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

Colloquium Home
Fall 2014 Schedule
Winter 2015 Schedule
How to Suggest a Speaker
Archives

 

Date
( Winter 2005 )
Speaker and Abstract
2005/01/14 Speaker: John Buchanan
Affiliation: Electronic Arts
Title: Computer Games
Abstract: Video games are now big business. In this talk I overview the state of the industry in terms of our cultural relevance and some business aspects. I introduce EA, briefly overviewing our history and how that affects our internal practices. Stories and experiences are important to us. I talk about stories and experiences in the context of videogames and how we are pushing the barriers to make better games. Attendees of this talk will walk away with an understanding of the size and dynamics of the industry, they will understand EA's position within the industry, and will have an understanding of some of the internal workings of EA. In addition to this they will have lots of questions that are important to think about when evaluating the quality of a game and/or designing one.
2005/01/21 Speaker: Ian Foster
Affiliation: University of Chicago
Area: Grid Computing
Title: The Grid: Reality, Technology, and Applications
Abstract: The Grid seems to be everywhere, with announcements of sales from major computer vendors, deployment in a wide range of application spaces, and many national and international scale infrastructure deployment. However, in spite of the popularity of the term, there is often confusion as to what the Grid is and what problems it solves. Is there any "there there" or is it all just marketing hype? In this talk, I will address these questions, describing what the Grid is, what problems it solves, and what technology has been developed to build Grid infrastructure and create Grid applications. I will review the current status of Grid infrastructure and deployment and give examples of where Grid technology is being used not only to perform current tasks better, but to provide fundamentally new types of capabilities that are not possible otherwise.
2005/02/11 Speaker: Scott Aaronson
Affiliation: Princeton Institute of Advanced Studies
Title: The Postselection Principle
Abstract: I'll propose a principle that unifies many results about classical and quantum computation, and also leads to many new results. The principle is that if an object belongs to a space of low dimension, and if we can repeatedly interact with or simulate the object with reliable results, then it is possible to duplicate the object by guessing a random object, and then postselecting on the new object behaving the same as the old one on a small number of carefully chosen test cases. I'll show how this principle sharply limits the power of quantum communication, quantum circuits, quantum proofs, and quantum advice. I'll also market the principle as an "intellectual export" from theoretical computer science to other areas of knowledge.
2005/02/18 Speaker: Stefan Brands
Affiliation: McGill University
Area: Cryptography
Title: Federated Identity Management & Cryptography
Abstract: Most individuals are registered in many hundreds if not thousands of databases scattered across disparate systems. Major industry initiatives in so-called "federated identity management" aim to enable organizations to seamlessly share "identity assertions" across organizational boundaries. I will start today's presentation with an overview of the severe security and privacy problems that these federated identity management initiatives cause for individuals and organizations. Elimination of these problems requires that each individual be made the middle man for the sharing of identity assertions about him or her. To prevent individuals from pooling, cloning, lending, discarding, reusing, or otherwise manipulating organization-made identity assertions about them, organization-made identity assertions must be adequately protected. In the remainder of the presentation, I will give an overview of twenty years of academic research in cryptography that has concerned itself with exactly this objective. To conclude the presentation, I will outline the use of the resulting user-centric identity management architecture in two example contexts: electronic health record management and government online.
2005/03/04 Speaker: Carsten Schürmann
Affiliation: Yale University
Area: Logic (Frameworks and Programming Languages), Theorem Proving
Title: Proofs 'R' Us
Abstract: The Logosphere project is designing and implementing a digital library of formal proof, bringing together different proof assistants and theorem provers, with the goal to facilitate the exchange of mathematical knowledge. Once completed, Logosphere enables industries that employ formal methods in various dialects, such as INTEL, AMD, and NASA to draw on formerly inaccessible mathematical facts and their meaning. Often underestimated, formal proofs are invaluable and essential when it comes to assign meaning to mathematical theorems. In my talk I will introduce the practical and theoretical foundations underlying the Logosphere digital library and discuss an implementation of an executable transformer between two popular proof development tools, HOL and Nuprl (joint work with Mark-Oliver Stehr).
2005/03/11 Speaker: Maja D'Hondt
Affiliation: Vrije Universiteit Brussel
Title: Rule-Based Knowledge in Object-Oriented Software
Abstract: Software applications often contain implicit knowledge for making decisions, giving recommendations, guiding processes and specifying policies or preferences. This knowledge can be represented as rules, or business rules, generally having the form "if condition then conclusion", or a variation thereof. Software applications containing rule-based knowledge, however, generally also consist of functionality that is most suitably represented in the object-oriented paradigm. As such, many recent approaches exist for developing object-oriented software applications with explicit and separated rule-based knowledge. At the implementation level, these approaches vary widely in the way they represent rules, going from ad-hoc representation of rules to full-fledged rule-based languages. Additionally, they provide mechanisms for integrating rules with object-oriented functionality in order to obtain one, fully functional application. In this talk, I introduce the state of the art in this domain and identify the shortcomings of existing approaches, more specifically the tight coupling of rules and object-oriented functionality. I show how Aspect-Oriented Programming - although a seemingly unrelated domain - can provide a solution.
2005/03/18 Speaker: Tim Merrett
Affiliation: McGill University
Title: CS++: Re-inventing Computer Science (for Secondary Storage)
Abstract: Secondary storage ("SS") offers a significantly different memory organization from RAM, which most of computer science is geared to support. This forces a revision of data structures and algorithms on one hand and of programming language on the other. Algorithms become simpler, but in the frequent situation that the data structure sizes exceed RAM capacity, even polynomial complexity becomes intractable unless it is sub-quadratic. Languages are obliged to abstract over looping, which presents opportunities for programming at a much higher level than allowed by the "von Neumann bottleneck" that restricts most languages geared for RAM. This talk will describe the challenges of developing a language which operates at the level demanded by SS. We will look at some benefits of abstracting over looping: enormous reductions in code for building large systems, intrinsic parallelization, and easy incorporation of the Internet. We will note a diversity of SS applications, including geographical information systems and expert systems as well as conventional organizational management. We will see that Computer Science SS also has benefits for ordinary RAM programming, and will touch on some data structure ideas as illustration.
2005/04/08 Speaker: Ulfar Erlingsson
Affiliation: Microsoft Research
Title: Software Security and Control-Flow Integrity
Abstract: Current software attacks often build on exploits that subvert machine-code execution. The enforcement of a basic safety property, Control-Flow Integrity (CFI), can prevent such attacks from arbitrarily controlling program behavior. CFI enforcement is simple, and its guarantees can be established formally, even with respect to powerful adversaries. Moreover, CFI enforcement is practical: it is compatible with existing software and can be efficiently implemented. Finally, CFI guarantees are a useful foundation for analysis, processing, and the establishment of further properties of the software. CFI derives its security benefits from constraining low-level machine operations to conform more closely to the high-level programming language semantics. CFI generalizes, or can subsume, popular mechanisms for reducing this discrepancy between the hardware and software semantics.