- talks - lists.cs.princeton.edu

R Dannenfelser MSE thesis talk
by Melissa M. Lawson 21 Apr '14

21 Apr '14

Ruth Dannenfelser will present her MSE thesis talk on Thursday April 24 at 3PM in room 301 (note room!). The members of her committee are Olga Troyanskaya, advisor and Kai Li, reader. Everyone is invited to attend her talk. Her abstract follows below. ----- Original Message ----- The immune system is a complex set of biological processes and structures that work together to protect the human body. A main component of this system are the lymphocytes, white blood cells, which are responsible for identifying and destroying invading cells, including tumors. In breast cancer it has been shown that the level of the immune infiltrate in the tumors is one of the key determinants of prognosis and survival. However, a more detailed view of the relative abundances of specific lymphocytes in these tumors and their functional role in breast cancer remains largely unexplored. Here we attempt to quantify the level of immune activity in patient tumors by examining the expression of four specific lymphocytes: B cells, cytotoxic T, T helper 1, and T helper 2 cells. We utilize a machine learning approach, in silico nano-dissection, to expand small lists of immune specific markers into larger cell type specific signatures. We then use these signatures to profile breast tumors across different clinical groupings revealing consistent variations in lymphocyte expression across subtypes, morphologies, and treatments.

1 0

Fwd: Keith Devlin, April 21, public lecture
by Jennifer Rexford 17 Apr '14

17 Apr '14

> Please join the  > Princeton University  > Council on Science and Technology, Mathematics Department, & McGraw Center for Teaching  > and Learning > > FOR A PUBLIC LECTURE WITH > > Keith Devlin,  > Stanley Kelley, Jr. Visiting Professor for Distinguished Teaching, Department of Mathematics > > DISCUSSING  > Leonardo and Steve:  > The Young Genius Who Beat Apple to Market by  > 800 Years > EVENT DETAILS  > Date: April 21st  > Time: 6:00pm  > Place: Computer Science 104 > >

1 0

Colloquium Speaker Nick Feamster, Thurs, April 17, 4:30pm
by Nicole E. Wagenblast 17 Apr '14

17 Apr '14

Bots, Bubbles, and Bottlenecks: Safeguarding the User’s Internet Experience Nick Feamster, Georgia Tech Thursday, April 17, 4:30pm Computer Science 105 A user's experience on the Internet rests in the hands of a large and increasingly diverse set of stakeholders. Internet service providers and content providers point fingers at one other about performance problems. Miscreants launch attacks against both other users and the Internet infrastructure itself. Content providers continually engage in practices to "personalize" what we see and when we see it. Many governments aim to control its citizens' access to information, while activists design circumvention tools. Safeguarding the user's Internet experience requires both gathering empirical network measurements to detect threats (typically in the absence of any "ground truth") and developing large-scale systems to mitigate them. In this talk, I will present three classes of safeguards against different threats to the user's Internet experience: (1) technologies to characterize and improve performance in the Internet's "last mile", including a worldwide deployment of home routers in around 200 home networks and ongoing studies with the Federal Communications Commission; (2) methods for lightweight and fast detection of message abuse, such as spam, that have since been widely adopted by industry; and (3) defenses against against information manipulation attacks, a new class of attacks against personalization algorithms. I will also discuss other such threats and how networking can draw from other disciplines to tackle them. Nick Feamster is an associate professor in the College of Computing at Georgia Tech. He received his Ph.D. in Computer science from MIT in 2005, and his S.B. and M.Eng. degrees in Electrical Engineering and Computer Science from MIT in 2000 and 2001, respectively. His research focuses on many aspects of computer networking and networked systems, with a focus on network operations, network security, and censorship-resistant communication systems. In December 2008, he received the Presidential Early Career Award for Scientists and Engineers (PECASE) for his contributions to cybersecurity, notably spam filtering. His honors include the Technology Review 35 "Top Young Innovators Under 35" award, the ACM SIGCOMM Rising Star Award, a Sloan Research Fellowship, the NSF CAREER award, the IBM Faculty Fellowship, the IRTF Applied Networking Research Prize, and award papers at the SIGCOMM Internet Measurement Conference (measuring Web performance bottlenecks), SIGCOMM (network-level behavior of spammers), the NSDI conference (fault detection in router configuration), Usenix Security (circumventing web censorship using Infranet), and Usenix Security (web cookie analysis).

1 1

Steven Salzberg special seminar today at 4:00 pm
by Nicole E. Wagenblast 15 Apr '14

15 Apr '14

Steven Salzberg, Johns Hopkins University School of Medicine, Computational Challenges of High Throughput Genome Sequence Analysis Apr 15, 2014, 4:00 p.m. Lewis Thomas Lab 003 Next-generation sequencing technology allows us to peer inside the cell in exquisite detail, revealing new insights into biology, evolution, and disease that would have been impossible to discover just a few years ago. The enormous volumes of data produced by NGS experiments present many computational challenges that we are working to address. In this talk, I will discuss some of our algorithmic solutions to two key alignment problems: (1) mapping sequences onto the human genome at very high speed, and (2) mapping and assembling transcripts from RNA-seq experiments. I will also discuss some of the problems that can arise during analysis of exome data, in which the gene-containing portions of the genome are sequenced in an effort to identify mutations responsible for disease. My group has developed algorithms to solve each of these problems, including the widely-used Bowtie program for fast DNA sequence alignment, the TopHat and Cufflinks programs for assembly of genes from transcriptome sequencing (RNA-seq) experiments, and the new DIAMUND program for detecting de novo mutations. This talk describes joint work with current and former lab members including Ben Langmead, Cole Trapnell, Daehwan Kim, Mihaela Pertea, and Geo Pertea. http://ccb.jhu.edu/people/salzberg/Salzberg/Salzberg_Lab_Home.html

1 0

Colloquium Speaker Dafna Shahaf Tues April 15, 4:30pm
by Nicole E. Wagenblast 15 Apr '14

15 Apr '14

The Aha! Moment: From Data to Insight Dafna Shahaf, Stanford University Tuesday, April 15, 4:30pm Computer Science 105 The amount of data in the world is increasing at incredible rates. Large-scale data has potential to transform almost every aspect of our world, from science to business; for this potential to be realized, we must turn data into insight. In this talk, I will describe two of my efforts to address this problem computationally: The first project, Metro Maps of Information, aims to help people understand the underlying structure of complex topics, such as news stories or research areas. Metro Maps are structured summaries that can help us understand the information landscape, connect the dots between pieces of information, and uncover the big picture. The second project proposes a framework for automatic discovery of insightful connections in data. In particular, we focus on identifying gaps in medical knowledge: our system recommends directions of research that are both novel and promising. I will formulate both problems mathematically and provide efficient, scalable methods for solving them. User studies on real-world datasets demonstrate that our methods help users acquire insight efficiently across multiple domains. Dafna Shahaf is a postdoctoral fellow at Stanford University. She received her Ph.D. from Carnegie Mellon University; prior to that, she earned an M.S. from the University of Illinois at Urbana-Champaign and a B.Sc. from Tel-Aviv university. Dafna 's research focuses on helping people make sense of massive amounts of data. She has won a best research paper award at KDD 2010, a Microsoft Research Fellowship, a Siebel Scholarship, and a Magic Grant for innovative ideas.

1 2

P Gopalan preFPO
by Melissa M. Lawson 11 Apr '14

11 Apr '14

Prem Gopalan will present his preFPO on Friday April 18 at 3PM in Room 402. The members of his committee are: David Blei, advisor; Rob Schapire and Jake Hofman (MSR), readers; Michael Freedman and John Storey (MOL), nonreaders. Everyone is invited to attend his talk. His abstract follows below. ---------------- Title: Scalable inference of discrete outcomes: networks, genotype and user consumption Latent variable models are probabilistic models that can be used to extract hidden structure in real data. They are important in many fields such as genetics, social network analysis and collaborative filtering. Data analysis using these models is useful in making predictions, exploring the data and in making better models. Will inference algorithms be able to cope with the scale of modern data sets? If yes, what properties of the model, data and the algorithms help in achieving scalability? In this talk, I will present advances in statistical models and scalable inference algorithms for identifying overlapping communities in networks, ancestral populations in human genetic variations, and latent structure in user consumption data. These algorithms lie in the framework of variational inference, an approach to approximate posterior inference that has been adapted to a variety of probabilistic models. As a detailed example, I will present hierarchical Poisson matrix factorization models for recommendation, and a corresponding variational inference algorithm. The algorithm scales to more than 100 million observations on a single CPU and predicts better than prior methods. A simple extension to this model allows for cold-start recommendations. I will end with a novel Bayesian nonparametric variant of Poisson matrix factorization that eases the burden of searching for the best number of latent components. This talk includes ongoing work.

1 0

Stephen Yi-Hsien Lin MSE thesis talk
by Melissa M. Lawson 11 Apr '14

11 Apr '14

Stephen Yi-Hsien Lin will present his MSE thesis talk on Wednesday April 16 at 3PM in Room 401 (note room!). The members of his committee are: Andrew Appel, advisor and Ed Felten, reader. Everyone is invited to attend his talk. His abstract follows below. --------------------- ROS, a flexible framework for developing robot software, has been widely used to implement many robots. ROS is highly decoupled into different function nodes in order to encourage collaborative development. However, these nodes communicate through non-authenticated messages over an insecure communication channel, which makes the robots highly susceptible to cyber attacks. Adversary might be able to inject malicious commands into the communication channel in order to take control over the robot. Therefore, a secure message authentication mechanism must be introduced to ROS in order to prevent such attack. We developed a message authentication mechanism based on HMAC to address this problem. Our design goal is to not only protect the integrity and authentication of the messages in ROS but also a) bring as few intrusion to the system as possible, b) maintain backward compatibility so the code developed in the past could also benefit from our message authentication package with very little modification and c) make the usage as simple as possible. Besides the message authentication mechanism, we also proposed a key generation and exchange protocol on ROS. Our measurements comparing the throughput and latency before and after introducing message authentication indicated that the performance of the whole system is still reasonable after introducing the message authentication functionality.

1 0

Marcela Melara MSE thesis talk
by Melissa M. Lawson 10 Apr '14

10 Apr '14

Marcela Melara will present her MSE thesis talk on Thursday April 17 at 2PM in Room 402. Her committee is: Ed Felten, advisor and Michael Freedman, reader. Everyone is invited to attend her talk. Her abstract follows below. --------------------------- Abstract: As a foundation for securing Internet communication, today’s public-key infrastructures attempt to certify that the “correct” principal owns a name and possesses a corresponding public key. Yet these assertions are not machine-verifiable, and their correctness can be and has been broken by error, malice, or external (and sometimes legal) coercion. Rather than focus on the “correctness” of name ownership, we argue that ensuring the continuity of that ownership, as defined by the binding of a name to a public key, is necessary and often sufficient for secure communication. To this end, we have designed CONIKS, a key management system that ensures online identity providers maintain and respond with valid bindings from names to cryptographic keys, even without users managing (or seeing) those keys. If providers disseminate invalid bindings or equivocate by presenting different bindings to different users, parties quickly detect this misbehavior and can publish irrefutable proof of this equivocation. The cryptographic mechanisms employed by CONIKS are highly efficient, supporting an Internet environment of millions of providers, any of which may have users in the billions. Our prototype Off-the-Record chat service, ConiksChat, demonstrates that one can achieve high usability while preserving the security and privacy of online communications.

1 0

Colloquium Speaker Dorit Aharonov Thurs April 10, 4:30pm
by Nicole E. Wagenblast 10 Apr '14

10 Apr '14

Quantum computation as a lens on quantum physics Dorit Aharonov ( Hebrew University, Jerusalem, Israel ) Thursday, April 10, 4:30pm Computer Science 105 Quantum computation inspires the study of quantum many body systems from a computational perspective. This approach leads to a remarkably rich set of insights and questions, with deep implications to both physics and quantum computation. This direction, now coined "quantum Hamiltonian complexity", had turned over the past decade into an exciting fast growing field. I will try to give a taste of some of its main achievements, e.g., unexpected hardness of certain physical systems, and testing quantum mechanics using interactive proofs. Dorit Aharonov did a BSc in Physics and Mathematics at the Hebrew university, an MSc in Physics at the Weizmann institute, and a PhD in Computer Science and Physics at the Hebrew university (1999). After a postdoc at IAS Princeton and UC Berkeley she joined the faculty of the CS department at the Hebrew university in 2001. In 2005, Aharonov was profiled in Nature as one of four "young theorists who are making waves in their chosen fields", and in 2006 she received the Krill Prize for Excellence in Scientific Research. In 2011 she was awarded an ERC starting grant from the European Research Council. Her current topics of interest include quantum algorithms and complexity, and the computational view on quantum mechanics and multiparticle entanglement.

1 2

Nick Johnson Pre-FPO
by Melissa M. Lawson 10 Apr '14

10 Apr '14

Nick Johnson will present his preFPO on Thursday April 17 at 9AM in Room 402. The members of his committee are: David August, advisor; David Wentzlaff and Jae Lee (SKKU), readers; David Walker and Andrew Appel, nonreaders. Everyone is invited to attend his talk. His abstract follows below. --------- Title: Static Dependence Analysis in an Infrastructure for Automatic Parallelization. Now that parallel architectures are common, software must exploit parallel functional units to fully utilize hardware resources and achieve efficient execution. Although developers may restructure their applications for explicit parallelism, that process requires developers to reason about low-level details of the execution environment to avoid concurrency bugs and achieve parallel performance. A favorable alternative is automatic thread extraction, since it relieves the developer of the arduous task of (re-)developing applications with each new architecture. This dissertation presents compiler middle-ware to support automatic thread extraction---analyses and transformations that allow the compiler to deliver parallel performance from sequentially-specified input programs. Specifically, this thesis contributes, 1. Collaborative Dependence Analysis: Compilers use static analysis to determine facts about the input code. These facts restrict compiler transformations to ensure correctness. The collaborative dependence analysis framework combines simple analysis algorithms into an ensemble algorithm. Such ensembles are structured for collaboration: the ensemble algorithm disproves dependence queries which no member analysis algorithm disproves alone. Preliminary results demonstrate that this framework has a large impact on even speculative thread extraction systems: for some benchmarks, the strength of analysis is the difference between a 28x speedup and an 11x slowdown (geomean) measured on 50 cores. 2. Scaling Analysis to Parallelize Larger Scopes: As scopes grow large, the burden of analysis becomes prohibitive. Observing that parallelization depends on the strongly-connected components of the PDG (the DAG_SCC), this dissertation presents a faster algorithm to compute the DAG_SCC. This algorithm greedily identifies dependence edges which cannot affect parallelization. It reduces analysis time by skipping such unimportant dependence edges, instead spending analysis effort on the remaining, important queries. Overall, this faster algorithm reduces analysis time by half while maintaining precision. The dissertation presents an end-to-end integration of these contributions. This infrastructure has proved robust and flexible; many works from the Liberty Research Group have built upon this infrastructure.

1 0