Designing and Building Technology to Empower People

September 21, 2010

Bio:
Following many years of research in theoretical computer science, Dr. Ladner works on accessibility technology research, especially technology for deaf, deaf-blind, hard-of-hearing, and blind people.

Analytical Modeling of Data Centers to Optimize Performance and Power

October 26, 2010

Bio:
Prof. Harchol-Balter's work focuses on designing new resource allocation policies (load balancing policies, power management policies, and scheduling policies) for server farms and distributed systems, spanning both queueing analysis and systems implementation.

Meeting Everyone's Need for Computing

Tuesday, November 23, 2010

Abstract:
While interest in computer science degrees has declined, interest in computer science continues to grow across campus. Some estimates suggest that by 2012 there will be some 13 million end-user programmers in the United States, compared to an estimated 3 million professional software developers. In this talk, I argue for more attention to that much greater number, for having an impact by making more successful the non-professional who uses computer science. I will present historical evidence that our field has had a goal of teaching the non-professionals about computer science for over 40 years, and recent evidence that end-user programmers want what we have to offer, and that we need to develop new kinds of classes and new kinds of approaches to teaching CS to meet their needs. I will present methods for teaching computing that have improved success rates for non-computing majors (while still including programming), such as contextualized computing education.

Bio:
Dr. Guzdial received his Ph.D. in Education and Computer Science from the University of Michigan in 1993. His research focuses on learning sciences and technology, specifically, computing education research. He has published several books on the use of media as a context for learning computing. He was the original developer of the "Swiki" — the first wiki designed for educational use. He serves on the Education Board of ACM and on the Board of the Special Interest Group on CS Education (SIGCSE). He is on editorial boards of the Journal of the Learning Sciences, ACM Transactions on Computing Education, and Communications of the ACM.

Social Networks Offline

Tuesday, January 18, 2011

Abstract:
Humans are embedded in social networks that affect every aspect of our lives. Work in the Christakis lab involves the application of network science and statistical and mathematical models to a variety of observational and experimental datasets in order to understand the structure and function of human networks. What social, biological, and mathematical principles help determine how and why human social networks form and how they operate? One stream of work focuses on the spreading dynamics of health-related phenomena (obesity, smoking, emotions, altruism) in longitudinally evolving networks ("contagion"). Another stream of work examines the genetic, social, and psychological processes that determine social network structure ("connection"). These investigations have meaningful implications for public policy and public health.

Christakis has spent the last ten years examining how and why humans assemble themselves into small and large social networks, and how our embeddedness in such networks affects our lives. Using data from diverse sources — including longitudinally followed cohorts, online interactions, and experiments — he has explored questions as diverse as the inter-personal spread of obesity and emotions, the use of humans within networks as 'sensors,' the 'pay-it-forward' property of human altruism and the genetic basis for social network structure.

Bio:
Dr. Christakis received his B.S. from Yale University in 1984, M.D. from Harvard Medical School and M.P.H. from the Harvard School of Public Health in 1989, and Ph.D. from the University of Pennsylvania in 1995. He is an internist and social scientist who conducts research on social factors (such as small and large social networks) that affect health, health care, and longevity. He is the co-author of Connected: The Surprising Power of Our Social Networks and How They Shape Our Lives. He is 2006 Fellow of the Institute of Medicine of the National Academy of Sciences, and is on the 2009 Time magazine’s list of the 100 most influential people in the world.

Computational Models of Common-Sense Theories: What People Know About the World, and How They Know It

Tuesday, February 15, 2011

Abstract:
How do humans come to know so much about the world from so little data? Even young children can infer the meanings of words, the hidden properties of objects, or the existence of causal relations from just one or a few relevant observations — far outstripping the capabilities of conventional learning machines. How do they do it — and how can we bring machines closer to these human-like learning abilities? I will argue that people's everyday inductive leaps can be understood in terms of (approximations to) probabilistic inference over generative models of the world. These models can have rich latent structure based on abstract knowledge representations, what cognitive psychologists have sometimes called "intuitive theories", "mental models", or "schemas". They also typically have a hierarchical structure supporting inference at multiple levels, or "learning to learn", where abstract knowledge may itself be learned from experience at the same time as it guides more specific generalizations from sparse data.

This talk will focus on models of learning and "learning to learn" about categories, word meanings and causal relations. I will show in each of these settings how human learners can balance the need for strongly constraining inductive biases — necessary for rapid generalization — with the flexibility to adapt to the structure of new environments, learning new inductive biases for which our minds could not have been pre-programmed. I will also discuss how this approach extends to common sense reasoning with richer forms of knowledge, such as intuitive psychology and social inferences, and intuitive physics. Time permitting; I will raise some challenges for the project of understanding how learning works in the brain and neural network models.

Bio:
Dr. Tenenbaum received his Ph.D. from MIT in 1999, and was a member of the Stanford University faculty in Psychology and in Computer Science in 1999–2002. He studies learning and inference in humans and machines, with the twin goals of understanding human intelligence in computational terms and bringing computers closer to human capacities. He is the recipient of the 2005 New Investigator Award from the Society for Mathematical Psychology, the 2007 Early Investigator Award from the Society of Experimental Psychologists, and the 2008 Distinguished Scientific Award from the American Psychological Association.