- talks - lists.cs.princeton.edu

Malte Moeser General Exam Presentation - May 16, 2018 at 5:00pm - CS402
by Barbara A. Mooring 14 May '18

14 May '18

Malte Moeser General Exam Presentation - May 16, 2018 at 5:00pm - CS402 Title: Anonymous Alone? On the Challenge of Anonymity in Cryptographic Currencies Abstract: Cryptographic currencies balance two seemingly contradictory goals: radical transparency of the ledger and anonymity of users. A variety of privacy-enhancing techniques have been developed to strengthen users’ anonymity using obfuscation and cryptography, yet to this day their adoption remains limited. A closer look at existing deployments shows how in practice these techniques often fall short of their promises due to inherent complexity and dependence on user behavior. One such example is Monero, a privacy-centric cryptocurrency that allows users to obfuscate the true origin of the coin they are spending by including other chaff coins. Analyzing the Monero blockchain, we found several significant weaknesses in it’s design that allowed to retroactively deanonymize users’ transactions. This is one further example that cryptocurrency privacy is hard: the data is necessarily public forever, and often a user’s actions can negatively influence the anonymity of others. Based on the insights from the analysis of Monero and other privacy-preserving techniques, I’ll put forward a set of recommendations for protocol designers and discuss options for policy makers to regulate cryptographic currencies based on their inherent linkability. Barbara A. Mooring Interim Graduate Coordinator Computer Science Department Princeton University

1 0

Lauren Pick General Exam Presentation - Thursday, May 10, 2018 at 3:30 pm - CS401
by Barbara A. Mooring 10 May '18

10 May '18

Lauren Pick General Exam Presentation - Thursday, May 10, 2018 at 3:30 pm - CS401 Title: Exploiting Synchrony and Symmetry in Relational Verification Abstract: Relational safety specifications describe multiple runs of the same program or relate the behaviors of multiple programs. Approaches to automatic relational verification often compose the programs and analyze the composition for safety, but a naively composed program can lead to difficult verification problems. In my talk, I will demonstrate two ways of exploiting relational specifications to simplify verification subtasks generated during relational verification. First, I will describe an algorithm for maximizing opportunities for synchronizing code fragments containing loops, allowing a verification procedure to avoid computing difficult loop invariants. Second, I will show how to compute symmetries in the specifications, allowing a verification procedure to avoid redundant subtasks. These enhancements have been implemented in a prototype for verifying k-safety properties on Java programs, and evaluation confirms that taking advantage of relational specifications as described leads to a consistent performance speedup on a range of benchmarks. Barbara A. Mooring Interim Graduate Coordinator Computer Science Department Princeton University

1 2

Kyle Genova - General Exam Presentaton, Thursday, May 10, 2018 at 2:30 pm - CS402
by Barbara A. Mooring 10 May '18

10 May '18

Kyle Genova - General Exam Presentaton, Thursday, May 10, 2018 at 2:30 pm - CS402 Title: Unsupervised Training for 3D Morphable Model Regression Abstract: We present a method for training a regression network from image pixels to 3D morphable model coordinates using only unlabeled photographs. The training loss is based on features from a facial recognition network, computed on-the-fly by rendering the predicted faces with a differentiable renderer. To make training from features feasible and avoid network fooling effects, we introduce three objectives: a batch distribution loss that encourages the output distribution to match the distribution of the morphable model, a loopback loss that ensures the network can correctly reinterpret its own output, and a multi-view identity loss that compares the features of the predicted 3D face and the input photograph from multiple viewing angles. We train a regression network using these objectives, a set of unlabeled photographs, and the morphable model itself, and demonstrate state-of-the-art results. Barbara A. Mooring Interim Graduate Coordinator Computer Science Department Princeton University

1 2

Robert MacDavid General Exam Presentation - Thursday, May 10, 2018 at 12:00 pm, CS302
by Barbara A. Mooring 10 May '18

10 May '18

Robert MacDavid General Exam Presentation - Thursday, May 10, 2018 at 12:00 pm, CS302 Title: Scaling Network Polices in the Fast Path with MACH-4 Abstract: As network speeds increase, operators desire to push applications like Firewalls, Network Address Translators, and Load Balancers into network hardware for better bandwidth, latency, and aggregation. Emerging programmable switches are accelerating this trend, allowing operators to define their own functions that can run entirely in the hardware fast path. Unfortunately, such modern switches are still severely limited by memory, only capable of storing hundreds of thousands of application rules, where modern applications require tens of millions. In this work we show that, by using intimate knowledge of the programmable switches, centralized applications can be split across multiple switches for little to no performance penalty, allowing applications to scale with the size of the network. We show how application caches in each switch can be updated without software intervention, greatly reducing latency and control-plane load as compared to other caching solutions. Barbara A. Mooring Interim Graduate Coordinator Computer Science Department Princeton University

1 2

Colloquium Speaker-Huacheng Yu, TODAY at 12:30
by Emily C. Lawrence 08 May '18

08 May '18

Huacheng Yu from Harvard University Tuesday, May 8, 2018 - 12:30pm Computer Science - Room 105 Host: Ran Raz "Better understanding of efficient dynamic data structures" Abstract: Data structures have applications and connections to algorithm design, database systems, streaming algorithms and other areas of computer science. Understanding what efficient data structures can do (and what they cannot do) is crucial to these applications. In this talk, I will present my work in analyzing efficient data structures and proving what they cannot accomplish. I will focus on the recent development in building new connections between dynamic data structures and communication complexity, as well as a new approach to analyzing dynamic data structures with Boolean outputs and super-logarithmic time. Bio: Huacheng Yu is a postdoctoral researcher in the Theory of Computing group at Harvard University. He obtained his PhD from Stanford University in 2017 under the supervision of Ryan Williams and Omer Reingold . He also holds a Bachelor's degree from Tsinghua University (2012). His primary research interests are data structure lower bounds. He also works in algorithm design and communication complexity.

1 0

Reminder: Colloquium Speaker-Huacheng Yu, tomorrow at 12:30
by Emily C. Lawrence 07 May '18

07 May '18

Huacheng Yu from Harvard University Tuesday, May 8, 2018 - 12:30pm Computer Science - Room 105 Host: Ran Raz "Better understanding of efficient dynamic data structures" Abstract: Data structures have applications and connections to algorithm design, database systems, streaming algorithms and other areas of computer science. Understanding what efficient data structures can do (and what they cannot do) is crucial to these applications. In this talk, I will present my work in analyzing efficient data structures and proving what they cannot accomplish. I will focus on the recent development in building new connections between dynamic data structures and communication complexity, as well as a new approach to analyzing dynamic data structures with Boolean outputs and super-logarithmic time. Bio: Huacheng Yu is a postdoctoral researcher in the Theory of Computing group at Harvard University. He obtained his PhD from Stanford University in 2017 under the supervision of Ryan Williams and Omer Reingold . He also holds a Bachelor's degree from Tsinghua University (2012). His primary research interests are data structure lower bounds. He also works in algorithm design and communication complexity.

1 0

May 11th Symposium - Frontiers in Quantitative Cancer Biology
by Emily Lawrence 02 May '18

02 May '18

Please join us on Friday, May 11th for the Frontiers in Quantitative Cancer Biology Symposium <https://lsi.princeton.edu/news-seminars/seminars/frontiers-quantitative-can cer-biology> . Talks will take place in the Icahn Laboratory Auditorium (Room 101) followed by a reception in the Icahn Atrium. Symposium Schedule: 9:00 - 9:30 a.m. Coffee and check-in (outside of Icahn 101) 9:30 - 10:45 a.m. Sean Morrison <https://cri.utsw.edu/scientists/sean-morrison-laboratory/> : "Metabolic regulation of melanoma metastasis" (With a short talk by Professor Josh Rabinowitz) 10:45 - 11:15 a.m. Break 11:15 a.m. - 12:30 p.m. Dana Pe <https://www.mskcc.org/research-areas/labs/dana-pe-er> 'er: "A single cell view into metastasis" (With a short talk by Professor Mona Singh) 12:30 - 2:00 p.m. Lunch break (independent, not included in symposium) 2:00 - 3:15 p.m. David Tuveson <http://tuvesonlab.labsites.cshl.edu/> : "Predictive Models for the Treatment of Pancreatic Cancer" (With a short talk by Professor Ben Raphael) 3:15 - 3:45 p.m. Break 3:45 - 4:45 p.m. Arlene Sharpe <https://sharpelab.hms.harvard.edu/> : "Biology of PD-1 checkpoint blockade" Reception to follow in the Icahn Laboratory Atrium No registration is required. For more information, contact Sara Thibeault at thibeault(a)princeton.edu <mailto:thibeault@princeton.edu> or 609-258-3930

1 0

Colloquium Speaker-Huacheng Yu, Tuesday, May 8
by Emily Lawrence 02 May '18

02 May '18

Huacheng Yu from Harvard University Tuesday, May 8, 2018 - 12:30pm Computer Science - Room 105 Host: Ran Raz "Better understanding of efficient dynamic data structures" Abstract: Data structures have applications and connections to algorithm design, database systems, streaming algorithms and other areas of computer science. Understanding what efficient data structures can do (and what they cannot do) is crucial to these applications. In this talk, I will present my work in analyzing efficient data structures and proving what they cannot accomplish. I will focus on the recent development in building new connections between dynamic data structures and communication complexity, as well as a new approach to analyzing dynamic data structures with Boolean outputs and super-logarithmic time. Bio: Huacheng Yu is a postdoctoral researcher in the Theory of Computing group at Harvard University. He obtained his PhD from Stanford University in 2017 under the supervision of Ryan Williams and Omer Reingold. He also holds a Bachelor's degree from Tsinghua University (2012). His primary research interests are data structure lower bounds. He also works in algorithm design and communication complexity.

1 0

Colloquium Speaker: Ben Lee, TODAY at 4:30pm
by Emily C. Lawrence 01 May '18

01 May '18

Ben Lee from Duke University Tuesday, May 1, 2018 at 4:30 p.m. Maeder Hall auditorium Title: Datacenters and Energy Efficiency: A Game-Theoretic Perspective Abstract: Sharing datacenter hardware improves energy efficiency, but whether strategic users participate in consolidated systems depends on management policies. Users who dislike allocations may refuse to participate and deploy private, less-efficient systems. We rethink systems management, drawing on game theory to model strategic behavior and incentivize participation. We illustrate this perspective for two fundamental challenges in datacenters. For power delivery, we design sprinting games to produce equilibria in which users selfishly draw power for performance boosts yet avoid oversubscribing the shared supply. For resource allocation, we use Cobb-Douglas utility functions to produce fair allocations that incentivize users to share cache and memory. These solutions provide foundations for rigorously managing systems shared by strategic, competitive participants. Biography: Benjamin Lee is an Associate Professor of Electrical and Computer Engineering at Duke University. He received his B.S. from the University of California at Berkeley (2004), his Ph.D. from Harvard University (2008), and his post-doctorate from Stanford University (2010). He has held visiting positions at Microsoft Research, Intel Labs, and Lawrence Livermore National Lab. Dr. Lee’s research interests include computer architecture, energy efficiency, and security / privacy. He pursues these interests by building interdisciplinary links to statistical inference and algorithmic economics. His research has been recognized by IEEE Micro Top Picks (4x), Communications of the ACM Research Highlights (3x), as well as paper honors from the ASPLOS, HPCA, MICRO, and SC conferences. He received the NSF Computing Innovation Fellowship, NSF CAREER Award, and Google Faculty Research Award.

1 0

Reminder: Ben Lee, tomorrow at 4:30pm
by Emily C. Lawrence 30 Apr '18

30 Apr '18

Ben Lee from Duke University Tuesday, May 1, 2018 at 4:30 p.m. Maeder Hall auditorium Title: Datacenters and Energy Efficiency: A Game-Theoretic Perspective Abstract: Sharing datacenter hardware improves energy efficiency, but whether strategic users participate in consolidated systems depends on management policies. Users who dislike allocations may refuse to participate and deploy private, less-efficient systems. We rethink systems management, drawing on game theory to model strategic behavior and incentivize participation. We illustrate this perspective for two fundamental challenges in datacenters. For power delivery, we design sprinting games to produce equilibria in which users selfishly draw power for performance boosts yet avoid oversubscribing the shared supply. For resource allocation, we use Cobb-Douglas utility functions to produce fair allocations that incentivize users to share cache and memory. These solutions provide foundations for rigorously managing systems shared by strategic, competitive participants. Biography: Benjamin Lee is an Associate Professor of Electrical and Computer Engineering at Duke University. He received his B.S. from the University of California at Berkeley (2004), his Ph.D. from Harvard University (2008), and his post-doctorate from Stanford University (2010). He has held visiting positions at Microsoft Research, Intel Labs, and Lawrence Livermore National Lab. Dr. Lee’s research interests include computer architecture, energy efficiency, and security / privacy. He pursues these interests by building interdisciplinary links to statistical inference and algorithmic economics. His research has been recognized by IEEE Micro Top Picks (4x), Communications of the ACM Research Highlights (3x), as well as paper honors from the ASPLOS, HPCA, MICRO, and SC conferences. He received the NSF Computing Innovation Fellowship, NSF CAREER Award, and Google Faculty Research Award.

1 0