Select a calendar:
Filter March Events by Event Type:
SUNMONTUEWEDTHUFRISAT
Conferences, Lectures, & Seminars
Events for March
-
CS Colloquium: Manu Sridharan (Samsung Research) - Analysis Tools for Reliable Software Everywhere
Tue, Mar 01, 2016 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Manu Sridharan, Samsung Research America
Talk Title: Analysis Tools for Reliable Software Everywhere
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium
Software is becoming ubiquitous in everyday life, from today's
smartphones and servers to tomorrow's self-driving cars, drones, and Internet of Things devices. However, the distributed, always-on nature of this software poses significant new challenges for reliability, security, and programmer productivity. Better programming tools are needed to enable next-generation applications to achieve their full transformative potential. I have helped design and develop several such tools in my recent research based on novel techniques in program analysis.
This talk will focus on EventRacer, the first tool for discovering and debugging non-determinism errors in event-driven programs. Event-driven programming has recently achieved a meteoric rise in popularity, as it is well-suited to the needs of modern interactive, client-server applications. However, event-driven programs often suffer from timing-based data races that can be fiendishly difficult to reproduce and debug. EventRacer adapts the notion of a "happens-before relation" from concurrent and distributed systems to give a clean definition of data races for event-driven programs. It also incorporates multiple novel techniques to achieve scalability and usability for real-world applications. With EventRacer, we found many errors in deployed Fortune 100 web sites, and its techniques have since been applied in a variety of other emerging domains.
Biography: Manu Sridharan is a senior researcher at Samsung Research America in the area of programming languages and software engineering. He received his PhD from the University of California, Berkeley in 2007, and he worked as a research staff member at IBM Research from 2008-2013. His research has drawn on, and contributed to, techniques in static analysis, dynamic analysis, and program synthesis, with applications to security, software quality, code refactoring, and software performance. His work has been incorporated into multiple commercial products, including IBM's commercial security analysis tool and Samsung's developer toolkit for the Tizen operating system. For further details, see http://manu.sridharan.net.
Host: CS Department
Location: Olin Hall of Engineering (OHE) - 136
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Iolanda Leite (Disney Research) - Long-term Human-Robot Interaction in the Real-World
Tue, Mar 01, 2016 @ 04:00 PM - 05:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Iolanda Leite, Disney Research
Talk Title: Long-term Human-Robot Interaction in the Real-World
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium
Most social robots and virtual characters are still unable to keep users engaged over repeated interactions because they lack social and adaptive capabilities that facilitate the interaction once the novelty effect fades away. In this seminar, I will present my past and current research on mechanisms that allow autonomous robots to be deployed in real-world social environments over weeks and months. These mechanisms include computational models of empathy, turn-taking and engagement. I will present evidence on the positive effects of implementing these models in robots and virtual characters interacting with people in several application domains, and discuss limitations of the current state of the art in robotic technology suitable for realistic social environments. An improved understanding of how robots should perceive and act depending on their surrounding social context can lead to more natural, enjoyable and useful long-term human-robot interactions.
The meeting will be available to stream HERE. Please open in new tab for best results.
Biography: Iolanda Leite is an Associate Research Scientist at Disney Research, Pittsburgh. She received her Ph.D in Information Systems and Computer Engineering from Instituto Superior Técnico, University of Lisbon, in 2013. From 2013 to 2015, she was a Postdoctoral Associate at the Yale Social Robotics Lab. Her doctoral dissertation, "Long-term Interactions with Empathic Social Robots", received an honorable mention in the IFAAMAS-13 Victor Lesser Distinguished Dissertation Award. Iolanda has published over 40 conference and journal in the areas of human-robot interaction, artificial intelligence and affective computing.
Host: CS Department
Location: Henry Salvatori Computer Science Center (SAL) - 101
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Philipp Kraehenbuehl (UC Berkeley) - The many ways to understand the pixels, and how to teach computers to do so
Wed, Mar 02, 2016 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Philipp Krahenbuhl , UC Berkeley
Talk Title: The many ways to understand the pixels, and how to teach computers to do so
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium
The field of computer vision is arguably seeing one of its most transformative changes in recent history. Convolutional neural networks (CNNs) have revolutionized the field, reaching super-human performance on some long-standing computer vision tasks, such as image classification. The success of these networks is fueled by massive amounts of human-labeled data. However this paradigm does not scale to a deeper and more detailed understanding of images, as it is simply too hard to collect enough human-labeled data. The issue is not that we humans don't understand the image, but we often struggle to convey enough information to successfully supervise a vision system.
In this talk I show how computer vision can go beyond massive human supervision. This involves designing better models that deal with fewer labels, exploiting easier and more intuitive annotations, or coming up with novel optimizations to train deep architectures with far fewer human annotations, or even without any at all. I'll focus on three long standing computer vision problems: semantic segmentation, intrinsic image decomposition and dense semantic correspondences.
Biography: Philipp Krahenbuhl is a postdoctoral researcher at UC Berkeley. He received a B.S. in Computer Science from ETH Zurich in 2009, and a PhD in Computer Science from Stanford University in 2014. Philipp's research interests lie in Computer vision, Machine learning and Computer Graphics. He is particularly interested in deep learning, efficient optimization techniques, and structured output prediction.
Host: CS Department
Location: Olin Hall of Engineering (OHE) - 136
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CAMS Colloquium: Shang-Hua Teng (USC) - Through the Lens of the Laplacian Paradigm: Big Data and Scalable Algorithms -- a Pragmatic Match Made On Earth
Mon, Mar 07, 2016 @ 03:30 PM - 04:30 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Shang-Hua Teng, USC
Talk Title: Through the Lens of the Laplacian Paradigm: Big Data and Scalable Algorithms -- a Pragmatic Match Made On Earth
Abstract: In the age of Big Data, efficient algorithms are in higher demand now more than ever before. While Big Data takes us into the asymptotic world envisioned by our pioneers, the explosive growth of problem size has also significantly challenged the classical notion of efficient algorithms:
Algorithms that used to be considered efficient, according to polynomial-time characterization, may no longer be adequate for solving today's problems. It is not just desirable, but essential, that efficient algorithms should be scalable. In other words, their complexity should be nearly linear or sub-linear with respect to the problem size. Thus, scalability, not just polynomial-time computability, should be elevated as the central complexity notion for characterizing efficient computation.
In this talk, I will discuss the emerging Laplacian Paradigm, which has led to breakthroughs in scalable algorithms for several fundamental problems in network analysis, machine learning, and scientific computing. I will focus on three recent applications: (1) PageRank Approximation (and identification of network nodes with significant PageRanks). (2) Random-Walk Sparsification. (3) Scalable Newton's Method for Gaussian Sampling.
Biography: Dr. Shang-Hua Teng has twice won the prestigious Godel Prize in theoretical computer science, first in 2008, for developing the theory of smoothed analysis , and then in 2015, for designing the groundbreaking nearly-linear time Laplacian solver for network systems. Both are joint work with Dan Spielman of Yale --- his long-time collaborator. Smoothed analysis is fundamental for modeling and analyzing practical algorithms, and the Laplacian paradigm has since led to several breakthroughs in network analysis, matrix computation, and optimization. Citing him as, "one of the most original theoretical computer scientists in the world", the Simons Foundation named Teng a 2014 Simons Investigator, for pursuing long-term curiosity-driven fundamental research. He and his collaborators also received the best paper award at ACM Symposium on Theory of Computing (STOC) for what's considered to be the "first improvement in 10 years" of a fundamental optimization problem --- the computation of maximum flows and minimum cuts in a network. In addition, he is known for his joint work with Xi Chen and Xiaotie Deng that characterized the complexity for computing an approximate Nash equilibrium in game theory, and his joint papers on market equilibria in computational economics. He and his collaborators also pioneered the development of well-shaped Dalaunay meshing algorithms for arbitrary three-dimensional geometric domains, which settled a long-term open problem in numerical simulation, also a fundamental problem in computer graphics. Software based on this development was used at the University of Illinois for the simulation of advanced rockets. Teng is also interested in mathematical board games. With his former Ph.D. student Kyle Burke, he designed and analyzed a game called Atropos , which is played on the Sperner's triangle and based on the beautiful, celebrated Sperner's Lemma. In 2000 at UIUC, Teng was named on the List of Teachers Ranked as Excellent by Their Students for his class, "Network Security and Cryptography". He has worked and consulted for Microsoft Research, Akamai, IBM Almaden Research Center, Intel Corporation, Xerox PARC, and NASA Ames Research Center, for which he received fifteen patents for his work on compiler optimization, Internet technology, and social network.
Host: USC CAMS
Location: Kaprielian Hall (KAP) - 414
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Andreas Haeberlen (U. of Pennsylvania) - Accountability for Distributed Systems
Tue, Mar 08, 2016 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Andreas Haeberlen, U. of Pennsylvania
Talk Title: Accountability for Distributed Systems
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium
Many of our everyday activities are now performed online - whether it is banking, shopping, or chatting with friends. Behind the scenes, these activities are implemented by large distributed systems that often contain machines from several different organizations. Usually, these machines do what we expect them to, but occasionally they 'misbehave' - sometimes by mistake, sometimes to gain an advantage, and sometimes because of a deliberate attack.
In society, accountability is widely used to counter such threats.
Accountability incentivizes good performance, exposes problems, and builds trust between competing individuals and organizations. In this talk, I will argue that accountability is also a powerful tool for designing distributed systems. An accountable distributed system ensures that 'misbehavior' can be detected, and that it can be linked to a specific machine via some form of digital evidence. The evidence can then be used just like in the 'offline' world, e.g., to correct the problem and/or to take action against the responsible organizations.
I will give an overview of our progress towards accountable distributed systems, ranging from theoretical foundations and efficient algorithms to practical applications. I will also present one result in detail: a technique that can detect information leaks through covert timing channels.
Biography: Andreas Haeberlen is a Raj and Neera Singh Assistant Professor at the University of Pennsylvania. His research interests are in distributed systems, networking, and security. Andreas received his PhD degree in Computer Science from Rice University in 2009; he is the recipient of a NSF CAREER award, and he was awarded the Otto Hahn Medal by the Max Planck Society.
Host: CS Department
Location: Olin Hall of Engineering (OHE) - 136
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Hristo Paskov (Stanford) -Learning with N-Grams: from Massive Scales to Compressed Representations
Tue, Mar 08, 2016 @ 04:00 PM - 05:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Hristo Paskov, Stanford
Talk Title: Learning with N-Grams: from Massive Scales to Compressed Representations
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium
N-gram models are essential in any kind of text processing; they offer simple baselines that are surprisingly competitive with more complicated "state of the art" techniques. I will present a survey of my work for learning with arbitrarily long N-grams at massive scales. This framework combines fast matrix multiplication with a dual learning paradigm that I am developing to reconcile sparsity-inducing penalties with Kernels. The presentation will also introduce Dracula, a new form of deep learning based on classical ideas from compression. Dracula is a combinatorial optimization problem, and I will discuss some its problem structure and use this to visualize its solution surface.
The lecture will be available to stream HERE. Open in new window or tab for best results.
Biography: Hristo Paskov was born in Bulgaria and grew up in New York. He received a B.S. and M.Eng. in Computer Science from MIT while conducting research at the MIT Datacenter and Tomaso Poggio's group (CBCL). He is currently finishing a Ph.D. in Computer Science at Stanford under the advisement of John Mitchell and Trevor Hastie. His research spans machine learning, optimization, and algorithms in order to build large-scale statistical methods and data representations. He is developing a new deep learning paradigm that uses compression to find compact data representations that are useful for statistical inference. His work has provided state of the art methods for security and natural language processing.
Host: CS Department
Location: Henry Salvatori Computer Science Center (SAL) - 101
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: XiaoFeng Wang (Indiana University at Bloomington) - Security Innovations in the Big-Data Era
Wed, Mar 09, 2016 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: XiaoFeng Wang, Indiana University at Bloomington
Talk Title: Security Innovations in the Big-Data Era
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium
The rapid progress in computing has produced a huge amount of data, which will continue to grow in the years to come. In this big-data era, we envision that tomorrow's security technologies will be data-centric: new defense will become smart and proactive by using the data to understand what the attackers have already done, what they are about to do, what their strategies and infrastructures are; effective protection will be provided for dissemination and analysis of the data involving sensitive information on an unprecedented scale. In this talk, I report our first step toward this future of secure computing. We show that through effective analysis of over a million Android apps, previously unknown malware can be detected within a few seconds, without resorting to conventional Anti-Virus means such as signatures and behavior patterns. Also, by leveraging trillions of web pages indexed by search engines, we can capture tens of thousands of compromised websites (including those of government agencies like NIH, NSF and leading education institutions world-wide) by simply asking Google and Bing right questions and automatically analyzing their answers through Natural Language Processing. Further, we found that an in-depth understanding about the unique features of human genomes and how they are used in biomedical research and healthcare systems can help us find a highly efficient way to protect patient privacy during a large-scale genome analysis. Our findings indicate that by unlocking the great value of data, we can revolutionize the security landscape, making tomorrow security technologies more intelligent and effective.
Biography: Dr. XiaoFeng Wang is a professor in the School of Informatics and Computing at Indiana University, Bloomington. He received his Ph.D. in Electrical and Computer Engineering from Carnegie Mellon University in 2004, and has since been a faculty member at IU. Dr. Wang is a well-recognized researcher on system and network security. His work focuses on cloud and mobile security, and data privacy. He is a recipient of 2011 Award for Outstanding Research in Privacy Enhancing Technologies (the PET Award) and the Best Practical Paper Award at the 32nd IEEE Symposium on Security and Privacy. His work frequently receives attention from media, including CNN, MSNBC, Slashdot, CNet, PC World, etc. Examples include his discovery of security-critical vulnerabilities in payment API integrations (http://money.cnn.com/2011/04/13/technology/ecommerce_security_flaw/) and his recent study of the security flaws on the Apple platform (http://money.cnn.com/2015/06/18/technology/apple-keychain-passwords/). His research is supported by the NIH, NSF, Department of Homeland Security, the Air Force and Microsoft Research. He is the director of IU's Center for Security Informatics.
Host: CS Department
Location: Olin Hall of Engineering (OHE) - 136
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Andrew Gordon Wilson (CMU) -Scalable Gaussian Processes for Scientific Discovery
Mon, Mar 21, 2016 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Andrew Gordon Wilson, CMU
Talk Title: Scalable Gaussian Processes for Scientific Discovery
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium
Every minute of the day, users share hundreds of thousands of pictures, videos, tweets, reviews, and blog posts. More than ever before, we have access to massive datasets in almost every area of science and engineering, including genomics, robotics, and astronomy. These datasets provide unprecedented opportunities to automatically discover rich statistical structure, from which we can derive new scientific discoveries. Gaussian processes are flexible distributions over functions, which can learn interpretable structure through covariance kernels. In this talk, I introduce a Gaussian process framework which is capable of learning expressive kernel functions on massive datasets. I will show how this framework generalizes a wide family of scalable machine learning approaches, leverages the inductive biases of deep learning architectures, and allows one to exploit existing model structure for significant further gains in scalability and accuracy, without requiring severe assumptions. I will then discuss how we can use this framework for reverse engineering human learning biases, crime prediction using point processes, image inpainting, video extrapolation, modelling change points and the impacts of vaccine introduction, and discovering the structure and evolution of stars.
Biography: Andrew Gordon Wilson is a Postdoctoral Research Fellow in the Machine Learning Department at Carnegie Mellon University working with Eric Xing and Alexander Smola. Andrew received his PhD in machine learning from the University of Cambridge in 2014, supervised by Zoubin Ghahramani. Andrew's research interests include probabilistic machine learning, scalable inference, Gaussian processes, kernel methods, Bayesian modelling, nonparametrics, and deep learning. Andrew's work has received several awards, including the G-Research Outstanding Dissertation Award in 2014 and the Best Student Paper Award at the Conference on Uncertainty in Artificial Intelligence in 2011.
Host: CS Department
Location: Olin Hall of Engineering (OHE) - 136
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Olga Russakovsky (CMU) - The Human Side of Computer Vision
Wed, Mar 23, 2016 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Olga Russakovsky, Carnegie Mellon University
Talk Title: The Human Side of Computer Vision
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium
Intelligent agents acting in the real world need advanced vision capabilities to perceive, learn from, reason about and interact with their environment. In this talk, I will explore the role that humans play in the design and deployment of computer vision systems. Large-scale manually labeled datasets have proven instrumental for scaling up visual recognition, but they come at a substantial human cost. I will first briefly talk about strategies for making optimal use of human annotation effort for computer vision progress. However, no dataset can foresee all the visual scenarios that a real-world system might encounter. I will argue that seamlessly integrating human expertise at runtime will become increasingly important for open-world computer vision. I will introduce, and demonstrate the effectiveness of, a rigorous mathematical framework for human-machine collaboration. Looking ahead, in order for such collaborations to become practical, the computer vision algorithms we design will need to be both efficient and interpretable. I will conclude by presenting a new deep reinforcement learning model for human action detection in videos that is efficient, interpretable and more accurate than prior art, opening up new avenues for practical human-in-the-loop exploration.
Biography: Olga Russakovsky recently completed her PhD in computer science at Stanford and is now a postdoctoral fellow at Carnegie Mellon University. Her research is in computer vision, closely integrated with machine learning and human-computer interaction. Her work was featured in the New York Times and MIT Technology Review. She served as a Senior Program Committee member for WACV'16, led the ImageNet Large Scale Visual Recognition Challenge effort for two years, and organized multiple workshops and tutorials on large-scale recognition at premier computer vision conferences ICCV'13, ECCV'14, CVPR'15, ICCV'15 and CVPR'16. In addition, she founded and directs the Stanford AI Laboratory's outreach camp SAILORS (featured in Wired and published in SIGCSE'16) designed to expose high school students in underrepresented populations to the field of AI.
Host: CS Department
Location: Olin Hall of Engineering (OHE) - 136
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Linh Thi Xuan Phan (U. of Pennsylvania) - Timing Guarantees for Cyber-Physical Systems
Wed, Mar 23, 2016 @ 04:00 PM - 05:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Linh Thi Xuan Phan, U. of Pennsylvania
Talk Title: Timing Guarantees for Cyber-Physical Systems
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium
Cyber-physical systems -- such as cars, pacemakers, and power plants -- need to interact with the physical world in a timely manner to ensure safety. It is important to have a way to analyze these systems and to prove that they can meet their timing requirements. However, modern cyber-physical systems are increasingly complex: they can involve thousands of tasks running on dozens of processors, many of which can have multiple cores or shared caches. Existing techniques for ensuring timing guarantees cannot handle this level of complexity. In this talk, I will present some of my recent work that can help to bridge this gap, such as overhead-aware compositional scheduling/analysis and multicore cache management. I will also discuss some potential applications, such as real-time cloud platforms and intrusion-resistant cyber-physical systems.
Biography: Linh Thi Xuan Phan is an Assistant Research Professor in the Department of Computer and Information Science at the University of Pennsylvania. Her interests include real-time systems, embedded systems, cyber-physical systems, and cloud computing. Her research develops theoretical foundations and practical tools for building complex systems with provable safety and timing guarantees. She is especially interested in techniques that integrate theory, systems, and application aspects. Recently, she has been working on methods for defending cyber-physical systems against malicious attacks, as well as on real-time cloud infrastructures for safety critical and mission-critical systems. Linh holds a Ph.D. degree in Computer Science from the National University of Singapore (NUS); she received the Graduate Research Excellence Award from NUS for her dissertation work.
Host: CS Department
More Info: https://bluejeans.com/333182321
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
Event Link: https://bluejeans.com/333182321
-
CS Colloquium: Joseph Lim (MIT) - Toward Visual Understanding of the Physical World for Interaction
Thu, Mar 24, 2016 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Joseph Lim, MIT
Talk Title: Toward Visual Understanding of the Physical World for Interaction
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium
Recently, the computer vision community has made impressive progress on object recognition with deep learning approaches. However, for any visual system to interact with objects, it needs to understand much more than simply recognizing where the objects are. The goal of my research is to explore and solve physical understanding tasks for interaction -- finding an object's pose in 3D, interpreting its physical interactions, and understanding its various states and transformations. Unfortunately, obtaining extensive annotated data for such tasks is often intractable, yet required by recent popular learning techniques.
In this talk, I take a step away from expensive, manually labeled datasets. Instead, I develop learning algorithms that are supervised through physical constraints combined with structured priors. I will first talk about how to build learning algorithms, including a deep learning framework (e.g., convolutional neural networks), that can utilize geometric information from 3D CAD models in combination with real-world statistics from photographs. Then, I will show how to use differentiable physics simulators to learn object properties simply by watching videos.
Biography: Joseph Lim is a postdoctoral researcher at Stanford University. He received a PhD in Electrical Engineering and Computer Science at Massachusetts Institute of Technology, where he was advised by Professor Antonio Torralba. His research interests are in computer vision and machine learning. He is particularly interested in deep learning, structure learning, and multi-domain data. Joseph graduated with BA in Computer Science from UC Berkeley, where he worked under Professor Jitendra Malik.
Host: CS Department
Location: Olin Hall of Engineering (OHE) - 136
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Jason Polakis (Columbia U.) -Protecting Users in the Age of the Social Web
Thu, Mar 24, 2016 @ 04:00 PM - 05:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Jason Polakis, Columbia University
Talk Title: Protecting Users in the Age of the Social Web
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium
In this talk I will focus on my research efforts to better understand and protect against such loss. I will start with a focused review on the importance of online privacy, and highlight the privacy risks of location proximity, which has been adopted by major web services and mobile apps. This work demonstrated novel threats that can neutralize existing countermeasures used by the industry and pinpoint a user's location with high accuracy within seconds. To protect users, I developed a practical defense in the form of privacy-preserving proximity that obfuscates the user's location, which has been adopted by Facebook and Foursquare. I will demonstrate how user privacy also affects security mechanisms, and present my analysis of the threat surface of Facebook's social authentication system. I will then present a novel social authentication system that is robust against advanced targeted attacks and prevents adversaries from compromising user accounts, and conclude by sharing my thoughts for future directions.
This lecture will be available to stream HERE.
Biography: Jason Polakis is a postdoctoral research scientist at Columbia University. He earned his PhD in 2014 from the Computer Science Department of the University of Crete, Greece, where he was supported by the Foundation of Research and Technology Hellas (FORTH). He is broadly interested in identifying the security and privacy limitations of Internet technologies, designing robust defenses and privacy-preserving techniques, and enhancing our understanding of the online ecosystem and its threats. His research has revealed significant flaws in popular services, and major vendors such as Google, Facebook and Foursquare have deployed his proposed defenses. His work has been published in top tier security conferences (Security and Privacy, CCS, and NDSS) as well as other top tier computer science conferences (WWW).
Host: CS Department
More Info: https://bluejeans.com/313531059
Location: Henry Salvatori Computer Science Center (SAL) - 101
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
Event Link: https://bluejeans.com/313531059
-
CS Colloquium: David Fouhey (CMU) -Towards A Physical and Human-Centric Understanding of Images
Mon, Mar 28, 2016 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: David Fouhey, CMU
Talk Title: Towards A Physical and Human-Centric Understanding of Images
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium
One primary goal of AI from its very beginning has been to develop systems that can understand an image in a meaningful way. While we have seen tremendous progress in recent years on naming-style tasks like image classification or object detection, a meaningful understanding requires going beyond this paradigm. Scenes are inherently 3D, so our understanding must also capture the underlying 3D and physical properties. Additionally, our understanding must be human-centric since any man-made scene has been built with humans in mind. Despite the importance of obtaining a 3D and human-centric understanding, we are only beginning to scratch the surface on both fronts: many fundamental questions, in terms of how to both frame and solve the problem, remain unanswered.
In this talk, I will discuss my efforts towards building a physical and human-centric understanding of images. I will present work addressing the questions: (1) what 3D properties should we model and predict from images, and do we actually need explicit 3D training data to do this? (2) how can we reconcile data-driven learning techniques with the physical constraints that exist in the world? and (3) how can understanding humans improve traditional 3D and object recognition tasks?
Biography: David Fouhey is a Ph.D. student at the Robotics Institute of Carnegie Mellon University, where he is advised by Abhinav Gupta and Martial Hebert. His research interests include computer vision and machine learning with a particular focus on scene understanding. David's work has been supported by both NSF and NDSEG fellowships. He has spent time at Microsoft Research and University of Oxford's Visual Geometry Group.
Host: CS Department
Location: Olin Hall of Engineering (OHE) - 136
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Tuo Zhao (Johns Hopkins University) - Compute Faster and Learn Better: Machine Learning via Nonconvex Model-based Optimization
Mon, Mar 28, 2016 @ 03:00 PM - 04:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Tuo Zhao , Johns Hopkins University
Talk Title: Compute Faster and Learn Better: Machine Learning via Nonconvex Model-based Optimization
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium
Nonconvex optimization naturally arises in many machine learning problems (e.g. sparse learning, matrix factorization, and tensor decomposition). Machine learning researchers exploit various nonconvex formulations to gain modeling flexibility, estimation robustness, adaptivity, and computational scalability. Although classical computational complexity theory has shown that solving nonconvex optimization is generally NP-hard in the worst case, practitioners have proposed numerous heuristic optimization algorithms, which achieve outstanding empirical performance in real-world applications.
To bridge this gap between practice and theory, we propose a new generation of model-based optimization algorithms and theory, which incorporate the statistical thinking into modern optimization. Particularly, when designing practical computational algorithms, we take the underlying statistical models into consideration (e.g. sparsity, low rankness). Our novel algorithms exploit hidden geometric structures behind many nonconvex optimization problems, and can obtain global optima with the desired statistics properties in polynomial time with high probability.
Biography: Tuo Zhao is a PhD student in Department of Computer Science at Johns Hopkins University (http://www.cs.jhu.edu/~tour). His research focuses on high dimensional parametric and semiparametric learning, large-scale optimization, and applications to computational genomics and neuroimaging. He was the core member of the JHU team winning the INDI ADHD 200 global competition on fMRI imaging-based diagnosis classification in 2011. He received Siebel scholarship in 2014 and Baidu's research fellowship in 2015
Host: CS Department
More Info: https://bluejeans.com/741584974
Location: Ronald Tutor Hall of Engineering (RTH) - 526
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
Event Link: https://bluejeans.com/741584974
-
CS Colloquium: Bogdan Vasilescu (UC Davis) - Lessons in Social Coding: Software Analytics in the Age of GitHub
Mon, Mar 28, 2016 @ 04:00 PM - 05:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Bogdan Vasilescu, UC Davis
Talk Title: Lessons in Social Coding: Software Analytics in the Age of GitHub
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium
Social media has forever changed the ways in which we communicate and work, programming included. This "social coding" movement (code is meant to be shared!) made popular by GitHub has come to represent a paradigm shift in software development, especially in the open-source world. For example, the "pull request" model has made it easier than ever before for newcomers to submit contributions to a project. As a result, teams are becoming increasingly larger, more distributed, and more diverse. At the same time, the incentives for contributing have evolved. For example, one's social coding activity is starting to replace one's resume, and directly influence their hourly wage. Today, GitHub reports 12 million users and over 30 million repositories, with popular projects having communities the size of small cities. These numbers are unprecedented in open-source!
This new, social way of developing software opens a great many questions. How do people choose which projects to contribute to? Does prior technical experience matter, or do people learn on the job? Is it efficient to work on many projects in parallel? How does diversity in software teams affect productivity and code quality? What are the main factors that slow down pull request reviews? How does automation help developers do more with less? Does continuous integration help to ensure higher quality code? I will try to answer some of these questions in this talk.
Biography: Bogdan Vasilescu is currently a postdoctoral researcher at University of California, Davis (USA), where he is a member of the Davis Eclectic Computational Analytics Lab (DECAL). He received his PhD and MSc in Computer Science at Eindhoven University of Technology, both with cum laude distinction. His PhD dissertation won the best dissertation award from the Dutch Institute for Programming Research and Algorithmics in 2015. Follow him on Twitter @b_vasilescu
Host: CS Department
More Info: https://bluejeans.com/958446198
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
Event Link: https://bluejeans.com/958446198
-
CS Colloquium: David Levin (Disney Research Boston) - Physically-Based Simulation for Animation and Fabrication
Tue, Mar 29, 2016 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: David Levin, Disney Research Boston
Talk Title: Physically-Based Simulation for Animation and Fabrication
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium
Physics-based simulation has become a transformative tool for solving problems in computer animation and computational fabrication. In this talk I will discuss how leveraging unique abstractions, new discretizations and data-driven techniques can allow us to animate and fabricate a wide-range of phenomena with improved performance, robustness and accuracy. I'll show how layered discretizations can enable photoshop like editing of physically-based animations, how Eulerian methods can be used to robustly simulate deforming objects in close contact, how 3D printing and simulation can produce new musical instruments, and more. I'll conclude by discussing the important challenges facing physics-based animation and fabrication now and in the future.
Host: CS Department
Location: Olin Hall of Engineering (OHE) - 136
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Nisarg Shah (CMU) - Optimal Social Decision Making
Tue, Mar 29, 2016 @ 04:00 PM - 05:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Nisarg Shah, Carnegie Mellon University
Talk Title: Optimal Social Decision Making
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium
How can computers help ordinary people make collective decisions about real-life dilemmas, like which restaurant to go to with friends, or even how to divide an inheritance? I will present an optimization-driven approach that draws on ideas from AI, theoretical computer science, and economic theory, and illustrate it through my research in computational social choice and computational fair division. In both areas, I will make a special effort to demonstrate how fundamental theoretical questions underlie the design and implementation of deployed services that are already used by tens of thousands of people (spliddit.org), as well as upcoming services (robovote.org).
Biography: Nisarg Shah is a Ph.D. candidate in the Computer Science Department at Carnegie Mellon University, advised by Ariel Procaccia. His broad research agenda in algorithmic economics includes topics such as computational social choice, fair division, game theory (both cooperative and noncooperative), and prediction markets. He focuses on designing theoretically grounded methods that have practical implications. Shah is the winner of the 2013-2014 Hima and Jive Graduate Fellowship and the 2014-2015 Facebook Fellowship.
Host: CS Department
More Info: https://bluejeans.com/246853239
Location: Henry Salvatori Computer Science Center (SAL) - 101
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
Event Link: https://bluejeans.com/246853239
-
CS Colloquium: Baris Kasikci (EPFL) - Stamping Out Concurrency Bugs
Thu, Mar 31, 2016 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Baris Kasikci, EPFL
Talk Title: Stamping Out Concurrency Bugs
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium
The shift to multi-core architectures in the past ten years pushed developers to write concurrent software to leverage hardware parallelism. The transition to multi-core hardware happened at a more rapid pace than the evolution of associated programming techniques and tools, which made it difficult to write concurrent programs that are both efficient and correct. Failures due to concurrency bugs are often hard to reproduce and fix, and can cause significant losses.
In this talk, I will first give an overview of the techniques we developed for the detection, root cause diagnosis, and classification of concurrency bugs. Then, I will discuss how the techniques we developed have been adopted at Microsoft and Intel. I will then discuss in detail Gist, a technique for the root cause diagnosis of failures. Gist uses hybrid static-dynamic program analysis and gathers information from real user executions to isolate root causes of failures. Gist is highly accurate and efficient, even for failures that rarely occur in production. Finally, I will close by describing future work I plan to do toward solving the challenges posed to software systems by emerging technology trends.
Biography: Baris Kasikci completed his Ph.D. in the Dependable Systems Laboratory (DSLAB) at EPFL, advised by George Candea. His research is centered around developing techniques, tools, and environments that help developers build more reliable and secure software. He is interested in finding solutions that allow programmers to better reason about their code, and that efficiently detect bugs, classify them, and diagnose their root cause. He especially focuses on bugs that manifest in production, because they are hard and time-consuming. He is also interested in efficient runtime instrumentation, hardware and runtime support for enhancing system security, and program analysis under various memory models.
Baris is one of the four recipients of the VMware 2014-2015 Graduate Fellowship. During his Ph.D., he interned at Microsoft Research, VMware, and Intel. Before starting his Ph.D., he worked as a software engineer for four years, mainly developing real-time embedded systems software. Before joining EPFL, he was working for Siemens Corporate Technology. More details can be found at http://www.bariskasikci.org/.
Host: CS Department
Location: Olin Hall of Engineering (OHE) - 136
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Konrad Kording (Northwestern University) - Neural Cryptography
Thu, Mar 31, 2016 @ 03:00 PM - 04:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Konrad Kording, Northwestern University
Talk Title: Neural Cryptography
Series: CS Colloquium
Abstract: Neuroscience is slowly transitioning into a data rich discipline and large data sets allow new approaches. Brain decoders use neural recordings to infer what someone is thinking, viewing, or their intended movement. The problem has always been phrased as a supervised learning problem. Here we introduce a new method for brain decoding that does not require supervised data, i.e. the knowledge of the intended movement while the neural activity is recorded. Our approach is inspired by code breaking techniques used in cryptography where it is asked which mapping from from encrypted to decrypted text leads to text that most resembles the known structure of language. Analogously, we find a transformation of neural data (decoder) that aligns the distribution of the decoder output with the distribution of the user's intended movement. On a standard primate center-out reaching task, we demonstrate that we can obtain similar performance with that of a decoder with access to supervised data. However, current datasets are still too small to ask many relevant questions about neural computation and I am collaborating with neuroengineers to change that.
Host: CS Department
More Info: https://bluejeans.com/942986114
Location: Ronald Tutor Hall of Engineering (RTH) - 526
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
Event Link: https://bluejeans.com/942986114
-
CS Colloquium: Cynthia Sung (MIT CSAIL) - Computational Tools for Robot Design: A Composition Approach
Thu, Mar 31, 2016 @ 04:00 PM - 05:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Cynthia Sung, MIT CSAIL
Talk Title: Computational Tools for Robot Design: A Composition Approach
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium
As robots become more prevalent in society, they must develop an ability to deal with more diverse situations. This ability entails customizability of not only software intelligence, but also of hardware. However, designing a functional robot remains challenging and often involves many iterations of design and testing even for skilled designers. My goal is to create computational tools for making functional machines, allowing future designers to quickly improvise new hardware.
In this talk, I will discuss one possible approach to automated design using composition. I will describe our origami-inspired print-and-fold process that allows entire robots to be fabricated within a few hours, and I will demonstrate how foldable modules can be composed together to create foldable mechanisms and robots. The modules are represented parametrically, enabling a small set of modules to describe a wide range of geometries and also allowing geometries to be optimized in a straightforward manner. I will also introduce a tool that we have developed that combines this composition approach with simulations to help human designers of all skill levels to design and fabricate custom functional robots.
Biography: Cynthia Sung is a Ph.D. candidate in the Computer Science and Artificial Intelligence Laboratory at the Massachusetts Institute of Technology (MIT). She received a B.S. in Mechanical Engineering from Rice University in 2011 and an M.S. in Electrical Engineering and Computer Science from MIT in 2013. Her research interests include computational design, folding theory, and rapid fabrication, and her current work focuses on algorithms for synthesis and analysis of engineering designs.
Host: CS Department
More Info: https://bluejeans.com/727861390
Location: Henry Salvatori Computer Science Center (SAL) - 101
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
Event Link: https://bluejeans.com/727861390
