RAIN schedule for Winter Quarter 2013-14
RAIN will be held on Wednesdays from 12:00-1:00pm in Y2E2 101. And yes, lunch will be provided!
|January 29||Derek Ruths||Control Profiles of Complex Networks|
|February 5||Amin Karbasi||From Small-World Networks to User-Driven Content Search|
|February 19||Arun Chandrasekhar||Savings Monitors|
|February 26||Ariel Procaccia||Computational Fair Division: From Cake Cutting to Cloud Computing|
|March 12||Niki Kittur||Big Thinking: Augmenting human cognition with crowds and computation|
|March 19||Sinan Aral||TBA|
Google Calendar for RAIN
Previous year's talks
Archived talks can be accessed here.
Talk AbstractsControl Profiles of Complex Networks
Derek Ruths, McGill University
Studying the control properties of complex networks provides insight into how designers and engineers can influence these systems to achieve a desired behavior. Topology of a network has been shown to strongly correlate with certain control properties; in this talk I will discuss the fundamental structures that explain the basis of this correlation. I'll then use these fundamental structures to construct a new statistic, the control profile, that quantifies the different proportions of control-inducing structures present in a network. Using the control profile, I'll show that standard random network models do not reproduce the kinds of control structures that are observed in real world networks. Furthermore, the profiles of real networks form three well-defined clusters that provide insight into the high-level organization and function of complex systems.
Bio: Derek Ruths is an assistant professor of Computer Science at McGill University. He joined the faculty in 2009 after completing his PhD in Computer Science at Rice University. A major research direction in his group considers the problem of characterizing and predicting the large-scale dynamics of human behavior in online social platforms. His ongoing work in this area includes quantitatively modeling how communities change over time, measuring and predicting group demographics from unstructured user-generated content, and computational methods for assessing discussion topics within a collection of users. His work has been published in top-tier journals and conferences including Science, EMNLP, ICWSM, and PLoS Computational Biology. His research is currently funded by a wide array of organizations including NSERC, SSHRC, tech companies, and the US National Science Foundation - underscoring the broad, interdisciplinary nature of his work.
From Small-World Networks to User-Driven Content Search
Amin Karbasi, ETH Zurich
In this talk, we will describe a new way of navigating through a database of similar objects using comparisons. In short, at each step, the database presents two objects to the user, who then selects among the pair the object closest to the target that she has in mind. This process continues until, based on the user’s answers, the database can uniquely identify the target she has in mind. This kind of interactive navigation (also known as exploratory search) and its variants have numerous real-life applications in content-based image/multimedia retrieval.
We will argue that the above problem is strongly related to the small-world network design: given a graph embedded in a metric space, how should one augment this graph by adding edges in order to minimize the expected cost of greedy forwarding. We will first show that the small-world network design problem is NP-hard. Given this negative result, we propose a novel mechanism for small-world design and provide an upper bound on its performance. This mechanism has a natural equivalent in the context of content search through comparisons, and we establish both (order optimum) upper and lower bounds for its performance. These bounds are intuitively appealing as they provide performance guarantees in terms of the bias of the distribution of targets, captured by the entropy, as well as the topology of their embedding, captured by the doubling dimension.
Bio: Amin Karbasi is currently a post-doctoral scholar in the Learning and Adaptive Systems Group at ETHZ, hosted by Prof. Krause. From 2008 until 2012, he was a Ph.D. student in the school of computer and communication sciences at Ecole Polytechnique Federale de Lausanne (EPFL). During his Ph.D. studies, he was working on the scaling laws and applications of Graph-Based Information Processing under the supervision of Prof. Urbanke and Prof. Vetterli. He received his M.Sc. in computer and communication sciences in 2007, and his B.Sc. in electrical engineering in 2004, both from EPFL.
As of July 2014, he will be joining Yale school of engineering and applied science where he leads the Inference, Information and Decision Systems Group.
Arun G. Chandrasekhar, Stanford University
We conduct a field experiment in India to explore two interventions to help individuals to increase their savings balances. First, we design a financial product based on the popular business correspondent model, which includes frequent reminders, assistance in account opening, and the setting of a six-month savings goal. Second, we measure the effectiveness of adding a peer monitoring component to this basic bundle and test whether the local social network can help to increase the penetration of the formal banking system. We ask whether having a monitor substitutes for a formal commitment device, whether individuals choose the most effective monitors, and moreover, whether some community members are better than others at encouraging financial capability.
Bio: Arun Chandrasekhar.
Computational Fair Division: From Cake Cutting to Cloud Computing
Ariel Procaccia, Carnegie Mellon University
I will present an exciting new interaction between computer science and fair division theory. On the one hand, I will show how computational thinking provides a novel perspective on classic problems in fair division, such as cake cutting and estate division. On the other hand, I will explain how fair division theory is informing the allocation of computational resources. I will also talk about the integration of some of these theoretical ideas into a not-for-profit fair division website that aims to make the world just a bit fairer.
Bio: Ariel Procaccia is an assistant professor in the Computer Science Department at Carnegie Mellon University. He received his Ph.D. in computer science from the Hebrew University of Jerusalem, and was subsequently a postdoc at Microsoft and Harvard. Procaccia is a recipient of the NSF CAREER Award (2014), the (inaugural) Yahoo! Academic Career Enhancement Award (2011), the Victor Lesser Distinguished Dissertation Award (2009), and the Rothschild postdoctoral fellowship (2009). He was named in 2013 by IEEE Intelligent Systems to their biennial list of AI's 10 to Watch. He is currently the editor of ACM SIGecom Exchanges, an associate editor of the Journal of AI Research (JAIR) and Autonomous Agents and Multi-Agent Systems (JAAMAS), and an editor of the upcoming Handbook of Computational Social Choice.
Big Thinking: Augmenting human cognition with crowds and computation
Niki Kittur, Carnegie Mellon University
Despite the growing availability of information online, individual human cognition is fundamentally limited in the speed and amount of information it can process at once. In this talk I will discuss ways of augmenting individual cognition by taking advantage of crowds and computation. Our approaches combine the flexibility of many human minds working together with the raw power of computational systems to accelerate learning, innovation, knowledge production, and scientific discovery. I will discuss several novel crowdsourcing, machine learning and visualization systems that augment the uniquely human abilities of analogy, schema induction, and problem solving. Examples include augmenting human thinking in product innovation (e.g., Quirky); knowledge production (Wikipedia); creative tasks such as writing, journalism and poetry; and making sense of multivariate and graph data.
Bio: Aniket Kittur is an Assistant Professor and holds the Cooper-Siegel Chair in the Human-Computer Interaction Institute at Carnegie Mellon University. His research on crowd-augmented cognition looks at how we can augment the human intellect using crowds and computation. He has authored and co-authored more than 70 peer-reviewed articles, 10 of which have received best paper awards or honorable mentions. Dr. Kittur has received an NSF CAREER award, major research awards from NSF, NIH, Google, Microsoft, and Bosch, and his work has been reported in venues including Nature News, The Economist, The Wall Street Journal, NPR, Slashdot, and the Chronicle of Higher Education. He received a BA in Psychology and Computer Science at Princeton, and a PhD in Cognitive Psychology from UCLA.
To be announced
Sinan Aral, MIT
Bio: Sinan Aral.