of pigment-protein complexes. Most of his work is related to studying molecular mechanisms of light-catalyzed photochemical reactions in photosynthetic organisms.In photosynthesis, water is oxidized to molecular oxygen, and ozone that protects us from UV radiation. The electrons and protons released in this process are then used to synthesize carbohydrates, the source of food.These reactions take place in a variety of the extremely complex enzymatic systems sharing a similar structure in which a large number of chromophores (chlorophylls, carotenoids and phycobilins) act as light-harvesting antenna to absorb and direct solar energy to a single reaction center where the charge separation that drives the chemistry of photosynthesis occurs. The process is carried out with nearly 100% efficiency and no toxic products. Photosynthetic systems have been fine tuned by 3 billions years of evolution to achieve such extremely high efficiency. His long-term goal is to understand physical mechanisms of these processes at the atomic level. He is studying how nature has solved this multidimensional optimization problem of light collection and utilization in photosynthesis applying biomolecular simulations and dynamic modeling of photophysical and photochemical processes at atomic level. More info

Biophysics

Towards more efficient and safer pharmaceuticals

Boris Zhorov, McMaster University

The research of Boris Zhorov, a Professor of Biochemistry and Biomedical Sciences at McMaster University, focuses on molecular modeling, the computer-based simulations of physical and chemical properties of molecules. Zhorov focuses on ion channels, biological molecules that allow permeation of specific inorganic ions (calcium, potassium, etc.) in human body cells. He studies ion selection, permeation and "gating" of ion channels and how drugs and toxins change these fundamental properties. The growing power of high performance computing installations like SHARCNET offer an unprecedented opportunity to model various ion channels and predict effects of drugs and toxins, ultimately leading to the development of new, more efficient and safer pharmaceuticals. Dr. Zhorov is currently collaborating with several experimental labs to verify results derived from SHARCNET, including Dr. Ging-Kuo Wang (Harvard University), Dr. Daniel Yang (McMaster University), Dr. Stefan Hering (University of Vienna), and Dr. Satomi Adachi-Akahane (University of Tokyo). More info

Biostatistics

Methodological research and applications in biostatistics

Yingwei Peng, Queen's University

Dr. Peng is working on methodological research and applications in biostatistics areas. In particular, he is working on some novel survival models that can be used to analyze survival data arising from clinical trials studies and observational studies. The models are capable of handling subjects that are immune to the failure events, which is often found in cancer studies, and potential cluster/grouping effects, which is often found in observational studies. Due to the complexity of the models developed, it is often difficult, if not impossible, to investigate the properties of an estimation method in theory, and the team must rely heavily on computer simulation to examine the properties. The simulation work can easily surpass the capacity of a personal computer or a small group of networked local computers, and it often has to run weeks to obtain the results. Sharcnet clusters fit the need of this simulation work perfectly. Although the programs that run for simulation cannot be parallelized easily, the nature of simulation makes it relatively easy to distribute the simulation work to a large number of machines in clusters simultaneously, which dramatically reduces the computing time required to complete the simulation. More info

Business To Business Process Modeling, Simulation and Control

Building New Models of Capital Market Behaviour

Louis Culumovic, Brock University

Dr. Culumovic is working on a long-term project that is building new models of capital market behaviour. He has developed some computational models that require intensive calculations. They involve significant iterative computations involving very large numbers and very large numbers of significant digits in order to get accurate calculations. It would be extremely difficult, if not impossible, to do these calculations without a high-performance computer such as SHARCNET.

Cargo Revenue Management

Revenue Management (RM) Systems

Mikhail Nediak, Queen's University

In the thirty years since the successes of revenue management (RM) systems in airlines were first widely reported, applications have spread steadily into other business areas. RM is now common in such service businesses as passenger railways, cruise lines, hotel and motel accommodation, and car rentals. Other applications have been proposed in such diverse areas as broadcast advertising, sports and entertainment event management, medical services, real estate, and manufacturing. RM and related dynamic pricing practices are also gaining popularity in the retail industry and have engendered a growing body of academic research in recent years. However, the general area of capacity and RM for the supply chains that retailers depend on has been developing much more slowly. Dr. Nediak's research group believes that the slow pace of development in the area is due, in part, to the computational complexity of cargo transportation capacity and RM problems. Although cargo capacity is a ‘perishable asset’ like passenger capacity, it is often random and is at least two-dimensional (weight and volume). This area is open for the introduction of new methodologies that can improve the operations of cargo-carriers in today's challenging economic environment. The general objective is to develop models and computational methods to increase the efficiency and profitability of companies that are controlling and selling perishable inventories over time. In the case of cargo, key decisions faced by carriers relate to the acceptance and scheduling of cargo shipments from diverse customers, negotiation of prices, and the coordination of shipments over route networks. In these settings, the perishable inventory is shipping capacity at specific times over particular routes, and the challenge is efficient (cost and fuel) utilization of the capacity as well as maximization of revenues. Dr. Nediak plans to develop a new class of models and methods to help companies address the challenges posed by cargo capacity. Successful implementation of such methods can improve utilization of cargo space, reduce fuel consumption and other costs, and enhance the profitability and long term viability of cargo carrier businesses.

Civil Engineering

Finite Element Analysis and Structural Dynamics

Ashraf El Damatty, University of Western Ontario

During the past ten years, Dr. El Damatty has developed his own computer software that can be used in performing stress analysis of various engineering problems. His research interests include the stress analysis of structures having a shell of revolution surfaces, e.g. tanks, domes, chimneys and pipes. The objective of his research is to understand the complex state of stresses developing in such structures under various environmental loads and to develop simple design procedure for these structures. His research in shell structures, especially in steel conical tanks, has gained a wide international recognition.

Computational Biology

Early identification of cancerous cells using computer science

Mark Daley, University of Western Ontario

Mark Daley, an Assistant Professor of Computer Science and SHARCNET Chair of Biocomputing at The University of Western Ontario, focuses on applying traditional computer science theory to the modeling of biological systems, particularly in the area of molecular genetics. Some of his most recent results include a computational model of genetic compression in viruses that has lead to a novel method for predicting certain classes of proto-oncogenes, which are linked to cancer. Currently, Professor Daley is working on a virus model, which will be deployed on SHARCNET, to scan the human genome for potential proto-oncogenes. A proto-oncogene is a gene which, when damaged, may lead to the development of cancer. Another area of investigation is a model of the "gene descrambling" process found in stichotrichous ciliates, single-celled organisms that maintain their DNA in a scrambled form and must de-scramble it in order to survive. Comparatively, (unscrambled) DNA is automatically converted to genes in the human body. Professor Daley intends to use SHARCNET to try and decipher the "descrambling" process of ciliates with the ultimate aim of being able to perform computations with bio-molecules, effectively creating programmable cells. Without SHARCNET, these models, which are highly computationally-intensive, would be infeasible. More info

Computational Biology

Understanding the causes of skeletal birth defects

J Stone, McMaster University

Jonathon Stone, a Professor in McMaster University's Biology Department and a SHARCNET Chair in Biocomputing, analyzes the development and evolution of skeletons in both vertebrates and invertebrates. Using computer simulations and models, he conducts investigations that would be impossible to perform experimentally on humans. His research has wide-reaching applications in the health sciences, as it will help uncover the causes of syndromes, birth defects, and cancers. More info

Computational Biology

Understanding the evolution of genes

G. Brian Golding, McMaster University

In the area of bioinformatics, McMaster University Professor Brian Golding is using SHARCNET to study molecular evolution and DNA sequence analysis. His research attempts to understand how the processes of evolution act to cause the changes actually observed between molecules, between genes and between genomes. The recent advances in molecular genetics are providing a storm of new data on DNA sequences, on gene structure and higher order genomic structure. However, the implications of these new data are not always clear. Professor Golding's research group makes use of computer based analysis, statistical analysis and mathematical models to answer broad questions about the biology of all organisms. More info

Computational Biology

Unraveling the mysteries that lie in the smallest strands of DNA

Lukens Lewis, University of Guelph

Researchers at the University of Guelph have been using SHARCNET to discover the functions of a new and little understood genetic component known as small RNAs. They suspect these tiny molecules, which assist in regulating protein production in both plants and animals, could be a key factor for healthy cell growth and development. Prof. Lewis Lukens, Department of Plant Agriculture, and graduate student Shuhua Zhan have discovered eight new types of small RNAs, now known as micro RNAs, in plants. He found their functions range from processing sugars - an essential element for plant survival - to assisting in plant development. Learning more about small RNAs in plants might one day help develop healthier or more efficient crops. More info

Computational Biology, Systems Biology

Gene mapping to help the environment

Gabriel Moreno-Hagelsieb, Wilfrid Laurier University

Gabriel Moreno-Hagelsieb, an Assistant Professor at Wilfrid Laurier University, is widely published in the area of genome research. Genomes are a complete map of all genetic information or hereditary material possessed by an organism. Using information from all publicly available microbial genomes (tiny or micro organisms, often classed as bacteria), Professor Moreno-Hagelsieb aims to determine how genes are organized into functional modules, or how genes work together to attain a particular biological function. Because of the many microbial genome projects, there is a great deal of information about the existence of many genes but there is little evidence as to what their functions or products are. In addition, because genes within species continue to evolve, it is crucial to be able to determine if changes have a functional consequence. Using SHARCNET, Professor Moreno-Hagelsieb is seeking to classify all genes from most stable to least stable. His research has applications to the development of antimicrobial agents and for finding the most efficient genes for pathway engineering, which has applications to environmental clean up. More info

Computational Chemistry

Computational Chemistry and Atomic Clusters

Rene Fournier, York University

Dr. Fournier and his co-investigators, Min Zhang and Yan Sun, are using computational chemistry to study small gold clusters, Au-Na mixed clusters and also clusters with closed geometric and electronic shells. Specifically, they are interested in the geometric structure of clusters, which ones are relatively stable, and the development of new computational methods for modeling chemical structure and reactivity of metal element clusters. The research conducted by Dr. Fournier is extremely exciting because it is on the computational chemistry forefront. The research results can then be used by experimentalists to guide them in their attempts to synthesize stable clusters and assemble them to create new materials with unique properties. In addition, the research could potentially provide the basic theoretical model for understanding the factors that govern cluster structure and stability. More info

Computational Chemistry

Developing more fuel efficient cars; environmentally-friendly motor oils; more durable plastics

Tom Woo, University of Ottawa

The research of Tom Woo, Associate Professor of Chemistry at the University of Ottawa, and Canada Research Chair in Catalyst Modeling and Computational Chemistry, lies in the area of computational quantum chemistry and molecular modeling: the investigation of chemical systems at the atomic and electronic level. His work has far-reaching applications in the design of new materials and the development of new chemical processes to improve material composition, safety and utility. To accurately model systems at the atomic and electronic level, the laws of quantum mechanics must be utilized, and such simulations are extremely computationally intensive. Using SHARCNET, this group has been able to simulate the high pressure compression of anti-wear materials protecting a car's internal engine surfaces and understand how they function at the fundamental atomic level. Such simulations, which may lead to the development of anti-wear materials resulting in more fuel-efficient cars, would not have been possible in Canada without SHARCNET. General Motors, Nova Chemicals, Mitsui Chemicals and Defense Research and Development Canada have all taken advantage of Professor Woo's tremendous breakthroughs. For example, Woo helped initiate a molecular modeling program at Nova Chemicals, a leader in the commodity chemical industry, which has aided in the development of new catalysts now used commercially to produce plastics in mass quantities. More info

Computational Chemistry

Developing new therapies for epilepsy, cystic fibrosis

Saul Goldman, University of Guelph

Professors Saul Goldman and Chris Gray, University of Guelph, in collaboration with Robert Guy (National Institutes of Health, Maryland) are investigating the crucial role of ion channelling in physiological processes like the control of blood flow to the brain. SHARCNET computing resources allow these researchers to simulate the mechanism by which ion channels select and transport their ions of choice, which may help scientists develop new therapies for diseases like hypothyroidism, epilepsy, and cystic fibrosis. More info

Computational Chemistry

Essential Enzyme Research

James Gauld, University of Windsor

Dr. James Gauld, Assistant Professor in the Department of Chemistry and Biochemistry at the University of Windsor, along with a team of researchers, is applying computational chemistry to gain a better understanding of how biocatalysts such as protein enzymes and ribonucleic acid catalysts (ribozymes) function. Their research focuses on the unique catalytic mechanism of the class of enzymes known as nitric oxide synthases (NOSs). These enzymes play an important role in the synthesis of nitric oxide(NO), a chemical that serves a vital function in life processes such as brain development, embryo formation and blood pressure regulation. In addition, Gauld and his team of researchers are interested in explaining the mechanisms of catalytic ribonucleic acids (RNA), ribozymes. RNA has long been known as an essential building block of life; now, it turns out it also has a key role in the life-cycles of some viruses, including hepatitis. In some cases, ribozymes have shown potential as therapeutic agents against life threatening viruses, such as HIV. More info

Computational Chemistry

No Barrier Theory

Peter Guthrie, University of Western Ontario

Peter Guthrie, a researcher at the University of Western Ontario’s Department of Chemistry, has developed a method of predicting the absolute rate constants for chemical reactions in solution and is working on simplifying the method in hopes of more widespread use by chemists. Over the years, Guthrie has been the recipient of several awards and fellowships. His current interests focus on the prediction of rate constants for reactions in solution, thermodynamics applied to organic reaction mechanisms, and computational methods for calculating rate and equilibrium constants in solution. More recently, Guthrie has developed the “No Barrier Theory” which he explains is “a qualitative way of thinking about chemical transformations, which makes many things easier to understand and is also a quantitative way to calculate rate constants.” More info

Computational Chemistry

Studies of nanometer-sized systems

Fedor Naumkin, University of Ontario Institute of Technology

Dr. Naumkin's expertise lies in the domains of Theoretical / Computational Chemistry & Chemical Physics, with a particular focus on interdisciplinary nanoscience. The objects of interest are generally specified as polyatomic systems (atomic and molecular complexes and clusters, atoms and molecule on surfaces, interfaces at molecular level). His primary research aims are to predict new stable structures and compositions, to investigate relationships between various properties, and ultimately to design systems with desirable characteristics. Current focus is on unique metal-halocarbon molecular systems with pentavalent C atoms, and novel core-shell cluster nano-systems. Potential practical applications of this work include development of new materials with unique properties, novel catalytic agents, elements of molecular electronics, molecular transport and storage, energy storage at the molecular level. More info

Computational Design and Manufacturing

Metal cutting process optimization

Allan Spence, McMaster University

Research efforts are concentrated in analysis of complex manufacturing geometry problems. For metal cutting process optimization, the cutting tool / part intersection is determined to high accuracy, force / deflections are calculated, and the final tool path feed rate is optimized with the goal of a minimum safe machining time. Normally the part geometry is known a priori, but in cases where distortion occurs, such as after casting cool down, or after die/mould/part heat treat hardening, the remaining uncut material geometric is not known, and conservative material removal rates must remain. To address this, current research is investigating laser digitizing to measure the actual uncut surface geometry, so that material removal rates can be further optimized. For sheet metal forming, laser digitizing and/or passive computer vision are used to measure both part geometry and surface gray level. By printing a regular grid on the undeformed sheet metal, forming, and then measuring the grid deformation, the actual strain values can be validated against Finite Element Model predictions. The metal cutting research work is also part of a five year, NSERC Strategic Network proposal, Canadian Network for Research and Innovation in Machining Technology (CANRIMT) that includes Ontario researchers from McMaster, Toronto, and Windsor. More info

Computational Fluid Dynamics

Nonlinear Internal Waves in the Ocean and in Lakes

Kevin Lamb, University of Waterloo

Dr. Lamb's research group has recently started several projects that require three-dimensional simulations and these rely on SHARCNET resources to various degrees. There projects include: internal wave interactions with eddies, idealized simulations of physical processes in small lakes, simulations of physical and bio-geochemical processes in Lake Erie. That latter, funded by an NSERC Strategic Grant, would not be possible without SHARCNET resources. A graduate student of Lamb has begun doing 3D simulations of tide-topography interactions in the presence of eddies. The tide topography interaction generates internal waves which will interact with the eddies. Says Lamb, "Without SHARCNET, I simply would not be able to do high resolution simulations of lake hydrodynamics. The Lake Erie project would likely be undoable". More info

Computer Algebra

Modeling differential equations

Alexandre Odesski, Landau Institute for Theoretical Physics

Dr. Odesskii is one of SHARCNET's Research Chairs in high performance symbolic computing. Mathematics, science and engineering all rely on differential equations. One type of differential equation, called an integrable equation, only analyzes one spatial dimension and time, and its applications can be limited as a result. Dr. Odesskii is hoping to change that paradigm by adding another dimension to these equations. Because actions unfold in three dimensions, adding another dimension into these differential equations will make the models more representative of the real world. He is working with SHARCNET to develop new equations that can help solve several real world problems by developing accurate models of real world phenomenon. Exactly where these new equations will be applied is not yet known. “A new equation in pure mathematics is like working on a piece of art,” says Odesskii. “You can’t focus on the practical applications; they will come once the equations have been sorted out.” More info

Computer Security And Privacy

Algebraic combinatorics, algorithms, and digital identity

Angele Hamel, Wilfrid Laurier University

A recent recipient of the Petro-Canada Young Innovator Award at Laurier, Angèle Hamel is an internationally recognized researcher in combinatorics--the mathematics behind computer science. Her research supports the creation, discovery, and application of new results in combinatorics and algorithms. Combinatorics and algorithms work hand in hand to increase understanding of various objects in computer science--networks, data structures, and permutations. Combinatorics provides the structure and framework necessary for these objects as well as the means of analyzing the structure. Algorithms provide the means of manipulating these objects. Combinatorics supplies a unique way of looking at the world; algorithms supply a unique way of interacting with the world. Together they allow fundamental insights and establish essential connections. More info

Computing Science

Developing computational methods for the analysis of high-throughput sequencing data

Lucian Ilie, University of Western Ontario

Lucian is working on an investigation of the mathematical properties of words (not necessarily English words, just any combination of letters, like "ATTTGTCCAA"). What he's doing actually has application to understanding the structure of DNA sequences, so a major breakthrough here could prove very interesting for molecular biologists as well. Lucian is collaborating on solving an important conjecture with researchers from France, UK, and Canada. More info

Condensed Matter Physics

Condensed Matter Theory, Statistical Mechanics

Michel Gingras, University of Waterloo

Michel Gingras' research group has been working on models of randomly disordered magnetic systems called "spin glasses", with a focus on models where the interactions between the magnetic moments are given by long-range magnetostatic dipole-dipole couplings. The main question that they are addressing on these systems is whether a thermodynamic spin freezing transition is occurring at nonzero temperature. The purpose of these studies is to make contact with experiments on magnetic materials where the principal interactions are such long-range magnetostatic dipole-dipole interactions. There have been discussions both on the experimental and numerical fronts for over twenty years as to whether or not there exists a thermodynamic spin freezing transition in random dipolar spin glass systems. Their results are the first ones which provide compelling evidence that there is indeed such a transition. More info

Condensed Matter Physics

Creating more efficient, cost-effective electronics

Eugene Kim, University of Windsor

The research interests of Professor Eugene Kim, a SHARCNET Chair of Biocomputing at the University of Windsor, lie in the electronic properties of condensed matter systems. With the desire to increase the speed of computers while reducing their size, modern-day manufacturers have been doubling the number of transistors on a computer chip approximately every eighteen months. Up until now, manufacturers have managed to do this simply shrinking the already existing technology. However, this approach will soon have to be abandoned in favour of new technology; new devices, which will replace silicon-based transistors. One possible route is molecular electronics; utilizing molecules in electronics devices. This is the ultimate step toward miniaturization, and it offers the potential of drastically reducing production costs. Professor Kim, who focuses on molecular electronics, as well as disordered electronic systems, quantum impurity problems, and quantum magnetism, is performing calculations to determine systems and materials that are promising for these device applications. More info

Condensed Matter Physics

Developing the conductive materials for infinitesimally small electronic devices

Erik Sorensen, McMaster University

Erik Sorensen, of the Department of Physics and Astronomy at McMaster University, is a SHARCNET Chair in Computational Materials. His research focuses on two major areas: the different states of matter and nanotechnology. With his analysis of the critical properties of matter in different states, he could potentially discover materials not currently in existence. Nanotechnology is the multi-billion dollar industry intent on developing devices and technology in sizes invisible to the human eye. In the area of nanotechnology, Sorensen is investigating the effects of impurities on the transport properties of materials; his research so far suggests the plausibility of creating a single-electron transistor. More info

Condensed Matter Physics

Improving optical devices

Marek Wartak, Wilfrid Laurier University

Marek Wartak, a Professor at Wilfrid Laurier University, studies photonics: specifically the modeling, design, and simulation of semi-conductor lasers with the ultimate goal of discovering more cost effective solutions in computational photonics (for use in laser printers, compact disc players etc). Wartak has plans to commercialize the results of his research with the Waterloo-based company Lasma Inc. More info

Condensed Matter Physics

Numerical modeling of linear and non-linear optical responses of semiconductors

Anatoli Chkrebtii, University of Ontario Institute of Technology

Prof. Chkrebtii is internationally renowned for his innovative research that correlate at unprecedented level dynamics, electronics and optics of modern materials. This includes bulk of crystalline and non-crystalline semiconductors, and 2D, 1D or zero dimensional nano-systems. Being a coauthor of two monographs and more that 100 publications in most distinguished research journals (Phys. Rev. Lett., Appl. Phys. Lett., Phys. Rev. B and other), Prof. Chkrebtii's "advanced materials properties in motion" research field is based on the combination of most advanced computational techniques in solid state physics. It uniquely combines parameter-free (i) molecular dynamics to track materials’ main properties and their evolution at the most detailed atomic level from absolute zero to a temperature above a melting point; (ii) electronic structure tools to follow distribution of an individual electron, its transitions and related single-electron chemistry; (iii) advanced signal processing techniques that decodes non-equilibrium temperature, pressure, and external radiation dependent characteristics, (iv) sophisticated computer visualization and (v) most comprehensive description of the optical response for various materials. In particular, his recent focus is on hydrogen-bonding, ubiquitous in physical, chemical, and biological sciences. Such comprehensive combination represents a turning point in a very accurate description and understanding of the structural, dynamical and chemical properties of hydrogen in confined systems, including nano-composite materials, including extreme conditions. His research, being fundamental in its nature, is always directed toward explanation of the experimental results or their application in various fields. For instance, our recent two year long “Sonus/PV (Photo-Voltaic) Highway Traffic Noise Barrier” research project has demonstrated a feasibility of combining the noise reduction barriers along 400 highways with solar cells to produce a green energy. Currently they are working on commercialization of the results.

Condensed Matter Theory

Quantum matters

Roger Melko, University of Waterloo

Roger Melko's interests involve strongly-correlated many-body systems, with a focus on material properties, ground state phases, phase transitions, and quantum criticality. His work has employed Monte Carlo simulations and Density Matrix Renormalization Group methods to explore the low-temperature physics of classical and quantum magnetic materials, cold atoms in optical lattices, bosonic fluids and low-dimensional systems. He is particularly involved in studying microscopic models that display interesting quantum behavior in the bulk, such as superconducting, spin liquid, topological, superfluid or supersolid phases. Roger is also interested in broader ideas in computational physics, the development of efficient algorithms for simulating quantum mechanical systems on classical computers, and the relationship of these methods to the field of quantum information science. See the article on Solid, Liquid, Gas... Superglass? More info

Cosmology

Modeling the universe

Hugh Couchman, McMaster University

Hugh Couchman is a SHARCNET Founder, Fellow of the CIAR cosmology program and Professor in the Department of Physics and Astronomy at McMaster University. His research aims to simulate the formation of cosmic structure and, in particular, to describe the formation and evolution of the cosmic population of galaxies. His work ranges from the extremely small, the constituents that comprise a galaxy, to the infinitesimally large, clusters and superclusters of galaxies themselves. More info

Dark Matter

Numerical Simulations of Dark Matter Halos

James E. Taylor, University of Waterloo

There is now overwhelming evidence from many different types of astronomical observations that our Universe consists almost entirely of exotic forms of energy and matter, unknown on Earth. In particular, observations of individual galaxies, clusters of galaxies, and of the Cosmic Microwave Background, together with a knowledge of the local abundance of light elements, make it clear that 85% or more of the matter in the universe is dark matter, an exotic substance that does not produce or interact with light in almost any form. The properties of dark matter are naturally hard to determine, given its effective invisibility; it manifests itself only through gravity. The main goal of Taylor's research programme is to model the spatial distribution of dark matter around visible structures such as galaxies and clusters. He has developed novel computational methods to do so, methods which provide a 10,000 increase in speed over traditional approaches. Taylor is now starting to apply these techniques to a broad range of astrophysical puzzles, ranging from the nature of high-redshift galaxies to the distribution dark matter in the vicinity of our solar system. More info

Distributed Systems

Computer and Information Science

Michael Bauer, University of Western Ontario

Dr. Bauer’s research in high performance computing includes the development of novel algorithms for the efficient uses of high performance computing grids, such as SHARCNET, and the development of novel parallel algorithms that can be used to analyze multiple, large multi-modal data sets, particularly in the environmental and medical informatics areas. A recent paper, “Job Co-Allocation Strategies for Multiple High Performance Computing Clusters” which appeared in the journal Cluster Computing, reported on new algorithms for the efficient distribution of multi-process jobs across different clusters when considering communication costs and differing communication patterns. Other recent work on a new approach for the parallel computation of the atmospheric diffusion of plumes has appeared in the International Journal of Computers and Applications. More info

Economic Modelling

International and Empirical Finance

Alex Maynard, University of Guelph

Alex Maynard specializes in both Applied and Time Series Econometrics with research interests in International and Empirical Finance. His research has focused on the forward premium puzzle, predictive testing in empirical finance, long-horizon regressions, causality testing, and the use of instrumental quantile regression methods to study the effect of public insurance on savings across the wealth distribution. More info

Ecosystem Simulation

Evolving predator-prey ecosystem simulation

robin gras, University of Windsor

Robin Gras and his research team study the evolutionary process and the emergence of species in a simulated ecosystem. They have conceived an individual-based evolving predator-prey ecosystem simulation. The agents evaluate their environment (e.g., distance to predator/prey, distance to potential breeding partner, distance to food, energy level), its internal state (e.g., fear, hunger, curiosity) and chooses among several possible actions such as evasion, eating or breeding. The behavioral model of each individual is unique and is the outcome of the evolution process. One major and unique contribution of this simulation is that it combines a behavioral, an evolutionary and a speciation mechanism. This is the only simulation modeling the fact that individual behaviors affect evolution and speciation. Their approach allows interesting studies on theoretical ecology in collaboration with biologists. For example, this approach is used to study the species abundance distribution, patterns and rates of speciation, the evolution of sexual and asexual populations, the interaction and diffusion of an invasive species or a disease in an existing ecosystem, etc. More info

Embryo Development

The Mechanics of Early Embryogenesis

G. Wayne Brodland, University of Waterloo

During early embryo development, cell monolayers called epithelia undergo self-driven changes of shape in order to form organs and other critical structures. Sometimes, for reasons not totally understood, irregularities arise in these morphogenetic movements giving rise to malformations, which in humans include spina bifida and cleft lip and palate. The goal of Dr. Brodland’s research program is to use computational modeling and experiments to better understand the mechanics of embryonic tissues and their movements. The group uses finite element methods, matrix algebra and optimization schemes in their work; currently active programs consist of some 250,000 lines of C++ code. Dr. Brodland collaborates with biologists and others from Princeton, Harvard, the Max Planck Institute, King’s College, Vanderbilt and University College London. This research has direct implications for embryology, wound healing, tissue engineering and cancer metastasis. They hope what they learn will lead to improved clinical strategies for preventing birth defects and advancing human health. More info

Environmental Modelling and Analysis

Understanding and preventing climate change

David Swayne, University of Guelph

David Swayne, a Professor in the Department of Computing and Information Science at the University of Guelph, is using SHARCNET to investigate and model regional lake thermal models. As anyone who lives near the Great Lakes knows, the heat storage of lakes means warmer, snowy weather in early winter and a cooling effect in the spring. Canada is home to over 1 million lakes, twenty percent of the world's entire supply of freshwater. Nearly eighteen percent of these lakes are located in Ontario. Professor Swayne's models gauge the temperature variations of these lakes and their ensuing environmental implications. Understanding climatic systems and taking precautions to prevent extreme climate change are critical to nearly all facets of our lives, from our water supply and the crops we grow to the quality of life in our cities. This type of environmental modeling involves the development of a mathematical model to simulate the main physical and temperature-related characteristics of a particular climate. However, climate change modeling is extremely computationally intensive and, according to Swayne, his work would be nearly impossible without SHARCNET. More info

Finance; Asset Pricing

Constructing an optimal theory for financial portfolios

Matheus Grasselli, McMaster University

Matheus Grasselli, SHARCNET Chair in Computational Finance and Assistant Professor, Department of Mathematics and Statistics at McMaster University, is studying financial mathematics and derivatives and relating them back to real-life applications. Grasselli's research is focused on pricing and hedging techniques in incomplete markets, where traditional arguments based on replicating portfolios are not applicable. His research looks at risk preferences of investors and market participants to identify the best trading strategies to be followed when non-standard financial instruments are created and introduced. On a more practical level, Grasselli's research can be used to address socio-economic problems. For example, an innovative United Nations program applies the concepts of derivatives to rainfall levels to help manage the risk of starvation caused by drought in Africa. SHARCNET helps by generating large-scale simulations of market conditions on which approximated solutions can be implemented and tested for efficiency and accuracy. More info

Financial Mathematics

Energy markets and Real-Options Theory

Matt Davison, University of Western Ontario

A tenured professor in the Departments of Applied Mathematics and Statistical & Actuarial Sciences at the University of Western Ontario, Dr. Matt Davison studies the mathematical modeling of uncertain energy markets. His main research interests are the application of financial mathematics to energy markets and the uses of Real-Options Theory applied to various areas of research ranging from property and casualty insurance to defense sciences to cancer research. Dr. Davison, a Canada Research Chair in Quantitative Finance and the project leader for the MITACS Modeling Trading and Risk in the Market project, uses mathematics to model the realistic fluctuations in the natural gas industry as governed by the invisible hand of the market. His research is based on decisions made under realistic conditions that dictate the ways in which energy is generated, used and sourced in the uncertain, continually changing future. More info

Fluid Dynamics

Improving weather prediction using fluid dynamics

Bartosz Protas, McMaster University

Bartosz Protas, an Assistant Professor of Mathematics and Statistics and SHARCNET Chair in HPC Tools at McMaster University, focuses on the integration of Control and Estimation Theory and Computational Fluid Dynamics into a single unified framework. Important applications of this research included data assimilation in weather prediction. High performance computational facilities, such as provided by SHARCNET, are indispensable in the study of such problems. More info

Fluid Dynamics

Turbulence and combustion modeling

Stephen Tullis, McMaster University

With growing international pressure for Canada to reduce its greenhouse emissions, the private and public sectors are both actively seeking out new energy strategies and technologies—from cleaner burning fuels to more effectively harnessing our renewable natural resources. Wind turbine research is a federal and a provincial priority area that is important to environmental science and technology as well as the natural resources and energy industries. Stephen Tullis, a researcher at McMaster University, is focused on using mechanical engineering with computational and programming skills to help create greener technologies. Tullis’ general research interests lie primarily in the field of applied fluid dynamics—the physics of liquids and gases in terms of motion, combustion, and with other bodies. In his research recent projects he makes use of traditional laboratory experiments and Computational Fluid Dynamics (CFD) to examine more complicated, large-scale computer simulations. CFD allows researchers to build realistic computational models for systems or devices of interest—essentially virtual prototypes—based on real-world fluid flow physics and chemistry. Recently, Tullis’ research has focused on examining turbine turbulence—one of the main limiting factors in turbines—and on reducing the emissions from gas turbines and spark ignition and diesel engines. Both of these research projects rely directly on the computational power provided by SHARCNET. His group is using CFD methods to model highly complex flows using large grids over simulated long periods of time. More info

Fluid Mechanics

Modeling the movement of urban pollution

Nicholas Kevlahan, McMaster University

A Professor in McMaster University's Department of Mathematics, Nicholas Kevlahan is developing a model for calculating turbulent flows around complicated shapes like buildings and airplanes. His research has implications for aerodynamics, combustion, urban pollution, weather prediction, and climate modeling. It even applies to such biological functions as the movement of blood around the human heart. More info

Fluid Mechanics

Urban dispersion of air pollutants

Fue-Sang Lien, University of Waterloo

In an ongoing project supported by Chemical, Biological, Radiological and Nuclear Research and Technology Initiative (“CRTI”), a prototype multi-scale modeling system to study urban dispersion of air pollutants (or toxic gases) has been developed. This system is currently being used to model dispersion of exhaust fumes from traffic in the ubiquitous street canyon in Toronto area, which is an important aspect of urban air-quality studies. This project is in collaboration with scientists from the Ministry of the Environment Ontario in Toronto. The mesoscale community model WRF (the Weather Research and Forecasting Model) developed by National Center for Atmospheric Research (NCAR) among other US organizations will be installed on SHARCNET in order to facilitate the coupling between the mesoscale WRF code and a microscale Computational Fluid Dynamics (CFD) code developed by Prof. Lien’s research group. More info

Fuzzy Option Pricing

Creating models to aide in human decision-making

Marc Kilgour, Wilfrid Laurier University

Kilgour, trained originally as an engineer, describes his research interests as lying at the intersection of mathematics, social science, and engineering. He is collaborating with two systems design engineers on modelling the costs and benefits of brownfield redevelopment, funded by an NSERC strategic grant. Brownfields are abandoned or underutilized properties that may contain hazardous materials left from previous use. Municipalities encourage brownfield redevelopment, but developers are often reluctant since the risks and attendant costs are difficult to establish. The purpose of their project is to devise a decision support system to provide more accurate estimates of costs. A UW graduate student, working with Kilgour, is using the SHARCNET HPC systems as part of this research. More info

Genetic Linkage Analysis

Genetic basis of specific human diseases

Robert A. Hegele, John P. Robarts Research Institute

Rob Hegele is an endocrinologist with an interest in lipidology and diabetes. His laboratory has studied the genetic basis of diabetes and atherosclerosis in Canadian sub-populations and aboriginal communities. His lab discovered the molecular genetic basis of 12 human diseases, including hepatic lipase deficiency, Oji-Cree type 2 diabetes, familial partial lipodystrophy and endocrine-cerebro-ostedysplasia and has described >100 human mutations causing dyslipidemia, diabetes and atherosclerosis. The identification of new genes that can cause heart disease, stroke and such high risk conditions as diabetes, hyperlipidemia and hypertension leads to a new understanding of the pathways for these diseases and is the first step in the development of new diagnostic methods and treatments. More info

Geophysical Fluid Dynamics

Calculating currents to save whales

Francis Poulin, University of Waterloo

Breaching the boundaries between academic disciplines, Francis Poulin has waded into physics, chemistry, biology, and computer science in an effort to understand, predict, and describe the current through which the whales navigate each spring. “I’m trying to understand the structure of the current, using mathematical equations to describe basic physical properties, such as momentum.” To solve such complex equations, he writes computer programs, draws on his background in physics, and is learning more about biology. Some day he may be able to create models to predict when plankton blooms will occur. As well, the research could lead to a greater understanding of other oscillating currents. Such currents are common in coastal waters, and have physical properties that are more complicated than currents flowing in one direction. More info

Global Ecological Change

Biodiversity, restoration ecology, global ecological change and conservation of forests

Madhur Anand, University of Guelph

Dr. Anand is an Associate Professor & Canada Research Chair, Department of Environmental Biology at the University of Guelph. The rapid pace and far-reaching potential of ecological change in the face of globalization and climate change is creating stress within ecosystems all over the world. There is an ever-increasing threat of extinctions of species, communities, and ecosystems along with the biodiversity services and functions that they support. Understanding the complex effects of ecological change at many scales (local to global, recent to historical) on biodiversity is thus of critical importance for predicting human-mediated changes to the environment, conserving biodiversity heritage and sustaining global economies. The CRC in Global Ecological Change recognizes the increasing international stature of ecological problems and their solutions. The research program will highlight similarities and differences between ecological events and practices on different continents with a view to fostering knowledge exchange. It will also provide tools for the application of emerging interdisciplinary theories, improved quantitative methods and predictive simulation models to the preservation, recovery and restoration of perturbed and vulnerable forest ecosystems worldwide. More info

Groundwater Flow and Transport Modeling

Protecting ground and drinking water

Edward Sudicky, University of Waterloo

The water tragedy in Walkerton, Ontario reminded North Americans of just how crucial water safety is, and more, how increasingly complex it is to protect this precious resource. Edward Sudicky, a Canada Research Chair (CRC) in Quantitative Hydrogeology and Professor of Earth Sciences at the University of Waterloo, is using High Performance Computing to do precisely that. Professor Sudicky, an internationally recognized leader in Hydrogeology (the occurrence, movement and quality of water beneath the surface of the Earth), uses computational models to describe fluid flow and the migration of contaminants in subsurface water (groundwater). The models developed by this group over the past 20 years are used worldwide at universities, research institutions, government agencies and in the private environmental consulting industry, but because hydrogeology deals with a complicated subsurface environment, it is extremely computationally demanding. Sudicky's research has wide-ranging implications, all of which, he confirms, are dependant on HPC and that demand is increasing exponentially. Future environmental issues such as the impact of climate change on water resources, or the safe disposal of high-level radioactive wastes, will be affected by the availability of HPC resources. More info

Image Processing

Improving image-guided surgery, minimizing invasive animal and human testing

Terry Peters, John P. Robarts Research Institute

In the research lab of Professor Terry Peters of the Robarts Research Institute, SHARCNET is being used to create a new technology-based surgical environment. The Laboratory for Virtual Augmentation and Simulation for Surgery and Therapy (or VASST Lab) is intended to provide an environment where teams of scientists, engineers and clinicians can combine their knowledge and expertise to simulate the characteristics of living organs and tissues, and test alternative treatment strategies without performing tests on living subjects. Ultimately, researchers will be able to simulate entire organ systems virtually and perform minor surgical procedures without the use of human or animal models. SHARCNET has and will continue to play a significant role in establishing these procedures, including algorithm development, implementation, adaptation and parallelization for tissue modeling and image registration (overlapping medical images fused to make comprehensive anatomical models). Though there are a multitude of applications for this kind of research and research facility, the VASST Lab focuses on the development of new biomedical procedures, the creation of a virtual environment to guide minimally-invasive surgery and the training of a new generation of clinicians in these cutting-edge applications. More info

Infectious Disease Modelling

Modelling influenza dynamics within hosts and in vitro

Catherine Beauchemin, Ryerson University

Influenza (flu) is a growing concern for health authorities worldwide. The annual cost of flu illness and the threat of an imminent pandemic make it all the more necessary to better understand the mechanisms that drive this disease. The objective of this research is to describe and understand how various factors can affect the spread of flu within an individual. This will be done by suggesting hypotheses in the form of mathematical and computer models and seeing how their behaviour compares to that of experimental systems. For example, fitting the models to experimental data will help determine key parameters of an influenza infection like viral production rate which would be difficult or impossible to obtain experimentally. The research concentrates on 4 particular aspects of the disease: 1) how anti-influenza drugs act on the various phases of virus reproduction and thus influence viral population growth; 2) how to accurately model and predict the emergence of drug resistance; 3) how the variety of cell types in the lungs affects how fast the disease spreads in certain parts of the lungs compared to others; and 4) how the chemical signals (chemokines) released by the cells infected with or combating the flu infection are affecting how sick a patient will feel or how quickly she/he will recover. Answers to these questions will help us get a better understanding of what drives this disease and how we can best combat it. Additionally, the models and approaches that will be used in this research should be applicable to other respiratory diseases such as SARS and the metapneumovirus. More info

Information Visualization

Bridging the linguistic visualization divide

Christopher Collins, University of Ontario Institute of Technology

Dr. Christopher Collins is an assistant professor at the University Of Ontario Institute of Technology and was recently named SHARCNET Research Chair in Information Visualization. As text visualization continues to permeate our lives we need to improve the way we present textual relationships. HPC allows for efficient processing of computationally taxing data sets such as digital libraries; a necessary tool when developing new large scale visualization systems. Using his background in computational linguistics, techniques from information visualization, and research in human-computer interaction, Dr. Collins is working towards creating a toolkit for analysts, scholars, scientists, and everyday internet users to better understand the content and relationships within large scale repositories. More info

Machine Learning

Artificial Intelligence and Machine Learning

Yoshua Bengio, Université de Montréal

Dr. Bengio's long-term goal is to understand intelligence; understanding the underlying principles would deliver artificial intelligence, and he believes that learning algorithms are essential in this quest. Machine learning algorithms attempt to endow machines with the ability to capture operational knowledge through examples, e.g., allowing a machine to classify or predict correctly in new cases. Machine learning research has been extremely successful in the past two decades and is now applied in many areas of science and technology, some well known examples including web search engines, natural language translation, speech recognition, machine vision, and data-mining. Yet, machines still seem to fall short of even mammal-level intelligence in many respects. One of the remaining frontiers of machine learning is the difficulty of learning the kind of complicated and highly-varying functions that are necessary to perform machine vision or natural language processing tasks at a level comparable to humans (even a 2-year old). More info

Machine Learning

Statistical machine learning and biologically-inspired computer vision

Graham Taylor, University of Guelph

The pervasiveness of computing has resulted in the production and storage of more data than ever before. Machine learning seeks to transform this deluge of data into intelligent systems that identify patterns and make decisions. It has revolutionized fields as diverse as computer vision, computational neuroscience, biology and the social sciences. But when faced with data that is increasingly complex, how does a machine know which parts are relevant? How does it structure the millions of components into organized units on which it can base decisions? Taylor's research, in the field of deep and unsupervised feature learning, confronts this challenge by developing algorithms that learn increasingly abstract layers of representation without human guidance. A theme throughout his work is applying deep learning to sequence modeling, in particular, interpreting human movement either through highly structured data such as motion capture (mocap), or weakly structured data such as video. Taylor has made fundamental contributions to deep and unsupervised feature learning, proposing models for complex, high-dimensional data and has leveraged these models to produce state-of-the-art results in traditional computer vision problems such as object recognition, person tracking, and activity recognition. He has also proposed novel, interdisciplinary applications, for example, active crowd-sourcing and immersive, vision-based crowd games. More info

Materials Simulations

Development of lightweight aluminums

Nikolas Provatas, McMaster University

Nikolas Provatas, an Associate Professor of Materials Science and Engineering at McMaster University, in collaboration with Alcan International and the Centre for Automotive Materials Manufacturing (CAMM), is using SHARCNET to investigate alloy microstructure, and how material properties can be modelled and simulated to ultimately produce stronger yet lighter-weight aluminums. There is significant competitive advantage to be gained by Ontario industries through the application of Provatas' research results. More info

Materials Simulations

Liquids and soft matter

Ivan Saika-Voivod, Memorial University of Newfoundland

Ivan’s research group focuses on three areas: nucleation and glassy dynamics in bulk liquids; thermodynamics of and nucleation in small clusters; and gelation and phase behaviour of colloidal dispersions. More recently, through collaboration with experimental groups, they have embarked on more applied projects including organic heterojunction solar cells, and modeling the structure of hydroxyapatite, the mineral component of bone. They have collaborations with researchers at the University of Rome La Sapienza on properties of colloids. Past projects that have primarily been supported by SHARCNET spawn new, related projects that are naturally continued best on SHARCNET infrastructure. Says Ivan, “The SHARCNET staff have always been helpful in getting code running effectively on the clusters. The staff are an important part in making good use of the facilities." More info

Mathematical Biology

Advancing the treatment of bacterial infections

Hermann Eberl, University of Guelph

A Professor at the University of Guelph's Department of Mathematics and Statistics and SHARCNET Chair in Biocomputing, Hermann Eberl uses computers and mathematical models to simulate what are called "bio-processes". These bio-processes involve the growth, reproduction, and decay of micro-organism communities that live in a layer of slime on top of water. His research could lead to the optimization of waste-water management and to a change in the treatment of bacterial infections. It also has implications for food safety controls. More info

Mathematical Biology

Investigating immune systems and drug resistance

Lindi Wahl, University of Western Ontario

In the field of Applied Mathematics, The University of Western Ontario Professor Lindi Wahl uses SHARCNET to perform Monte Carlo simulations of HIV and immune systems dynamics and large scale simulations of experimental evolution. A central problem in immune system dynamics is the fact that imperfect adherence to prescription protocols may cause the emergence of drug-resistant viral strains. Wahl's research may have applications in the development of new drug therapies to both treat and prevent drug resistance for devastating diseases like HIV. More info

Mathematical Biology

Understanding outbreaks of diseases, such as SARS

David Earn, McMaster University

David Earn, McMaster University Professor of Mathematics and CHIR New Investigator, uses mathematics to model infectious disease transmission. He has recently worked closely with SARS researchers. While it is impossible to predict when the next outbreak might occur, according to Earn, it is important to learn more about the spread of diseases, and our ability to resist them, in order to meet the next threat head-on. Earn and his group have recently shown that simple models can explain the outbreak patterns of rubella, chicken pox and whooping cough. He has also developed a new way to design vaccination programs. More info

Mathematical Modeling & Simulation

Improving coats and crystals and human cells

Huaxiong Huang, York University

Huaxiong Huang, an Associate Professor in the Department of Mathematics and Statistics at York University is using SHARCNET to collaborate with researchers across a wide range of disciplines and industries to produce innovative mathematical models. Dr. Huang's collaboration with researchers at City University of Hong Kong and Hong Kong Polytechnic University is yielding mathematical models that will help the textile industry test new fabrics that are less resistant to moisture removal. In his work on human cells, specifically the cellular mechanisms that allow the mammalian brain to function, Dr. Huang is attempting to construct mathematical models to help researchers better understand the workings of the brain. Dr. Huang is also collaborating with Firebird Semiconductor Ltd., a global leader in the production of semiconductor wafers for radiation detection, to model crystal growth. This project, also supported by researchers at York, UBC and Penn State, will enable industry to detect how defects form in crystals during the growth process and enable the many industries to grow better, more cost-effective crystals for electronics applications. More info

Mathematical Modeling & Simulation

Modeling the Spread of H1N1 Flu Virus

Chris Bauch, University of Guelph

Prof. Chris Bauch is one of a handful of researchers in Ontario who were approached last fall by the Ontario Agency for Health Protection and Promotion (OAHPP) to create mathematical models to help with the province's pandemic preparedness planning. In recent weeks, they've stepped up the collaborative effort to focus on the transmission of H1N1 in order to predict the spread of the disease. More info

Mathematical Modeling & Simulation

Spatial Stochastic Models

Michelle Boué, Trent University

Spatial stochastic models for the spread of infectious diseases have been studied for several decades. However, it is only recently that researchers have taken into account the vital role that the movement of individuals plays in the spread of epidemics. Dr. Boue is working on simulations for one such model, where healthy and infected particles perform simple random walks on a lattice and follow some infection rules. The purpose of the research is to understand how the extinction probability for the epidemic (that is, the probability that after some time there are no infected particles) changes as the parameters in the model vary. Moreover, the effect of immunization, recovery, mutations and control measures are incorporated in the model in different ways. The ultimate goal of this research is to contribute to the realistic mathematical modeling of epidemics.

Mathematics

Development of new financial models to manage risk

Joe Campolieti, Wilfrid Laurier University

In the area of Financial Mathematics, Professor Giuseppe Campolieti, SHARCNET Chair at Wilfrid Laurier University, is investigating the development and application of new analytical and computational methodologies for tackling the pricing of complex financial derivatives. His research advocates innovative methods and algorithms with applications largely impacting the frontiers of derivatives pricing and the management of risk. These results will ultimately be of use in commercial software applications for financial engineering and better management of derivative risk. More info

Medical Physics, Radiation Sciences

Managing Radiation Therapy for Cancer Patients

Emily Heath, Ryerson University

Dr. Heath's research is concerned with improving the accuracy of radiation therapy for the treatment of cancer. This involves identifying different uncertainties in the radiation treatment processes and developing methods to minimize their effect. One current area of research Dr. Heath is pursuing is the question of how to manage respiratory motion of patients during radiation treatment. In radiation therapy they try to deliver a very high dose of irradiation to the tumour but at the same time it is critical to minimize the dose deposited in surrounding tissues to avoid side effects from the treatment. Organ motion can blur or distort the very conformal dose distributions that are delivered to the patient. For proton therapy and treatment with heavy charged particles, the effects are more severe. With the use of 4D optimization methods and simulations of dynamic dose deliveries this research group is developing treatment planning methods to account for respiratory motion. Another area of the research addresses the accuracy of different tools that are used for 4D treatment planning. This includes investigating different approaches to calculating dose in deforming anatomy and also the use of deformable image registration methods in radiation therapy planning. More info

Molecular Simulation

Furthering the field of nanotechnology

Martin Muser, University of Western Ontario

SHARCNET Chair in Computational Materials and Professor in The University of Western Ontario's Department of Applied Mathematics, Martin Muser studies the effects of extreme conditions, such as pressure, on minerals. His work with silica and other similar materials have applications for the booming industry of nanotechnology (the science and technology of building electronic devices from single atoms and molecules). Nanotechnology has the potential to revolutionize electronic devices in every industry but most significantly in health care, automobile manufacture, computing, and aerospace. More info

Music Data Analysis

The people's playlist

Matthew Woolhouse, McMaster University

Last October, Prof. Woolhouse started a five-year project funded by telecommunications giant, Nokia. Inside McMaster's new Digital Music Lab, Woolhouse and his team peer into the company's closely guarded data and study the music-downloading habits of its mobile-phone users. More info

Nanoparticles And Nanomaterials

Nano-medicine targets diseased cells

Pu Chen, University of Waterloo

How to cure the disease without killing the patient: that is a major hurdle facing medicine today and is one that Pu Chen is helping to resolve. As a Canada Research Chair in Nano-biomaterials, the chemical engineering professor conducts research straddling the fields of nanotechnology and biomedicine. More info

Neural Networks And Computational Neuroscience

Neural Computation, Aging, and Memory (NCAM)

Diano Marrone, Wilfrid Laurier University

A major shift in demographics occurred as life expectancy lengthened from 40-50 years in the early 1900s to nearly 80 years at the end of the last century. While succeeding generations live longer, it has become clear that discernible cognitive impairment (i.e. memory loss) afflicts a substantial proportion of the elderly, even in the absence of disease. An important tool for understanding the effects of normal aging is the use of laboratory animals, since they develop similar memory deficits to humans, but do not naturally develop neurodegenerative conditions such as Alzheimer's disease. Thus, the naturally occurring impairment in aged rodents is a powerful model for age-related cognitive decline. While neuron loss was once thought to be an inevitable consequence of aging, recent research demonstrates no significant loss in medial temporal lobe structures of aged rats or primates, including humans, in the absence of disease. Thus, the critical challenge is to uncover physiological factors that cause cognitive impairment. ...modeling may provide the critical ability to not only manipulate each of these changes individually in order to characterize their impact on the instability of activity in the aged hippocampus, but also to systematically examine the combination of these changes in order to pinpoint emergent properties of these collective alterations in hippocampal function.

Neuroimaging, Brain Mapping

CBRAIN

Alan Evans, McGill University

Dr. Evans' group is developing a Canada-wide platform for distributed processing, analysis, exchange and visualization of brain imaging and other neuroscience data. Even though the focus of CBRAIN is providing tools for use by brain imaging researchers, the platform is generalizable to other imaging domains, such as radiology, surgical planning, heart and body imaging, with profound consequences for Canadian medical research. The expected result is a middleware platform that will render the processing environment (hardware, operating systems, storage servers, etc) transparent to a remote user. Interaction with a standard web browser allows applying complex algorithm "pipelines" to large datasets stored at remote locations using a mixture of network available resources such as small clusters, neuroimaging tools and databases as well as Compute Canada's High Performance Computing Centres (HPC). Dr. Evans also heads an international initiative, GBRAIN. More info

Numerical Methods

Computer and Information Science/Business

Peter Forsyth, University of Waterloo

The current market meltdown is widely attributed to poor hedging and mispricing of credit derivatives. Current models make the assumptions that credit quality moves randomly, but changes are small in small time intervals; and it is always possible to trade credit derivatives and bonds, so that a dynamic hedging strategy can be carried out. Most people realize (and have for a long time) that these two assumptions are unrealistic. The underlying process for credit quality should include jump processes, i.e. large discontinuous changes. Think of the credit quality of Lehman Brothers the day before they went bankrupt. Many traders have been quoted as saying "this market meltdown is a five sigma event," meaning that the current events are highly unlikely. Actually, if one assumes that the process followed by credit includes jump processes, the current meltdown is not so improbable; we can expect such situations about once every ten years. Markets can be ill-liquid, which means that it may not be possible to sell something at any value except zero. It is possible to reformulate the standard models as a problem in optimal stochastic control, and solve this problem for an optimal hedging strategy. This more general approach can handle jump processes and liquidity effects. The availability of SHARCNET computing resources makes it possible to test and validate these more general market models. More info

Numerical Solution Of Differential Equations

Calculations take flight

Lilia Krivodonova, University of Waterloo

Lilia Krivodonova use computers to tackle scientific problems that have proved impossible to solve for more than 100 years. “Most equations of practical interest can’t be solved exactly,” she says. “It’s not only too difficult, but theoretically impossible. When we build or compute something, it’s never exact. There’s always some error.” Krivodonova’s current focus is gas dynamics, a field that uses differential equations to describe the flow of air over an object, such as an airplane. Using high-powered computers, she calculates how to make air flow around the plane more efficiently — providing more lift and less drag. More info

Parallel Statistical Computing

Long-term Mortality Forecast

Hao Yu, University of Western Ontario

The prediction performance of the Lee-Carter (LC) model for long-term mortality forecast is the focus. In order to make a sound assessment, Dr. Yu's group set up a backtesting methodology to evaluate the prediction performance of the LC model. They propose to use the Kolmogorov-Smirnov (KS) test to assess how close the percentile histogram resembles uniform distribution, which can complement the assessment of probabilistic prediction. Two issues are addressed with implementing the LC model: robustness and drift uncertainty. Dr. Yu's group proposes quantile regression (QR) for robust parameter estimation of the model for time varying index kt, using the boot- strap method to incorporate the drift uncertainty. Finally, they illustrate their proposed methods through examining the model performance on simulated data as well as actual mortality data from different countries. The findings of this study suggest that the QR method improves the prediction performance of the LC model and there exists evidence for trend changes in male mortality in the last century. More info

Physics

Few-Body Systems

Alexei Frolov, University of Western Ontario

Many of Dr. Frolov's research projects have lead him and his co-authors to solve the eigenvalue equations, out of necessity, $(\hat{A} - E \hat{B}) \vec{C} = 0$ with extremely ill-conditioned matrices $\hat{B}$ and $\hat{A}$. The ill-conditioning means that some of the eigenvalues of these two matrices are very close to zero, while other eigenvalues can be very large. There are many effective numerical methods which allow one to operate with such ill-conditioned matrices and some of these algorithms have been created by Frolov in collaboration with different co-authors, such as David H. Bailey--who has worked for the NASA Ames Research Center, received multiple awards, and is now the Chief Technologist of the Computational Research Department at the Lawrence Berkeley Laboratory. It is almost certain that this research group now has the most complete Library of Numerical Algorithms to work with extremely ill-conditioned matrices. More info

Physics

Sudbury Neutrino Observatory (SNO)

Clarence Virtue, Laurentian University

Neutrinos are thought to be vital to the understanding of the universe. It’s believed millions of neutrinos exist per cubic foot, filling the universe. So even with the smallest mass, they can induce gravitational effects in a huge way. Neutrinos are emitted by radiation sources. They’re known to exist in three types related to three different charged particles -- the electron-neutrino, muon-neutrino and tau-neutrino. Transformations called neutrino oscillation occur between them. Researchers Clarence Virtue and Melin Huang at the SNOLAB are studying a specific oscillation parameter, theta_13. The name comes from the fact that it is related to the transformation between the first generation electron-neutrino, to the third generation tau-neutrino. “Without theta_13, the phenomena employed to describe neutrino oscillation behavior is incomplete and could lead to misinterpretations in the particle physics world,” says Huang. New technology has made it possible to model and detect the three different types of neutrinos accurately, and how they oscillate. Theta_13 is the only missing parameter in the oscillation equation. It’s usually overlooked, being the smallest and hardest to assess. Currently, only a range exists for the parameter. Further study of this oscillation parameter could have profound effects on particle physics understanding. More info

Planetary Interior Dynamics

Dynamic Evolution of the Earth's Mantle

Julian Lowman, University of Toronto

Dr. Lowman's principal research area is the dynamic evolution of the Earth's mantle. Mantle convection is the primary means by which heat is removed from the Earth's mantle and core. It is the driving force responsible for plate tectonics, volcanism, earthquakes and orogenesis. It influences the Earth's topography and gravitational field, the geodynamo, the formation of mineral and hydrocarbon resources and the planet's climatic system, thus influencing cycles of environmental and biological evolution. Accordingly, the study of each of these disciplines reveals some expression of mantle convection and, conversely, evolving understanding of convection in the mantle affects conventional views of a diverse range of fields in the Earth Sciences. This specific research interest focuses on using numerical models to investigate the coupling between mantle convection and plate motion. In addition to an interest in the surface expression of mantle convection and its time-dependence, Dr. Lowman's research focuses on the viscosity profile of the mantle, the long-wavelength heterogeneity of the density of the mantle, the origin of hotspots, and influences on the Earth's heat loss (e.g., the supercontinent cycle, episodic plate reorganisation and the fate of old subduction zones). During his time as a faculty member at the Universities of Toronto and Leeds, Dr. Lowman has been and remains involved in several international collaborations focussing on developing parallelised computer codes for studying mantle convection in the Earth and terrestrial planets. More info

Quantum Computing

Master of the quantum dance

Raymond Laflamme, University of Waterloo

Raymond Laflamme choreographs the world’s most complicated dance: a pirouette of atoms. Using radio-wave pulses, he flips the nuclei of hydrogen and carbon-13 atoms inside a nuclear magnetic resonance machine, causing the nuclei to reverse their positive and negative poles. The fascination of this dance lies in its ability to represent the values of zero/one (off/on), like the switches in conventional computers. But there’s a difference. A binary circuit is open or closed; a bit, the smallest data unit, has a value of zero or one. But an atom’s positive pole can be up, down, or both; a quantum bit (qubit) can have both values at once. Laflamme, Director of the Institute for Quantum Computing, is a University of Waterloo physics professor and holds the Canada Research Chair in Quantum Information. As a scientist, he is celebrated worldwide. His goal: to build large, robust quantum devices. “But we still need to better control what we have,” Laflamme says. “The first step is to understand quantum properties; the second, to learn to control them; the third is to use them for something interesting.” More info

Quantum Gravity

Algebraic Topology and Quantum Gravity

Dan Christensen, University of Western Ontario

Dr. Christensen is examining the theory of quantum gravity from topological perspective. Topology is a branch of mathematics that examines certain properties of geometrical figures and, more broadly, the nature of space. For example, topology considers two objects equivalent, such as a coffee mug and a doughnut, if they can be continuously deformed into each other without cutting or adding any pieces. Although somewhat abstract and difficult to comprehend for non-mathematicians, topological methods have proven extremely successful in mathematics and physics. In his most recent paper, "Dual Computations of Non-Abelian Yang-Mills on the Lattice", Christensen and his co-investigators J. Wade Cherrington and Igor Khavkine use new mathematical techniques to perform calculations in Lattice Gauge Theory. Lattice Gauge Theory is a widely used method for studying three of the four fundamental forces that exist in the universe, the strong, weak and electromagnetic forces. The fourth force, gravity, has proven significantly difficult to unify with the other three forces. Using SHARCNET, Christensen says, puts him and his team in an excellent position to be the first to make these computations which will lead to new developments in the field. More info

Scientific Computing

Computer and Information Science

Hans De Sterck, University of Waterloo

Dr. De Sterck's group has made important contributions to algorithms that are used at the Center for Applied Scientific Computing of the US Lawrence Livermore National Laboratory and scale on the world's most powerful computers. They are currently extending these multilevel methods to the calculation of stationary vectors of large sparse Markov chains, with applications in computer network modeling and information retrieval (Google's PageRank). This research will lead to fast and scalable methods to solve very large systems, with applications in economically important areas like web searching and data mining for social networks. The second focus area is the development of parallel computational fluid dynamics simulation methods for simulation of solar system plasmas and space weather prediction. They plan to develop advanced solution-adaptive finite volume codes with acceleration on clusters with Graphics Processing Units (GPUs) and Cell processors. This project is collaboration with Natural Resources Canada and is funded by the Canadian Space Agency, and aims at developing tools for forecasting the detrimental effects that solar-initiated Space Weather events may have on the Earth environment. These effects include radiation exposure to astronauts and airline passengers, damage to satellite infrastructure, and the potential for catastrophic damage to power and telecommunications infrastructure. Canada and Ontario are leaders in observation of high-latitude Space Weather effects, and our project contributes to building out simulation capabilities for Ontario and Canada that can match the leading role in observations. More info

Scientific Computing

Human Computer Interaction

Ed Sykes, Sheridan College Institute of Technology and Advanced Learning

Dr. Sykes’ Human Computer Interaction (HCI) project involves inter-institutional research between Sheridan College and the University of Guelph. Dr. Sykes is currently working with Dr. Stacey and Dr. McCuaig (University of Guelph) on a research project involving Human Computer Interaction detecting interruption timings for computer-based tasks. The algorithm he is designing uses Bayesian Networks which are scalable and parallelizable through the use of MPI on large scale cluster supercomputers. He is also using SHARCNET resources for visualization of the interruptible moment which relies on large and complex datasets of the human-computer interactions. For research involving Human Computer Interaction and interruption detection, Dr. Sykes is working on establishing national and international collaborators where SHARCNET would play a major role in developing the algorithms and providing the resources for visualization analysis. This is also part of his role as SHARCNET's first College Research Chair.

Simulation and Modeling

Infectious Disease Modeling to Support Public Health

Yulia Gel, University of Waterloo

The main research interests of Dr. Gel concern time series analysis and spatio-temporal modeling, with applications to weather and climate forecasting, modeling dynamics of infectious diseases and analysis of fMRI data. Dr. Gel's work on infectious diseases is on creating highly reliable and comprehensive disease mapping to support public health and municipal disease mapping. She has just started a collaboration with the Infonaut and Peel Public Health Unit on developing software that enables one to model and forecast spatio-temporal spread of infectious diseases. More info

Soft Matter Physics

Unfolding the mystery of Mad Cow, Alzheimers and Parkinsons

Jeff Z. Y. Chen, University of Waterloo

Bovine spongiform encephalopathy or BSE. The term strikes fear into the hearts of farmers and economists alike. In 2003, a single case of BSE, more commonly known as Mad Cow disease, closed borders to Canadian beef exports internationally and resulted in more than billion in losses for Canada's cattle industry. A prion, an abnormal protein that changes its shape, or folds, when it is performing its biological functions, is believed to be the cause of BSE and a number of human diseases. For unknown reasons, prions sometime fold into the wrong shape. Misshapen prions have been attributed, in humans, to diseases like Alzheimer's and Parkinson's, but the causes of improper protein folding remain a mystery. Jeff Chen, a Professor of Physics at the University of Waterloo is using SHARCNET to investigate the folding properties of prions at the molecular level. Professor Chen uses SHARCNET's immense computing power to visualize the natural patterns of these complex protein systems and attempts to reconstruct the process by which prions mysteriously convert themselves from a normal form into a disease-rich structure. The research has both significant medical and economic implications. More info

Theoretical Neuroscience

Development of the neural engineering framework

Chris Eliasmith, University of Waterloo

The power of modern computers grows, as does our understanding of the human brain, but can a computer ever model the brain? Traditionally there are two approaches to modelling the brain – one that simulates human thought, and the other that models its low-level network structure. Recently, these two approaches have been unified by a new method called the Neural Engineering Framework (NEF). Chris Eliasmith is a professor in the Department of Systems Design Engineering and in the Department of Philosophy at the University of Waterloo. His NEF method is implemented in a software environment called Nengo (an amalgam of neural engineering objects). It simulates how neurons and neural systems interact to represent information, perform computations, and behave in perceptual, motor, and cognitive tasks. More info

Theoretical Physics

Tools for extremely accurate measurement

Gordon Drake, University of Windsor

Gordon Drake, Chair of the Department of Physics at the University of Windsor, is combining high performance computing on SHARCNET with high precision theory of fundamental atomic systems. According to his peers, Drake has set a new standard for high precision atomic theory with this work. His results provide tools for performing more accurate measurement of short-lived atomic nuclei that do not live long enough to be studied by other more traditional means. This research can be applied to technology, like geographic position systems (GPS), which require extremely precise measures of time and accuracy. More info

Chemistry and Biochemistry

Computer-assisted molecular design

Wely Floriano, Lakehead University

SHARCNET Research Chair, Dr. Floriano's research involves the development and application of computational tools to simulate biologically relevant systems. In the context of Medical Biotechnology, these computational techniques are applied for the development of new and modified chemical entities, such as new medicinal drugs and taste modifiers. In the context of Environmental Biotechnology and as part of the Biorefining Research Initiative at Lakehead University, computer-assisted molecular design methods are used to study genetic sequences of biomass-degrading microorganisms for selection and/or modifications aimed at improving efficiency and reducing waste of biorefining processes. These techniques are also used to identify economically relevant targets for chemical compounds produced as by-products of biorefining processes or isolated from forest resources, and to suggest chemical modifications to primary compounds to make them more valuable (e.g., a primary compound found to be a weak agonist of a pharmacologically relevant target may be further designed into a marketable medicinal drug). More info

Psychology

Conflict and Cognitive Control: a Role for the Anterior Cingulate Cortex in Reading

Serje Robidoux, University of Waterloo

Technological advances have revolutionized the field of visual word recognition. With the power to solve complex computational problems sitting on every desk, computer models of reading have become increasingly complex. A seminal paper by McClelland and Rumelhart (1981) has led to a strong tradition of computational modeling of word recognition. Indeed, their original Interactive Activation Model is a significant component of the currently dominant model of visual word recognition: the Dual-Route Cascaded model (Coltheart et al., 2001). Later work by McClelland & Rumelhart (1986) introduced principles of trainable parallel distributed processing (PDP) networks to the field. Today, the strengths and weaknesses of these two classes of models (Interactive Activation (IA) vs. PDP models) are hotly debated in the literature (Plaut et al., 1996; Plaut & Booth, 2000, 2006; Harm & Seidenberg, 2004; Borowsky & Besner, 2006). My proposed research will work towards resolving this debate. The ubiquity of powerful computers has a downside: reducing computational constraints vastly expands the range of testable models. I am a firm believer that modeling must be strongly informed by empirical and neurological research: it is empirical and neurological results that provide constraints on the range of possible models. I believe this approach will provide a more sophisticated understanding of the effects described here, leading to more effective computational models.

Humanities

Digital History

John Bonnett, Brock University

Dr. Bonnett's research in digital history touches on two areas: the appropriation of digital environments to support academic publishing; and the potential applicability of agent-based simulations to support historical analysis. He is currently working on a project titled HistorySpace, which is devoted to the development of tools and workflows to support the expression and documentation of historic research using virtual environments, and specifically environments that express combinations of text, sound, 2D, 3D and 4D objects, and treat topics in history where space is an important variable. SHARCNET will be an important support for this research because the datasets associated with the project will be quite large. Rendering and dissemination of high resolution reconstructions of historic environments, such as cities, will require significant computational power. The academic significance of this research centers on developing methods to express scholarly content in multi-media environments. While print-based repositories such as journals and books are the beneficiaries of a rich array of expressive and attestive practices and workflows, the same cannot be said for digital environments such as Google Earth. Until such methods are developed, scholars, particularly humanities scholars, will not appreciate the expressive and analytical potential that HPC and multi-media content afford. More info

Biological and Life Sciences

Integrative Biology

Cortland Griswold, University of Guelph

Dr. Griswold is developing gene genealogy models that require large amounts of memory and physiological models that require large numbers of processors. His research focuses on the development of models of multivariate trait evolution. He is particularly interested in determining how gene-genealogical structure and the history of mutation affects multivariate trait evolution within a population. Complementary to this genealogical work, he is developing models of multivariate traits from underlying biophysical models. The biophysical work is focused on multivariate traits that are determined, in part, by calcium signaling. More info

Environmental and Earth Science

Lake ice augurs climate change

Claude Duguay, University of Waterloo

Studies of receding glaciers and melting sea ice are suddenly a hot topic as earthlings brace for a warmer planet. For Claude Duguay, an interest in the cryosphere — Earth’s ice masses and snow deposits — predates current concerns, going back to the backyard ice rink of his childhood in Montreal. Duguay’s research focuses on lake ice as an indicator of climate variability and change. As well, he looks at how changes in lake ice conditions affect regional climate. Charting seasonal ice formation and breakup across the northern hemisphere, he has identified some startling trends. Between 1965 and 1995, the average date of ice breakup on lakes in western Canada was a full day earlier each successive year. His results were published in the Intergovernmental Panel on Climate Change Fourth Assessment Report of 2007. More info

Physics

Modeling the properties of nanoparticles

Ralf Meyer, Laurentian University

Dr. Ralf Meyer is a SHARCNET Research Chair in mathematical and computational materials science. Dr. Meyer is writing a computer program which can use parallel computers to carry out DFT calculations on larger sample sizes more efficiently, and at higher speeds than previously capable. By developing more timely and advanced computer simulation programs, this will allow scientists to closely analyze and understand the properties of infinitely small particles in many different materials. Tiny particles invisible to the naked eye may be the key to creating more effective medicines, and stronger engineering materials. More info

Biological and Life Sciences

Modelling the efficacy of drug combinations to treat influenza

Keith Poore, Ryerson University

Keith Poore is a 4th year undergraduate student in the Medical Physics program at Ryerson University. As part of his 4th year project, supervised by Dr. Catherine Beauchemin, Keith is modelling the efficacy of drug combinations in treating influenza infections. In the treatment of HIV, antiviral drug combinations are often used to slow down the development of resistance during therapy. Given the emergence of influenza strains highly resistant to the two main anti-influenza drugs, amantadine and oseltamivir, it is time a similar strategy be explored for influenza. In his work, Keith is using a Monte Carlo computer model capable of replicating the course of an influenza infection under treatment with amantadine and oseltamivir at different concentrations. This allows him to explore a wide range of dual drug combinations, something that would be too costly to do in a lab. The computer model is written in Octave and takes about 25 min to simulate 14 days of infection. Keith requires HPC for his project because the model has to be run over and over again, for a wide range of parameter values and drug concentrations to monitor how each combination affects every aspect of the infection. The work is therefore highly serial and CPU intensive making the Goblin cluster appropriate for this task. "I'm really excited to be using SHARCNET for my undergraduate thesis project. I got to learn about cluster computing and tools like ssh and bash and I think they're very useful skills to have. I'm also very happy I don't have to wait forever for all my simulations to finish. It's great!" More info

Physics

Simulations of advanced carbon based nano-materials

Eric Heritage, University of Ontario Institute of Technology

Eric, an NSERC Undergraduate Student Research Award (USRA) recipient, is working on extensive computer simulations of advanced carbon based nano-materials. In particular, he studies graphene functionalization by controlled adsorption of hydrogen. (Discovery of graphene in 2005 was awarded by 2010 Nobel Prize in Physics). He also studies clean and hydrogen modified diamond (111) surfaces, the main microcrystal orientation for nanodiamond particles. Eric models electronic, vibrational and optical properties of such modern materials to enable their non-destructive in-situ characterization for various promising applications. These include, for instance, microelectronics and renewable energy sources. The research involves international collaboration with research groups at the University of Rome Tor Vergata (Italy), Friedrich Schiller Universitat in Jena (Germany) and Centro de Investigaciones en Optica en Leon (Mexico). The main numerical tools are based on the first principles Density Functional Theory (DFT) Molecular Dynamics (MD) as implemented in the Quantum Espresso computer package, and used for calculation of the electron band structure, density of the electron states, optical transitions and vibrational spectra. The subsequent post-processing is done using software tools both SHARCNET based and custom developed (such as visualization, frequency analysis, etc.), which allows extraction from the microscopic MD output, the dynamics of the chemical bond formation, structural instability, temperature dependent optical properties, all of them to be compared to macroscopic experimentally measured properties and their interpretation. The above calculations rely heavily on parallel implementation of the computer code (using up to 128 CPUs per simulation), and SHARCNET systems such as saw, requin, narwhal etc., are perfectly suited for such type of research work. Eric's recent results were included in two oral talks on partially hydrogenated graphene and diamond (111) surface, and were presented by his supervisor A. Chkrebtii at the International Conference on Formation of the Semiconductor Interface (ICFSI13) in Prague, Czech Republic on July 3, 2011.

Humanities

The Orlando Project

Susan Brown, University of Guelph

Traditionally, high-performance computing has been used primarily in the academic disciplines of physics, astronomy, chemistry, and computer science - fields requiring complex simulations from very large datasets. Recently however, humanities researchers have also started using new computing technologies and, as a result, adding new and unexplored dimensions to the nature and scope of humanities work. Specifically, in 2006 women's literature scholars were presented with a new and exciting computing tool, a type of database capable of allowing researchers to investigate complex research questions. It was in 2006 that "Orlando: Women's Writing in the British Isles from the Beginnings to the Present" debuted and changed the relationship between computing and the humanities. Orlando is a unique "textbase" containing articles on the history, literature, and culture relevant on women writers in Britain. The semantic encoding of its content, allows users to track such matters as the reception of an author's writings, related intertexuality and influence, literary responses, textual features, and other aspects of the author's literary career and life, not just in the entries directly on her, but throughout the textbase. Susan Brown, a co-founder of Orlando, is now collaborating with SHARCNET to develop software that will generate three-dimensional information networks based around search results. Brown and SHARCNET will push the Orlando's capabilities even further by enhancing the standard search-and-retrieval text browser with a visual modelling tool for illustrating relationships, which scholars will use to explore new and unexpected paths and patterns. More info