September 2020 - talks - lists.cs.princeton.edu

Naorin Hossain will present her General Exam "TransForm: Formally Specifying Transistency Models and Synthesizing Enhanced Litmus Tests" on Thursday, September 24, 2020 at 4:30pm via Zoom
by Nicki Mahler 02 Oct '20

02 Oct '20

Naorin Hossain will present her General Exam "TransForm: Formally Specifying Transistency Models and Synthesizing Enhanced Litmus Tests" on Thursday, September 24, 2020 at 4:30pm via Zoom. Zoom address: [ https://princeton.zoom.us/j/96195237138 | https://princeton.zoom.us/j/96195237138 ] The members of her committee are as follows: Margaret Martonosi (Adviser), Aarti Gupta, and Amit Levy Link to Abstract and Reading list: [ https://docs.google.com/document/d/15INZKwkigiriz8vCc2ntx2833kFAkAiQT9viVGc… | https://docs.google.com/document/d/15INZKwkigiriz8vCc2ntx2833kFAkAiQT9viVGc… ] [ https://docs.google.com/document/d/1smKDgB4jcnIaS6lpGBaWcFgMW_Mxv6rm8-_4PVM… | https://docs.google.com/document/d/1smKDgB4jcnIaS6lpGBaWcFgMW_Mxv6rm8-_4PVM… ] Everyone is invited to attend the talk, and those faculty wishing to remain for the oral exam following are welcome to do so. Nicki Mahler Graduate Coordinator Computer Science Department Princeton University

1 1

CITP Special Event: Election Security and Transparency in 2020 - TONIGHT at 7pm ET
by Emily C. Lawrence 17 Sep '20

17 Sep '20

Center for Information Technology Policy Special Event: Election Security and Transparency in 2020 Date: Thursday, September 17, 2020 Time: 7:00 pm - 8:30 pm Event page: [ https://citp.princeton.edu/event/election-security-and-transparency/ | https://citp.princeton.edu/event/election-security-and-transparency/ ] This event is co-sponsored by the League of Women Voters of Berkeley Heights, New Providence, and Summit, New Jersey. Please register here: [ https://princeton.zoom.us/webinar/register/WN_g_huaPqzSqmGrfTTtKUxfQ | https://princeton.zoom.us/webinar/register/WN_g_huaPqzSqmGrfTTtKUxfQ ] “What can we as voters do to protect our elections and our representative government?”, asks the League of Women Voters of Berkeley Heights, New Providence, and Summit, New Jersey. Some of the answers include, be aware of election processes in use, and if those are inadequate, press for reforms. Be aware of election technology in use, and if that is insecure, press for reforms. Those processes and technology must ensure every citizen has the substantive opportunity to vote, and must count every vote (but just once!). Participate, as citizens, in running and observing elections. In the 21st century, new technology has led to new challenges in election security and transparency, but the 2020 pandemic has added special challenges of its own. Andrew Appel will offer his thoughts and insight on the 2020 election process. Bio: Andrew Appel is the Eugene Higgins Professor Computer Science (CS) at Princeton University, and served from 2009-2015 as Chair of Princeton’s CS department. His research is in software verification, computer security, programming languages and compilers, and technology policy. He received his A.B. summa cum laude in physics from Princeton University in 1981, and his Ph.D. in computer science from Carnegie Mellon University in 1985. Professor Appel has been editor in chief of ACM Transactions on Programming Languages and Systems and is a fellow of the ACM (Association for Computing Machinery). He has worked on fast N-body algorithms (1980s), Standard ML of New Jersey (1990s), Foundational Proof-Carrying Code (2000s), and the Verified Software Toolchain (2010-present).

1 0

Nick Giannarakis will present his FPO "An Intermediate Language for Network Verification" on Friday, 9/11/2020 at 3pm via Zoom.
by Nicki Mahler 04 Sep '20

04 Sep '20

Nick Giannarakis will present his FPO "An Intermediate Language for Network Verification" on Friday, 9/11/2020 at 3pm via Zoom. Link to Zoom: [ https://princeton.zoom.us/j/8048625570 | https://princeton.zoom.us/j/8048625570 ] The members of his committee are as follows: Examiners: David Walker (Adviser), Andrew Appel, and Jennifer Rexford; Readers: Aarti Gupta and Ratul Mahajan (University of Washington) A copy of his thesis, is available upon request. Please email ngotsis(a)cs.princeton if you would like a copy of the thesis. Everyone is invited to attend his talk. Abstract follows below: [ https://drive.google.com/file/d/1ejMK4RSAX0lF72EALcMxxPfjXZ4Hh-Mr/view?usp=… | https://drive.google.com/file/d/1ejMK4RSAX0lF72EALcMxxPfjXZ4Hh-Mr/view?usp=… ]

1 0

Cyril Zhang will present his FPO "Regret-Minimizing Algorithms Beyond Classical Optimization and Control" on Thursday, 9/10/2020 at 4:40pm via Zoom.
by Nicki Mahler 03 Sep '20

03 Sep '20

Cyril Zhang will present his FPO "Regret-Minimizing Algorithms Beyond Classical Optimization and Control" on Thursday, 9/10/2020 at 4:40pm via Zoom. Link to Zoom: [ https://princeton.zoom.us/j/95603512563 | https://princeton.zoom.us/j/95603512563 ] The members of his committee are as follows: Readers: Elad Hazan (Adviser) and Yoram Singer; Examiners: Karthik Narasimhan, Mark Braverman, and Sanjeev Arora A copy of his thesis, is available upon request. Please email [ mailto:ngotsis@cs.princeton | ngotsis(a)cs.princeton ] if you would like a copy of the thesis. Everyone is invited to attend his talk. Abstract follows below. Abstract: Temporally correlated data and non-convex programs are two core challenges which lie at the frontier of reconciling theory and practice in machine learning. In this thesis, we present varied perspectives on algorithms and provable learning guarantees in stateful and non-convex environments, such as those encountered in reinforcement learning, control theory, time-series analysis, and deep learning. These ideas are unified by the framework of online convex optimization, which provides robust algorithmic primitives and regret guarantees under minimal assumptions on the data. The recurring theme of leveraging regret minimization beyond classical settings will manifest in several ways, eclectic in their scope. First, we develop a solution concept and ecient algorithms for online non-convex optimization in its most generic form. Next, in a line of theoretical work on prediction and control in the presence of linear time-invariant state transitions, beyond the linear-quadratic-Gaussian assumptions of classical control theory, we use online learning with convex relaxations to bypass the computational and statistical brittleness of the usual system identification pipelines. Finally, we use online convex optimization as a testbed for principled algorithm design, towards improving and better understanding ubiquitous training heuristics in deep learning.

1 0

Jordan Ash will present his FPO "Towards Flexible Active and Online Learning With Neural Networks" on Thursday, 9/10/2020 at 2:00pm via Zoom.
by Nicki Mahler 03 Sep '20

03 Sep '20

Jordan Ash will present his FPO "Towards Flexible Active and Online Learning With Neural Networks" on Thursday, 9/10/2020 at 2:00pm via Zoom. Link: [ https://princeton.zoom.us/j/7443698693?pwd=TXBSektKRVZ4ODhoeWpwVWtOMHVEQT09 | https://princeton.zoom.us/j/7443698693?pwd=TXBSektKRVZ4ODhoeWpwVWtOMHVEQT09 ] Meeting ID: 744 369 8693 Passcode: jordan2020 The members of his committee are as follows: Readers: Ryan Adams (Adviser) and Barbara Engelhardt; Examiners: Szymon Rusinkiewicz, Robert Schapire, and Akshay Krishnamurthy (Microsoft Research). A copy of his thesis, is available upon request. Please email [ mailto:ngotsis@cs.princeton | ngotsis(a)cs.princeton ] if you would like a copy of the thesis. Everyone is invited to attend his talk. Abstract follows below. Abstract: Deep learning has elicited breakthrough successes on a wide array of machine learning tasks. Outside of the fully-supervised regime, however, many deep learning algorithms are brittle and unable to reliably perform across model architectures, dataset types, and optimization parameters. As a consequence, these algorithms are not easily usable by non-machine-learning experts, limiting their ability to meaningfully impact science and society. This thesis addresses some nuanced pathologies around the use of deep learning for active and passive online learning. We propose a practical active learning approach for neural networks that is robust to environmental variables: Batch Active learning by Diverse Gradient Embeddings (BADGE). We also discuss the deleterious generalization effects of warm-starting the optimization of neural networks in sequential environments, and why this is a major problem for deep learning. We introduce a simple method that remedies this problem, and discuss some important ramifications of its application.

1 0