Skip to content. Skip to navigation
McGill Home SOCS Home
Personal tools
You are here: Home Announcements and Events Colloquium Archive

Colloquium Home
Fall 2013 Schedule
Winter 2014 Schedule
How to Suggest a Speaker
Archives

 

Date
( Fall 2010 )
Speaker and Abstract
2010/09/03 Speaker: SOCS Professors
Affiliation: McGill University
Area: Computer Science
Title: SOCS Research Overview
Abstract: This first colloquium of the semester aims at giving a overview of the research being done within the various groups at SOCS. Several professors are going to present short 5 minute talks about their current research. The attendance to the colloquium is mandatory for new M.Sc. students, but all students and profs are invited to hear about what's happening in our department. Biography of Speaker:


2010/10/01 Speaker: James O'Brien
Affiliation: University of California at Berkeley
Area: Computer Animation
Title: Mesh Modification and Real-Time FEM Simulation
Abstract: This talk will discuss a pair of simulation systems that been developed to model complex deformable behaviors in real-time, interactive contexts. Based around fast finite element methods, the systems are designed to address two specific applications: destructible environments in Star Wars: The Force Unleashed, and interactive modeling of prostate brachytherapy. Although dynamic remeshing is often dismissed as impractically slow, in both cases it was key part to making the algorithms work effectively in a real-time setting. The results presented will include captured footage from the live game and comparisons of simulated needle insertion to footage with gel tissue phantoms. Biography of Speaker:

James F. O'Brien is an Associate Professor of Computer Science at the University of California, Berkeley. His primary area of interest is Computer Animation, with an emphasis on generating realistic motion using physically based simulation and motion capture techniques. He has authored numerous papers on these topics. In addition to his research pursuits, Prof. O'Brien has worked with several game companies on integrating advanced simulation physics into game engines, and his methods for destruction modeling were recently used in the film Avatar. He received his doctorate from the Georgia Institute of Technology in 2000, the same year he joined the Faculty at U.C. Berkeley. Professor O'Brien is a Sloan Fellow and ACM Distinguished Scientist, Technology Review selected him as one of their TR-100 for 2004, and he was recently awarded research grants from the Okawa and Hellman Foundations. He is currently serving as ACM SIGGRAPH Director at Large.


2010/10/08 Speaker: Wolfgang Heidrich
Affiliation: University of British Columbia
Area: Computer Graphics
Title: Visible Light Tomography in Computer Graphics
Abstract: Tomographic methods are the standard approach for obtaining volumetric measurements in medicine, science, and engineering. Typical tomography setups acquire 2D X-ray images of an object, and reconstruct a 3D voxel representation from this data. Unfortunately, for many applications in computer graphics, such X-ray setups are not feasible due to cost and/or safety concerns. In this presentation, I will introduce our recent work on visible light tomography, which has much more modest hardware requirements. I will discuss tomographic methods in the presence of refraction, and show applications to the scanning of transparent objects, and the capture of gas flows. I will also discuss first results on a new solver for generic tomography problems. Biography of Speaker:

Professor Wolfgang Heidrich holds the Dolby Research Chair in Computer Science at the University of British Columbia. He received a PhD in Computer Science from the University of Erlangen in 1999, and then worked as a Research Associate in the Computer Graphics Group of the Max-Planck-Institute for Computer Science in Saarbrucken, Germany, before joining UBC in 2000. Heidrich's research interests lie at the intersection of computer graphics, computer vision, imaging, and optics. In particular, he has worked on High Dynamic Range imaging and display, image-based modeling, measuring, and rendering, geometry acquisition, GPU-based rendering, and global illumination. Heidrich has written over 100 refereed publications on these subjects and has served on numerous program committees. He was the program co-chair for Graphics Hardware 2002, Graphics Interface 2004, and the Eurographics Symposium on Rendering, 2006.


2010/10/15 Speaker: Joelle Pineau
Affiliation: SOCS McGill
Area: AI
Title: Automatic detection and suppression of epileptiform activity
Abstract: Electrical stimulation has recently emerged as a promising therapy for patients with medically intractable epilepsy. However little is known about the best stimulation patterns to use, such that we get maximal seizure reduction, while also minimizing long-term damage to the brain. The overall goal of this project is to automatically optimize a closed-loop strategy for the control of deep brain stimulation using reinforcement learning methods. In this talk I will present some of the key computational challenges that arise, including: (1) the use of machine learning methods to automatically detect seizures from EEG recordings, (2) methods for building a dynamical model of epilepsy to provide synthetic training data for the reinforcement learning agent, (3) initial results of applying reinforcement learning using an in vitro model of epilepsy. Biography of Speaker:

Joelle Pineau is associate professor of Computer Science at McGill University, where she heads the Reasoning and Learning Laboratory. She received her PhD (2004) and MSc (2000) in Robotics from Carnegie Mellon University, and her B.A.Sc. in Engineering from the University of Waterloo (1998). Her research focuses on developing new methodologies in artificial intelligence, statistical machine learning, and robotics. She is also actively exploring the application of these ideas to optimizing dynamic treatment regimes for chronic illnesses.


2010/10/22 Speaker: Matthew Dwyer
Affiliation: University of Nebraska
Area: Software Engineering
Title: Exploiting Partial Success in Applying Automated Formal Methods
Abstract: The past decades have produced a wide-range of formal techniques for developing and assessing the correctness of software systems. In practice, when applied to large modern software systems all existing automated formal methods come up short. They might produce false error reports, exhaust available human or computational resources, or be incapable of reasoning about some set of important properties. Whatever their shortcoming, the goal of proving a system correct remains elusive. Nevertheless, the conventional wisdom that software systems are mostly correct appears to be true - systems have much more correct behavior than incorrect behavior. Given this, we explore what it means to shift from focusing on proving correctness, to developing automated formal methods that produce evidence that a subset of system behaviors are consistent with system specifications. We describe the challenges in producing a rich suite of evidence-producing formal methods, where the weakness of one method is masked by the strength of another, and where evidence can be combined to produce an explicit characterization of system correctness. Biography of Speaker:

Matthew B. Dwyer is the Henson Professor of Software Engineering in the Department of Computer Science and Engineering at the University of Nebraska - Lincoln. He received the BS in Electrical Engineering in 1985 from the University of Rochester, worked for six years as a Senior Engineer with Intermetrics Inc. developing compilers and software for safety-critical embedded systems, then completed his PhD in Computer Science at the University of Massachusetts at Amherst in 1995. Dr. Dwyer is an active member of the software engineering, computer-aided verification and program analysis research communities and has published widely in those areas. He has chaired the top meetings in the field of software engineering, including FSE in 2004 and ICSE in 2008. He was named an ACM Distinguished Scientist in 2007 and has received the ICSE Most Influential Paper award and the SIGSOFT Impact Paper award, both in 2010.


2010/11/12 Speaker: Eitan Grinspun
Affiliation: Columbia University
Area: Computer Animation
Title: From Sorcery to Science: how Hollywood Physics impacts the Sciences
Abstract: Cinema uses computers to animate physics. Special effects such as explosions and lifelike depictions of imaginary characters are made possible by mathematical and computational models that capture qualitative, characteristic behavior of a mechanical system. This is scientific computing with a twist. I will describe the process by which we derive and compute models of physics, and show actual examples of resulting technologies in film, consumer products, physics, and medicine. Our research group develops scientific computing tools by focusing on the underlying geometry of the mechanical system. I will describe a process in which we build a discrete picture from the ground up, mimicking the axioms, structures, and symmetries of the smooth setting. I will survey the problems we address using this methodology, such as computing the motion of flexible surfaces, cloth, hair, honey, and solids experiencing mechanical contact. Industry and academia has adopted these methods to improve products such as Adobe Photoshop, films such as Disney's Tangled (whose release date coincides with this talk), train surgeons, and understand nonlinear soft-matter phenomena. Biography of Speaker:

Eitan Grinspun is Associate Professor of Computer Science at Columbia University in the City of New York. He was Professeur d'Université Invité at l'Université Pierre et Marie Curie in 2009, a Research Scientist at the Courant Institute of Mathematical Sciences from 2003-2004, and a graduate student at the California Institute of Technology from 1997-2003. He was an NVIDIA Fellow in 2001, an Everhart Distinguished Lecturer in 2003, an NSF CAREER Award recipient in 2007, and is currently an Alfred P. Sloan Research Fellow.


2010/11/19 Speaker: Alan Mislove
Affiliation: Northeastern University, USA
Area: AI
Title: Leveraging social networks in systems design
Abstract: Recently, online social networking sites have exploded in popularity; numerous sites are dedicated to exchanging end-user generated content, such as photos on Flickr, videos on YouTube, and status updates on Twitter. These sites represent a new type of information system, where links between users (as opposed to links between content items) are the primary manner in which information is structured. But, the enabling of all users to become publishers is resulting in a deluge of data that only makes the fundamental problems of finding relevant content and trustworthy users even harder. The social networks that these sites are based on, however, offer a potential solution, because the relationships among users in the social network may indicate the relevance of content or the trustworthiness of other users in the network. In this talk, I discuss two systems that leverage social networks to solve some of the challenges facing information systems. First, I present WebCloud, a system that addresses the challenge of content distribution in online social networks by recruiting individual users to help serve their own content. WebCloud represents the first step towards decentralized content exchange over centralized social networks. Second, I describe Bazaar, a system that strengthens user reputations in online marketplaces. Bazaar addresses the reputation manipulation that is possible today because accounts on online marketplaces are usually free to obtain. Both of these systems have been evaluated on data from real online social networks or in a deployment with real users. Biography of Speaker:

Alan Mislove is an Assistant Professor in the College of Computer and Information Science at Northeastern University. He holds a Ph.D. in computer science from Rice University. His primary research interests are online social networks, distributed systems, and Internet measurement, and his recent work includes studies of structural properties, information propagation, and the privacy implications of social networks, as well as the design of multiple systems that leverage social networks in novel ways. The social network data that he collected, anonymized, and released is now in-use by over 500 researchers worldwide. His work has been widely covered in the press, including New Scientist, the Associated Press, Reuters, the New York Times, and the CBS Evening News.


2010/12/03 Speaker: Sebastian Will
Affiliation: Massachusetts Institute of Technology
Area: Bioinformatics
Title: LocARNA-P: Alignment reliabilities for RNA structural analysis
Abstract: RNA plays a crucial role in living cells going far beyond coding for proteins. Elucidating the functional roles of non-coding RNAs (ncRNAs) has thus become a central research interest in molecular biology. Current genomic screens for ncRNAs predict a large number of genomic regions containing potential structural ncRNAs. The analysis of this data requires highly accurate prediction of ncRNA boundaries and discrimination of promising candidate ncRNAs from weak predictions. In this talk, I present the novel approach LocARNA-P, which overcomes the main limitations of existing approaches for predicting boundaries and discriminating ncRNA candidates. The approach is based on efficiently computing reliabilities of sequence-structure alignments. In general, reliability profiles of alignments provide a versatile tool for the manual and automatic analysis of ncRNAs. The ready-to-use software tool LocARNA-P is freely available at http://www.bioinf.uni-freiburg.de/Software/LocARNA/ and produces high-quality multiple sequence-structure alignments together with associated reliability information and prediction of accurate boundaries. This is joint work with Tejal Joshi, Ivo Hofacker, Peter Stadler, and Rolf Backofen. Biography of Speaker:

Sebastian Will is Post-Doc in Bonnie Berger's Computation and Biology Lab at the Massachusetts Institute of Technology (MIT), receiving a Research Scholarship from the German Research Foundation. He received his doctorate in Bioinformatics/Computer Science from the University of Jena and then worked as post-doctoral researcher at the University of Freiburg. His research interests comprise bioinformatics, constraint programming, and structural analysis and kinetics of RNA and proteins. He published numerous papers in these areas, repeatedly co-chaired workshops on constraints in bioinformatics (WCB series), and developed the widely used RNA analysis tool LocARNA.