- talks - lists.cs.princeton.edu

Colloquium Speaker Yann LeCun Weds Oct 22, 4:30pm
by Nicole E. Wagenblast 16 Oct '14

16 Oct '14

Yann LeCun, Facebook AI Research and Center for Data Science, New York University The Unreasonable Effectiveness of Deep Learning Wednesday, October 22, 4:30pm Computer Science 105 The emergence of large datasets, parallel computers, and new machine learning methods, have enabled the deployment of highly-accurate computer perception systems and are opening the door to a wide deployment of AI systems. A key component in AI systems is a module, sometimes called a feature extractor, that turns raw inputs into suitable internal representations. But designing and building such a module requires a considerable amount of engineering efforts and domain expertise. Deep Learning methods have provided a way to automatically learn good representations of data from labeled or unlabeled samples. Deep architectures are composed of successive stages in which data representations are increasingly global, abstract, and invariant to irrelevant transformations of the input. Deep learning enables end-to-end training of these architectures, from raw inputs to ultimate outputs. The convolutional network model (ConvNet) is a particular type of deep architecture somewhat inspired by biology, which consists of multiple stages of filter banks, interspersed with non-linear operators, and spatial pooling. ConvNets have become the record holder for a wide variety of benchmarks, including object detection, localization and recognition in image, semantic segmentation and labeling, face recognition, acoustic modeling for speech recognition, drug design, handwriting recognition, biological image segmentation, etc. The most recent systems deployed by Facebook, Google, NEC, IBM, Microsoft, Baidu, Yahoo and others for image understanding, speech recognition, and natural language processing use deep learning. Many of these systems use very large and very deep ConvNets with billions of connections, trained in supervised mode. But many new applications require the use of unsupervised feature learning. A number of such methods based on sparse auto-encoder will be presented. Several applications will be shown through videos and live demos, including a category-level object recognition system that can be trained on the fly, a scene parsing system that can label every pixel in an image with the category of the object it belongs to (scene parsing), an object localization and detection system, and several natural language processing systems. Specialized hardware architectures that run these systems in real time will also be described. Yann LeCun is Director of AI Research at Facebook, and Silver Professor of Data Science, Computer Science, Neural Science, and Electrical Engineering at New York University, affiliated with the NYU Center for Data Science, the Courant Institute of Mathematical Science, the Center for Neural Science, and the Electrical and Computer Engineering Department. He received the Electrical Engineer Diploma from Ecole Supérieure d'Ingénieurs en Electrotechnique et Electronique (ESIEE), Paris in 1983, and a PhD in Computer Science from Université Pierre et Marie Curie (Paris) in 1987. After a postdoc at the University of Toronto, he joined AT&T Bell Laboratories in Holmdel, NJ in 1988. He became head of the Image Processing Research Department at AT&T Labs-Research in 1996, and joined NYU as a professor in 2003, after a brief period as a Fellow of the NEC Research Institute in Princeton. From 2012 to 2014 he directed NYU's initiative in data science and became the founding director of the NYU Center for Data Science. He was named Director of AI Research at Facebook in late 2013 and retains a part-time position on the NYU faculty. His current interests include AI, machine learning, computer perception, mobile robotics, and computational neuroscience. He has published over 180 technical papers and book chapters on these topics as well as on neural networks, handwriting recognition, image processing and compression, and on dedicated circuits and architectures for computer perception. The character recognition technology he developed at Bell Labs is used by several banks around the world to read checks and was reading between 10 and 20% of all the checks in the US in the early 2000s. His image compression technology, called DjVu, is used by hundreds of web sites and publishers and millions of users to access scanned documents on the Web. Since the mid 1980's he has been working on deep learning methods, particularly the convolutional network model, which is the basis of many products and services deployed by companies such as Facebook, Google, Microsoft, Baidu, IBM, NEC, AT&T and others for image and video understanding, document recognition, human-computer interaction, and speech recognition. LeCun has been on the editorial board of IJCV, IEEE PAMI, and IEEE Trans. Neural Networks, was program chair of CVPR'06, and is chair of ICLR 2013 and 2014. He is on the science advisory board of Institute for Pure and Applied Mathematics, and Neural Computation and Adaptive Perception Program of the Canadian Institute for Advanced Research. He has advised many large and small companies about machine learning technology, including several startups he co-founded. He is the lead faculty at NYU for the Moore-Sloan Data Science Environment, a $36M initiative in collaboration with UC Berkeley and University of Washington to develop data-driven methods in the sciences. He is the recipient of the 2014 IEEE Neural Network Pioneer Award.

1 0

Colloquium Speaker Trevor Darrell Today 4:30pm
by Nicole E. Wagenblast 16 Oct '14

16 Oct '14

Large-Scale Detector Adaptation and Other Recent Results Trevor Darrell , University of California, Berkeley Thursday, October 16, 4:30pm Computer Science 105 In this talk I'll review recent progress towards robust and effective perceptual representation learning. I'll describe new methods for large-scale detection, whereby robust detectors can be learned from weakly labeled training data, following paradigms of domain adaptation and multiple instance learning. I'll discuss how such models can be used not only for detection but also for pose prediction and further for effective fine-grained recognition, extending traditional convolutional neural network models to include explicit pose-normalized descriptors. Finally, and time permitting (pardon the pun), I'll review our recent work on anytime recognition, which provides methods that strive to provide the best answer possible, even with a limited (and unknown) time budget. Prof. Trevor Darrell’s group is co-located at the University of California, Berkeley, and the UCB-affiliated International Computer Science Institute (ICSI), also located in Berkeley, CA. Prof. Darrell is on the faculty of the CS Division of the EECS Department at UCB and is the vision group lead at ICSI. Darrell’s group develops algorithms for large-scale perceptual learning, including object and activity recognition and detection, for a variety of applications including multimodal interaction with robots and mobile devices. His interests include computer vision, machine learning, computer graphics, and perception-based human computer interfaces. Prof. Darrell was previously on the faculty of the MIT EECS department from 1999-2008, where he directed the Vision Interface Group. He was a member of the research staff at Interval Research Corporation from 1996-1999, and received the S.M., and PhD. degrees from MIT in 1992 and 1996, respectively. He obtained the B.S.E. degree from the University of Pennsylvania in 1988, having started his career in computer vision as an undergraduate researcher in Ruzena Bajcsy's GRASP lab.

1 0

Colloquium Speaker Trevor Darrell Thurs Oct 16, 4:30pm
by Nicole E. Wagenblast 15 Oct '14

15 Oct '14

Large-Scale Detector Adaptation and Other Recent Results Trevor Darrell , University of California, Berkeley Thursday, October 16, 4:30pm Computer Science 105 In this talk I'll review recent progress towards robust and effective perceptual representation learning. I'll describe new methods for large-scale detection, whereby robust detectors can be learned from weakly labeled training data, following paradigms of domain adaptation and multiple instance learning. I'll discuss how such models can be used not only for detection but also for pose prediction and further for effective fine-grained recognition, extending traditional convolutional neural network models to include explicit pose-normalized descriptors. Finally, and time permitting (pardon the pun), I'll review our recent work on anytime recognition, which provides methods that strive to provide the best answer possible, even with a limited (and unknown) time budget. Prof. Trevor Darrell’s group is co-located at the University of California, Berkeley, and the UCB-affiliated International Computer Science Institute (ICSI), also located in Berkeley, CA. Prof. Darrell is on the faculty of the CS Division of the EECS Department at UCB and is the vision group lead at ICSI. Darrell’s group develops algorithms for large-scale perceptual learning, including object and activity recognition and detection, for a variety of applications including multimodal interaction with robots and mobile devices. His interests include computer vision, machine learning, computer graphics, and perception-based human computer interfaces. Prof. Darrell was previously on the faculty of the MIT EECS department from 1999-2008, where he directed the Vision Interface Group. He was a member of the research staff at Interval Research Corporation from 1996-1999, and received the S.M., and PhD. degrees from MIT in 1992 and 1996, respectively. He obtained the B.S.E. degree from the University of Pennsylvania in 1988, having started his career in computer vision as an undergraduate researcher in Ruzena Bajcsy's GRASP lab.

1 0

Colloquium Speaker Alex Halderman- Wed Oct 15, 4:30pm
by Nicole E. Wagenblast 15 Oct '14

15 Oct '14

Colloquium Speaker J. Alex Halderman, University of Michigan Wednesday, October 15, 4:30pm Computer Science 105 The Internet's size, and the diversity of connected hosts, create difficult challenges for security. Conventionally, most vulnerabilities are discovered through labor-intensive scrutiny of individual implementations, but this scales poorly, and important classes of vulnerabilities can be hard to detect when considering hosts in isolation. Moreover, the security of the Internet as a whole is affected by management decisions made by individual system operators, but it is difficult to make sense of these choices--or to influence them to improve security--without a global perspective. In recent work, I have been developing new approaches to these challenges, based on the analysis of large-scale Internet measurement data. By collecting and analyzing the public keys used for HTTPS and SSH, my team discovered serious weaknesses in key generation affecting millions of machines, and we were able to efficiently factor the RSA moduli used by almost 0.5% of all HTTPS servers. By clustering and investigating the vulnerable hosts, we exposed flawed cryptographic implementations in network devices manufactured by more than 60 companies and uncovered a critical design flaw in the Linux kernel. To help other researchers apply similar techniques, we developed ZMap, a tool for performing Internet-wide network surveys that can probe the entire IPv4 address space in minutes, thousands of times faster than prior approaches. ZMap has become a thriving open-source project and is available in major Linux distributions. We've used it to develop defenses against compromised HTTPS certificate authorities, to study the Internet's response to the infamous OpenSSL Heartbleed vulnerability, and to significantly increase the global rate of patching for vulnerable hosts. Ultimately, measurement-driven approaches to Internet security may help shift the security balance of power to favor defenders over attackers. J. Alex Halderman is an assistant professor of computer science and engineering at the University of Michigan and director of Michigan's Center for Computer Security and Society. His research focuses on computer security and privacy, with an emphasis on problems that broadly impact society and public policy. Prof. Halderman's interests include application security, network security, anonymous and censorship-resistant communication, electronic voting, digital rights management, mass surveillance, and online crime, as well as the interaction of technology with law, regulatory policy, and international affairs. Prof. Halderman is widely known for developing the "cold boot" attack against disk encryption, which altered widespread security assumptions about the behavior of RAM, influenced computer forensics practice, and inspired the creation of a new subfield of theoretical cryptography. A noted expert on electronic voting security, he helped lead the first independent review of the election technology used by half a billion voters in India, which prompted the national government to undertake major technical reforms. He has authored more than 50 publications, and his work has won numerous distinctions, including two best paper awards from Usenix Security, a top systems security venue.

1 2

Colloquium Speaker Trevor Darrell Thurs Oct 16, 4:30pm
by Nicole E. Wagenblast 10 Oct '14

10 Oct '14

Large-Scale Detector Adaptation and Other Recent Results Trevor Darrell , University of California, Berkeley Thursday, October 16, 4:30pm Computer Science 105 In this talk I'll review recent progress towards robust and effective perceptual representation learning. I'll describe new methods for large-scale detection, whereby robust detectors can be learned from weakly labeled training data, following paradigms of domain adaptation and multiple instance learning. I'll discuss how such models can be used not only for detection but also for pose prediction and further for effective fine-grained recognition, extending traditional convolutional neural network models to include explicit pose-normalized descriptors. Finally, and time permitting (pardon the pun), I'll review our recent work on anytime recognition, which provides methods that strive to provide the best answer possible, even with a limited (and unknown) time budget. Prof. Trevor Darrell’s group is co-located at the University of California, Berkeley, and the UCB-affiliated International Computer Science Institute (ICSI), also located in Berkeley, CA. Prof. Darrell is on the faculty of the CS Division of the EECS Department at UCB and is the vision group lead at ICSI. Darrell’s group develops algorithms for large-scale perceptual learning, including object and activity recognition and detection, for a variety of applications including multimodal interaction with robots and mobile devices. His interests include computer vision, machine learning, computer graphics, and perception-based human computer interfaces. Prof. Darrell was previously on the faculty of the MIT EECS department from 1999-2008, where he directed the Vision Interface Group. He was a member of the research staff at Interval Research Corporation from 1996-1999, and received the S.M., and PhD. degrees from MIT in 1992 and 1996, respectively. He obtained the B.S.E. degree from the University of Pennsylvania in 1988, having started his career in computer vision as an undergraduate researcher in Ruzena Bajcsy's GRASP lab.

1 0

Colloquium Speaker: Rahul Sukthankar, Thurs. Oct. 9 4:30pm
by Nicole E. Wagenblast 09 Oct '14

09 Oct '14

The CouchPotato Project: Learning Visual Concepts by Watching YouTube Rahul Sukthankar ( Google Research ) Thursday, October 9, 4:30pm Computer Science 105 What could computers learn about visual concepts, such as objects and actions, by watching large quantities of internet video, with minimal human supervision? In this talk, I will present several of our recent explorations in this area, including spatio-temporal object segmentation in video, weakly-supervised learning of human actions and unsupervised discovery of motion patterns in animal videos. The talk includes research contributions from many colleagues at Google, interns and faculty colleagues: L. Del Pero, I. Essa, L. Fei-Fei, V. Ferrari, M. Grundmann, G. Hartmann, J. Hoffman, A. Karpathy, T. Leung, V. Kwatra, O. Madani, J. Rehg, S. Ricco, S. Shetty, K. Tang, G. Toderici, D. Tsai, J. Yagnik. Rahul Sukthankar is a scientist at Google Research, an adjunct research professor in the Robotics Institute at Carnegie Mellon and courtesy faculty at UCF. He was previously a senior principal researcher at Intel Labs, a senior researcher at HP/Compaq and research scientist at Just Research. Rahul received his Ph.D. in Robotics from Carnegie Mellon in 1997 and his B.S.E. in Computer Science from Princeton in 1991. His current research focuses on computer vision and machine learning, particularly in the areas of object recognition and video understanding.

1 2

Andrej Risteski will present his Generals on Tuesday, Oct. 14th at 10:30am in CS 302
by Nicki Gotsis 07 Oct '14

07 Oct '14

Andrej Risteski will present his Generals on Oct. 14th at 10:30am in CS 302 The members of his committee are Sanjeev Arora, Moses Charikar, and David Blei. Everyone is invited to attend the talk, and those faculty wishing to remain for the oral exam following are welcome to do so. His abstract and reading list follow below. Abstract: Variational methods are a fairly commonly used tool in machine learning to solve inference problems of various kinds. The way they work is to reduce the problem of calculating some posterior probability to a suitable, usually non-convex, mathematical program. This is then solved using some form of alternating minimization. The vanilla form of this is expectation maximization (EM), and some more sophisticated flavors are variational Bayes, expectation propagation, etc. Unfortunately, all of these rarely come with any guarantees, and even in practice are prone to local optima. I will discuss some simple cases where these methods, along with suitable initialization, converge to the right value, mostly in the setting of topic models. Reading list: 0. Algorithm Design; Kleinberg, Tardos 1. Graphical Models, Exponential Families, and Variational Inference; Jordan, Wainwright [Chapters 1,2,3,4,7,9] 2. Latent Dirichlet Allocation; Blei, Ng, Jordan 3. Learning Topic Models - Going beyond SVD; Arora, Ge, Moitra 4. A Practical Algorithm for Topic Modeling with Provable Guarantees; Arora et al. 5. Computing a Nonnegative Matrix Factorization --- Provably; Arora, Ge, Kannan, Moitra 6. Complexity of Inference in Graphical Models; Chandresekaran, Srebro, Harsha 7. Understanding Belief Propagation and its Generalizations; Yedidia, Freeman, Weiss 8. Convexity Arguments for Efficient Minimization of the Bethe and Kikuchi Free Energies; Heskes 9. Convexifying the Bethe Free Energy; Meshi, Jaimovich, Globerson, Friedman 10. Counting independent sets up to the tree threshold; Weitz 11. Counting Independent Sets Using the Bethe Approximation; Chandrasekaran, Chertkov, Gamarnik, Shah, Shin 12. Clustering with Spectral Norm and the k-means Algorithm; Kumar, Kannan

1 0

4:30pm Tue Oct 7 talk on multipath routing in CS 402
by Jennifer Rexford 06 Oct '14

06 Oct '14

Speaker: Deep Medhi, University of Missouri-Kansas City Title: Is Multipath Routing Beneficial? Date/time: 4:30pm Tuesday October 7 Location: CS 402 Abstract: It's often believed that multipath routing is always beneficial. Taking a traffic engineering perspective, we consider commonly used goals such as congestion minimization, minimum average delay, and minimum cost routing. Using a well-known result from linear programming, we show that numbers of paths taken by all demands at optimality is limited by the total number of demands and links in a network. When all node pairs (demands) in a network have traffic, multipath routing essentially becomes single-path routing, especially as the network becomes large. Under certain traffic conditions, single-path routing is found to be optimal. We will also present results on a number of traffic scenarios and load conditions using topologies used by ISPs and in data center networks. These observations are counter-intuitive due to our commonly held belief about multipath routing. Bio: Deep Medhi is Curators' Professor in the Department of Computer Science and Electrical Engineering at the University of Missouri-Kansas City, USA. He received B.Sc. in Mathematics from Cotton College, Gauhati University, India, M.Sc. in Mathematics from the University of Delhi, India, and his Ph.D. in Computer Sciences from the University of Wisconsin-Madison, USA. Prior to joining UMKC in 1989, he was a member of the technical staff at AT&T Bell Laboratories. He was an invited visiting professor at the Technical University of Denmark, a visiting research fellow at Lund Institute of Technology, Sweden, a research visitor at University of Campinas, Brazil under the Brazilian Science Mobility Program and served as a Fulbright Senior Specialist. He is the Editor-in-Chief of Springer’s Journal of Network and Systems Management, and is on the editorial board of IEEE/ACM Transactions on Networking, IEEE Transactions on Network and Service Management, and IEEE Communications Surveys & Tutorials. He is co-author of the books, Routing, Flow, and Capacity Design in Communication and Computer Networks (2004) and Network Routing: Algorithms, Protocols, and Architectures (2007), both published by Morgan Kauffman/Elsevier.

1 0

Fwd: EE Seminar - October 7, 2014, 12:00 pm, E-Quad B327 - Dr. Wei Wang
by Jennifer Rexford 06 Oct '14

06 Oct '14

> > Speaker: Dr. Wei Wang > Beijing Jiaotong University > Title: Automated Detection and Analysis of Malicious Android Apps > Date: Tuesday, October 7, 2014 > Time: 12:00 p.m. > Room: E-Quad, Room B327 > Host: Prof. Ruby Lee > > > Abstract: Android has been a major target of malicious applications (malapps). How to automate the detection and analysis of malapps is an important issue. We will present three parts of our recent work. First, we explore the permission-induced risk in Android apps in a systematic manner. Second, we extract as many as possible static features from Android apps for the detection and analysis of malapps. Third, we extract structural features from apps for the analysis. The preliminary results show the automated process of the detection and analysis of malicious apps. > > Biography: Dr. Wei Wang is currently associate professor in the Department of Information SecuriBeijing Jiaotong University, China. He earned his Ph.D. degree from Xi’an Jiaotong University, in 2006. He was a postdoctoral researcher in University of Trento, Italy, during 2005-2006. He was a postdoctoral researcher in TELECOM Bretagne and in INRIA, France, during 2007-2008. He was a European ERCIM Fellow in Norwegian University of Science and Technology (NTNU), Norway, and in Interdisciplinary Centre for Security, Reliability and Trust (SnT), University of Luxembourg, during 2009-2011. He visited INRIA, ETH, NTNU, CNR, and New York University Polytechnic. He has authored or co-authored over 40 peer-reviewed papers in various journals and international conferences. His main research interests include mobile, computer and network security. >

1 0

Update: Wenhao Jia (ELE) presents FPO "Analysis and Optimization Techniques for Massively Parallel Processors" Tuesday, October 7 at 9am, Equad Room J401
by Nicki Gotsis 01 Oct '14

01 Oct '14

Wenhao Jia FPO Date and Location: Tuesday, October 7 at 9am, Equad Room J401 Thesis Committee: Co-advisors: Margaret Martonosi and Kelly Shaw; Reader: Niraj Jha; Non-readers: David August and Sharad Malik Thesis Title: Analysis and Optimization Techniques for Massively Parallel Processors Thesis Abstract: Heterogeneous parallelism has emerged as a widespread computing paradigm in the past decade or so. In particular, massively parallel processors such as graphics processing units (GPUs) have become the prevalent throughput computing element in heterogeneous systems. However, GPUs are difficult to program and design, primarily due to developers' general unfamiliarity with GPUs, a lack of hardware resource virtualization, and inadequate tailoring of conventionally designed components--such as general-purpose caches--to the massively parallel nature of GPU workloads. To overcome these challenges, this thesis proposes statistical analysis techniques and software and hardware optimizations that improve the performance, power efficiency, and programmability of GPUs. These proposals make it easier for programmers and designers to produce optimized GPU software and hardware designs. The first part of the thesis describes how statistical analysis can help users explore a GPU software or hardware design space with performance or power as the metric of interest. In particular, two fully automated tools--linear regression-based Stargazer and regression tree-based Starchart--are developed and presented. They help identify important design parameters and automate design decisions, while saving design exploration time by 300-3000 times compared to exhaustive approaches. Then, the second part of the thesis proposes two compile-time algorithms to automatically make two key GPU software design decisions: cache configuration and thread block size selection. The first algorithm uses instruction-level memory access pattern analysis and cache control to improve the performance benefits of caching from 5.8% to 18%. The second algorithm runs programs with thread counts estimated to trigger performance saturation, reducing GPU core resource usage by 27-62% while improvin g performance by 5-10%. Finally, the third part of the thesis proposes and evaluates the memory request prioritization buffer (MRPB). MRPB uses request reordering to reduce cache thrashing and uses cache bypassing to reduce resource stalls, improving performance by 1.3-2.7 times and easing GPU programming. In summary, using GPUs as an example, the high-level statistical tools and the more focused software and hardware studies presented in this thesis demonstrate how to use automation techniques to effectively improve the performance, power efficiency, and programmability of emerging heterogeneous computing platforms. _______________________________________________ talks mailing list talks(a)lists.cs.princeton.edu To edit subscription settings or remove yourself, use this link: https://lists.cs.princeton.edu/mailman/listinfo/talks

1 0