- talks - lists.cs.princeton.edu

Zoe Ashwood will present her Pre FPO, "Probabilistic models for characterizing animal learning and decision-making" on Friday, April 1, 2022 at 9am via Zoom.
by Nicki Mahler 24 Mar '22

24 Mar '22

Zoe Ashwood will present her Pre FPO, "Probabilistic models for characterizing animal learning and decision-making" on Friday, April 1, 2022 at 9am via Zoom. Zoom link: [ https://princeton.zoom.us/j/98847197070?pwd=U3ZiMFZDbnlYQ1IrYmZkRjQ0VHR6dz09 | https://princeton.zoom.us/j/98847197070?pwd=U3ZiMFZDbnlYQ1IrYmZkRjQ0VHR6dz09 ] The members of Zoe's committee are as follows: Readers: Alex Pouget (University of Geneva), Barbara Engelhardt Non-readers: Jonathan Pillow (advisor), Ryan Adams, Tom Griffiths Abstract: A central goal in neuroscience is to identify the strategies used by animals and humans as they make decisions, and to characterize the learning rules that shape these policies in the first place. In this talk, I will first discuss recent modeling work that revealed that — contrary to common wisdom — mice and humans use multiple strategies over the course of a single session to perform perceptual decision-making tasks. Our modeling framework, based on hidden Markov models with Bernoulli Generalized Linear Model observations, allows us to characterize the nature of these policies and to label the trials associated with each strategy. Next, I will discuss work that sought to uncover the learning rules used by mice and rats as they learned to perform this type of task. Our model tracked trial-to-trial changes in the animals’ choice policies, and separated these changes into components explainable by a reinforcement learning rule, and components that remained unexplained. Whereas the average contribution of the conventional REINFORCE learning rule to the policy update for mice learning a common task was just 30%, we found that adding baseline parameters allowed the learning rule to explain 92% of the animals' policy updates under our model. Finally, I will discuss work in progress that applies inverse reinforcement learning to the trajectories of mice exploring a maze environment. Unlike artificial agents, mice are very good at exploring novel environments, and in this work, we will seek to uncover the extrinsic and intrinsic reward functions that power such exploration.

1 0

ECE SEMINAR: Maria Bauza Villalonga, MIT, TODAY at 4:30pm
by Emily C. Lawrence 23 Mar '22

23 Mar '22

ECE SEMINAR Speaker: Maria Bauza Villalonga, MIT Title: Learning to interact with the world: when generality meets precision Day: March 23, 2022 Time: 4:30 pm Location: B205 Engineering Quadrangle Host: Jaime Fernández Fisac Abstract: Robotics stands as one of the most impactful and promising endeavors of our times. Learning to interact with the world is fundamental for solving some of our most pressing societal challenges: from taking care of our aging population and aiding with labor-intensive jobs to assisting in climate-related disasters and rescue emergencies. In this talk, I will argue that such a level of autonomy and performance requires robots that can excel across diverse settings while remaining accurate and reliable. My talk will focus on how we can develop learning algorithms that foster robotic generalization while ensuring the desired task performance. First, I will present a learning approach to pose estimation for novel objects based on visuo-tactile sensing that doesn't rely on real data and results in accurate pose distributions. Then, I will demonstrate how this approach enables precise robotic pick-and-place using task-aware grasping. The robotic system reasons over the models for grasping, planning, and perception in order to optimize its actions based only on simulated data. In real experiments, we demonstrate that our approach learned purely in simulation, allows robots to successfully manipulate new objects and perform highly accurate placements. Bio: Maria Bauza Villalonga is a PhD student in Robotics at the Massachusetts Institute of Technology, working with Professor Alberto Rodriguez. Before that, she received Bachelor's degrees in Mathematics and Physics from CFIS, an excellence center at the Polytechnic University of Catalonia. Her research focuses on achieving precise robotic generalization by learning probabilistic models of the world that allow robots to reuse their skills across multiple tasks and environments. Maria has received several fellowships, including Facebook, NVIDIA, and LaCaixa fellowships. Her research has obtained awards such as Best Paper Finalist in Service Robotics at ICRA 2021, Best Cognitive Paper award at IROS 2018, and Best Paper award finalist at IROS 2016. She was also part of the MIT-Princeton Team participating in the Amazon Robotics Challenge, winning the stowing task in 2017 and receiving the 2018 Amazon Best Systems Paper Award in Manipulation.

1 0

Fwd: James Baldwin Lecture (3/28) [POSTPONED]
by Elizabeth Wang 22 Mar '22

22 Mar '22

Greetings, Please see below that the James Baldwin Lecture will be postponed. It has not yet been rescheduled. Best, Elizabeth From: "Anthony K Gibbons Jr." <akg2(a)princeton.edu> To: "Elizabeth Wang" <ew0001(a)cs.princeton.edu> Sent: Tuesday, March 22, 2022 10:17:18 AM Subject: James Baldwin Lecture (3/28) [POSTPONED] Hello Elizabeth , It is with the deepest regret that we announce this year’s [ https://aas.princeton.edu/events/2022/james-baldwin-lecture-series-limits-q… | James Baldwin Lecture ] will be postponed for AY23. We extend our sincere apologies for any inconvenience this may cause, and will update you as soon as we can reschedule. Thank you for your continued support of African American Studies. Thank You, Anthony Gibbons Jr. Communications & Media Specialist Department of African American Studies Princeton University 003 Morrison Hall, Princeton, NJ 08544 [ https://aas.princeton.edu/ | aas.princteon.edu ]

1 0

CS Colloquium speaker Sherry Wu: TODAY at 12:30pm
by Emily C. Lawrence 22 Mar '22

22 Mar '22

CS Colloquium Speaker Speaker: Sherry Tongshuang Wu, University of Washington Date: Tuesday, March 22, 2022 Time: 12:30pm EST Location: CS 105 Hosts: Andrés Monroy-Hernández & Adam Finkelstein Event page: [ https://www.cs.princeton.edu/events/26176 | https://www.cs.princeton.edu/events/26176 ] This talk will be live-streamed at [ https://mediacentrallive.princeton.edu/ | https://mediacentrallive.princeton.edu/ ] Title: Interactive AI Model Debugging and Correction Abstract: Research in Artificial Intelligence (AI) has advanced at an incredible pace, to the point where it is making its way into our everyday lives, explicitly and behind the scenes. However, beneath their impressive progress, many AI models hide deficiencies that amplify social biases or even cause fatal accidents. How do we identify, improve, and cope with imperfect models, while still benefiting from their use? I will discuss my work empowering humans to interact with AI models in order to debug and correct them. I will describe both (1) how I help experts run scalable and testable analyses on models in development, and (2) how I help end users collaborate with deployed AI in a transparent and controllable way. In my final remarks, I will discuss my future research perspectives on building human-centered AI through data-centric approaches. Bio: Sherry Tongshuang Wu is a final year Ph.D. Candidate in Computer Science & Engineering at the University of Washington, advised by Jeffrey Heer and Dan Weld. She received her B.Eng in CSE from the Hong Kong University of Science and Technology. Her research lies at the intersection of Human-Computer Interaction (HCI) and Natural Language Processing (NLP), and aims to empower humans to debug and correct AI models interactively, both when the model is under active development, and after it is deployed for end users. Sherry has authored 19 papers in top-tier NLP, HCI and Visualization conferences and journals such as ACL, CHI, TOCHI, TVCG, etc, including a best paper award (top-1) and an honorable mention (top-3). You can find out more about her at [ https://homes.cs.washington.edu/~wtshuang/ | https://homes.cs.washington.edu/~wtshuang/ ] .

1 0

ECE SEMINAR: Shuwen Deng, Yale University TODAY at 4:30pm
by Emily C. Lawrence 21 Mar '22

21 Mar '22

ECE SEMINAR Speaker: Shuwen Deng, Yale University Title: Processor Microarchitecture Security Day: March 21, 2022 Time: 4:30pm Location: B205 Engineering Quadrangle Host: David Wentzlaff Abstract: As computer systems grow more and more complicated, various performance optimizations can unintentionally introduce security vulnerabilities in these systems. The vulnerabilities can lead to user information and data being compromised or stolen. Many of the computer processor optimizations often focus on sharing or re-using the processor hardware between different users or programs. This can lead to different types of timing-based security attacks where the sharing or re-using of hardware components influences the timing of the operations performed on the processor. Different timing can be exploited by malicious users to gain information about operations or data of the co-located victim program or user. To address the security threats, this talk will present research on developing a theoretical model for cache and TLB timing attacks to help evaluate the susceptibility of different architectures and microarchitectures to timing attacks on caches and TLBs. Further, it will present a benchmark suite built upon the model to test if processor caches, or secure cache designs, are vulnerable to the theoretical vulnerabilities found using the theoretical model. The talk will also present defenses for the TLBs which were developed to address the newly found vulnerabilities. Going beyond caches and TLBs, the presentation will also discuss research on, and evaluation of, timing and power-based attacks in the processor frontend. Bio: Shuwen Deng is a final-year Ph.D. candidate at Yale University, advised by Prof. Jakub Szefer. Her research interests span computer architecture and security. Specifically, she focuses her research on developing secure processor microarchitectures by proposing side-channel vulnerability checking schemes and combining the theoretical modeling with practical benchmarking analysis. She also researches new microarchitectural vulnerabilities, including those of value predictors and processor frontend. She is the recipient of the 2020 Google PhD Fellowship in Privacy and Security, and an Elihu Elias Dickerman Fellowship. She has been named a Rising Star in EECS by the University of California Berkeley. Her work has been selected as Top Picks 2021 in Hardware and Embedded Security.

1 0

CS Colloquium speaker Xinyun Chen: TODAY at 12:30pm
by Emily C. Lawrence 21 Mar '22

21 Mar '22

CS Colloquium Speaker Speaker: Xinyun Chen, University of California, Berkeley Date: Monday, March 21, 2022 Time: 12:30pm EST Location: CS 105 Host: Jia Deng Event page: https://www.cs.princeton.edu/events/26175 This talk will be live-streamed at https://mediacentrallive.princeton.edu/ Title: Learning-Based Program Synthesis: Learning for Program Synthesis and Program Synthesis for Learning Abstract: With the advancement of modern technologies, programming becomes ubiquitous not only among professional software developers, but also for general computer users. However, gaining programming expertise is time-consuming and challenging. Therefore, program synthesis has many applications, where the computer automatically synthesizes programs from specifications such as natural language descriptions and input-output examples. In this talk, I will present my work on learning-based program synthesis, where I have developed deep learning techniques for various program synthesis problems. Despite the remarkable success of deep neural networks for many domains, including natural language processing and computer vision, existing deep neural networks are still insufficient for handling challenging symbolic reasoning and generalization problems. My learning-based program synthesis research lies in two folds: (1) learning to synthesize programs from potentially ambiguous and complex specifications; and (2) neural-symbolic learning for language understanding. I will first talk about program synthesis applications, where my work demonstrates the applicability of learning-based program synthesizers for production usage. I will then present my work on neural-symbolic frameworks that integrate symbolic components into neural networks, which achieve better reasoning and generalization capabilities. In closing, I will discuss the challenges and opportunities of further improving the complexity and generalizability of learning-based program synthesis for future work. Bio: Xinyun Chen is a Ph.D. candidate at UC Berkeley, working with Prof. Dawn Song. Her research lies at the intersection of deep learning, programming languages, and security. Her recent research focuses on learning-based program synthesis and adversarial machine learning. She received the Facebook Fellowship in 2020, and Rising Stars in Machine Learning in 2021. Her work SpreadsheetCoder for spreadsheet formula prediction was integrated into Google Sheets, and she was part of the AlphaCode team when she interned at DeepMind.

1 0

CS Colloquium speaker Pang Wei Koh: TODAY at 12:30pm
by Emily C. Lawrence 17 Mar '22

17 Mar '22

Speaker: Pang Wei Koh, Stanford University Date: Thursday, March 17, 2022 Time: 12:30pm EST Location: CS 105 Host: Sanjeev Arora Event page: [ https://www.cs.princeton.edu/events/26174 | https://www.cs.princeton.edu/events/26174 ] This talk will be live-streamed at [ https://mediacentrallive.princeton.edu/ | https://mediacentrallive.princeton.edu/ ] Title: Reliable machine learning in the wild Abstract: Machine learning systems are widely deployed today, but they are unreliable. They can fail – and with catastrophic consequences – on subpopulations of the data, such as particular demographic groups, or when deployed in different environments from what they were trained on. In this talk, I will describe our work towards building reliable machine learning systems that are robust to these failures. First, I will show how we can use influence functions to understand the predictions and failures of existing models through the lens of their training data. Second, I will discuss the use of distributionally robust optimization to train models that perform well across all subpopulations. Third, I will describe WILDS – a benchmark of in-the-wild distribution shifts spanning applications such as pathology, conservation, remote sensing, and drug discovery – and show how current state-of-the-art methods, which perform well on synthetic distribution shifts, still fail to be robust on these real-world shifts. Finally, I will describe our work on building more reliable COVID-19 models, using anonymized cellphone mobility data, to inform public health policy; this is a challenging application as the underlying environment is often changing and there is substantial heterogeneity across demographic subpopulations. Bio: Pang Wei Koh is a PhD student at Stanford, advised by Percy Liang. He studies the theory and practice of building reliable machine learning systems. His research has been published in Nature and Cell, featured in media outlets such as The New York Times and The Washington Post, and recognized by best paper awards at ICML and KDD, a Meta Research PhD fellowship, and the Kennedy Prize for best honors thesis at Stanford. Prior to his PhD, he was the 3rd employee and Director of Partnerships at Coursera.

1 0

Danqi Liao will present her MSE talk "Selective Feature Aggregation for Semi-supervised Temporal Action Localization" on March 22, 2022 at 11:30am in Friend 125 and Zoom.
by Nicki Mahler 16 Mar '22

16 Mar '22

Danqi Liao will present her MSE talk "Selective Feature Aggregation for Semi-supervised Temporal Action Localization" on March 22, 2022 at 11:30am in Friend 125 and Zoom. Zoom link: [ https://princeton.zoom.us/j/3846231080 | https://princeton.zoom.us/j/3846231080 ] The members of her committee are as follows: Olga Russakovsky (adviser) and Jia Deng (reader) All are welcome to attend. Abstract: Action detection in untrimmed video has been a long standing goal in computer vision. Recently, single-frame annotation has emerged as a promising direction that bridges the gap between the video-level weak-supervision and costly full supervision. In this talk, I will be discussing Selective Feature Aggregation module that selectively aggregates frame features. The module utilizes temporal contextual information, leading to performance boosts in single frame-supervised temporal action localization without complex heuristic-based system designs.

1 0

CS Colloquium speaker Gilbert Bernstein: TODAY at 12:30pm
by Emily C. Lawrence 15 Mar '22

15 Mar '22

Speaker: Gilbert Bernstein, University of California, Berkeley and MIT Date: Tuesday, March 15, 2022 Time: 12:30pm EST Location: CS 105 Host: Felix Heide Event page: [ https://www.cs.princeton.edu/events/26172 | https://www.cs.princeton.edu/events/26172 ] This talk will be live-streamed at [ https://mediacentrallive.princeton.edu/ | https://mediacentrallive.princeton.edu/ ] Title: High-Performance Languages for Visual Computing Abstract: Computer Graphics and Visual Computing problems challenge us with a near inexhaustible demand for more resolution and scale in order to simulate the climate, reconstruct 3d environments, train neural networks, and produce games & movies. Building such applications requires integrating disciplinary expertise (e.g. physics, numerical methods, geometry and custom hardware) into a single system. However, abstraction barriers are regularly discarded in the name of higher-performance, leading to code that must be written and maintained by super-experts—programmers who simultaneously possess deep knowledge of all relevant disciplines. Programming languages, especially Domain Specific Languages (DSLs) are perhaps the most promising approach to recovering a separation of concerns in such high-performance systems. In this talk, I will first describe my work on DSLs to enable parallel portability of physical simulation and optimization programs, including the use of relational algebra and automatic differentiation to structure these problem domains. Then I will discuss more recent work on “horizontal DSLs” designed to dig into specific sub-problems: maximizing utilization of novel hardware accelerators, formally verifying optimizations of tensor programs, and extending automatic differentiation to correctly handle discontinuities in inverse rendering and simulation problems. Bio: Gilbert Bernstein is a Postdoctoral Scholar at the University of California, Berkeley and MIT CSAIL, working with Professor Jonathan Ragan-Kelley. His research lies in Computer Graphics and Programming Languages, especially the design of high-performance domain specific languages for numeric computing applications such as physical simulation, optimization and inverse problems. His work spans the gamut from user interfaces, to differentiable programming, parallel-portability, and new hardware design languages. His work has been published at SIGGRAPH, POPL, PLDI, & OOPSLA, as well as being incorporated into products at Adobe, Autodesk, and Disney. He holds a Ph.D. in Computer Science from Stanford University, where he was advised by Pat Hanrahan.

1 0

CS Colloquium speaker: Kostis Kaffes TODAY at 12:30pm
by Emily C. Lawrence 14 Mar '22

14 Mar '22

Speaker: Kostis Kaffes, Stanford University Date: Monday, March 14, 2022 Time: 12:30pm EST Location: Zoom Webinar Host: Ravi Netravali Event page: [ https://www.cs.princeton.edu/events/26171 | https://www.cs.princeton.edu/events/26171 ] Please register here: [ https://princeton.zoom.us/webinar/register/WN_Nn4iOb-mRqSEJbGsbzg5uQ | https://princeton.zoom.us/webinar/register/WN_Nn4iOb-mRqSEJbGsbzg5uQ ] Title: Solving the Cloud Efficiency Crisis with Fast and Accessible Scheduling Abstract: Operating systems (OS) specialization is necessary as the one-size-fits-all approach of fundamental OS operations such as scheduling is incompatible with today's diverse application landscape. Such specialization can improve application performance and cloud platform efficiency by an order of magnitude or more. Towards this goal, I will first discuss Shinjuku, a specialized OS that supports an order of magnitude higher load and lower tail latency than state-of-the-art systems by enabling better scheduling. Shinjuku leverages hardware support for virtualization to preempt as often as every 5 microseconds and disproves the conventional wisdom that interrupts are incompatible with microsecond timescales. Then, I will present Syrup, a framework that enables everyday application developers to specify custom scheduling policies easily and safely deploy them across different layers of the stack over existing operating systems like Linux, bringing the benefits of specialized scheduling to everyone. For example, Syrup allowed us to implement policies that previously required specialized dataplanes in less than 20 lines of code and improve the performance of an in-memory database by 8x without needing any application modification. Bio: Kostis Kaffes is a final-year Ph.D. candidate in Electrical Engineering at Stanford University, advised by Christos Kozyrakis. He is broadly interested in computer systems, cloud computing, and scheduling. His thesis focuses on end-host, rack-scale, and cluster-scale scheduling for microsecond-scale tail latency with the goal of improving efficiency in the cloud. Recently, he has been looking for ways to make it easier to implement and deploy custom scheduling policies across different layers of the stack. Kostis's research has been supported by a Facebook Research Award and various scholarships and fellowships from Stanford, A.G. Leventis Foundation, and Gerondelis Foundation. Prior to Stanford, he received his undergraduate degree in Electrical and Computer Engineering from the National Technical University of Athens in Greece.

1 0