- talks - lists.cs.princeton.edu

ECE Seminar 2/15/2023 - Priyanka Raina, Stanford University
by Emily C. Lawrence 08 Feb '23

08 Feb '23

ECE SEMINAR Speaker: Priyanka Raina, Stanford University Title: Agile Design of Domain-Specific Accelerators and Compilers Day: February 15, 2023 Time: 4:30pm Location: B205 Engineering Quadrangle Host: Naveen Verma Abstract: With the slowing of Moore's law, computer architects have turned to domain-specific hardware accelerators to improve the performance and efficiency of computing systems. However, programming these systems entails significant modifications to the software stack to properly leverage the specialized hardware. Moreover, the accelerators become obsolete quickly as the applications evolve. What is needed is a structured approach for generating programmable accelerators and for updating the software compiler as the accelerator architecture evolves with the applications. In this talk, I will describe a new agile methodology for co-designing programmable hardware accelerators and compilers. Our methodology employs a combination of new programming languages and formal methods to automatically generate the accelerator hardware and its compiler from a single specification. This enables faster evolution and optimization of accelerators, because of the availability of a working compiler. I will showcase this methodology using Amber, a coarse-grained programmable accelerator for imaging and machine learning (ML) we designed and fabricated using our flow in TSMC 16 nm technology. I will show how we agilely evolved Amber into Onyx, our next generation accelerator, using an application-driven design space exploration framework called APEX enabled by our hardware-compiler co-design flow. Bio: Priyanka Raina is an Assistant Professor of Electrical Engineering at Stanford University. She received her B.Tech. degree in Electrical Engineering from the IIT Delhi in 2011 and her S.M. and Ph.D. degrees in Electrical Engineering and Computer Science from MIT in 2013 and 2018. Priyanka's research is on creating high-performance and energy-efficient architectures for domain-specific hardware accelerators in existing and emerging technologies. She also works on methodologies for agile hardware-software co-design. Her research has won best paper awards at VLSI, ESSCIRC and MICRO conferences and in the JSSC journal. She has also won the Intel Rising Star Faculty Award, Hellman Faculty Scholar Award and is a Terman Faculty Fellow.

1 0

Princeton Robotics Seminar: Feb 10, 2022 - Ken Goldberg
by Emily C. Lawrence 07 Feb '23

07 Feb '23

Speaker: Ken Goldberg, William S. Floyd Jr. Distinguished Chair in Engineering, UC Berkeley Date: Friday Feb 10, 2022 Time: 11am EST Location: Computer Science 105 Event page: [ https://robo.princeton.edu/seminar | https://robo.princeton.edu/seminar ] Title: The New Wave in Robot Grasping Abstract: Despite 50 years of research, robots remain remarkably clumsy, limiting their reliability for warehouse order fulfillment, robot-assisted surgery, and home decluttering. The First Wave of grasping research is purely analytical, applying variations of screw theory to exact knowledge of pose, shape, and contact mechanics. The Second Wave is purely empirical: end-to-end hyperparametric function approximation (aka Deep Learning) based on human demonstrations or time-consuming self-exploration. A "New Wave" of research considers hybrid methods that combine analytic models with stochastic sampling and Deep Learning models. I'll present this history with new results from our lab on grasping diverse and previously-unknown objects. Bio: Ken Goldberg is the William S. Floyd Distinguished Chair in Engineering at UC Berkeley and an award-winning roboticist, filmmaker, artist and popular public speaker on AI and robotics. Ken trains the next generation of researchers and entrepreneurs in his research lab at UC Berkeley; he has published over 300 papers, 3 books, and holds 9 US Patents. Ken’s artwork has been featured in 70 art exhibits including the 2000 Whitney Biennial. He is a pioneer in technology and artistic visual expression, bridging the “two cultures” of art and science. With unique skills in communication and creative problem solving, invention, and thinking on the edge, Ken has presented over 600 invited lectures at events around the world. Ken has been interested in robots, rockets, and rebels since he was a kid. He’s skeptical about claims that humans are on the verge of being replaced by Superintelligent machines yet optimistic about the potential of technology to improve the human condition. Ken developed the first provably complete algorithm for part feeding and the first robot on the Internet. In 1995 he was awarded the Presidential Faculty Fellowship and in 2005 was elected IEEE Fellow: "For contributions to networked telerobotics and geometric algorithms for automation." Ken founded UC Berkeley's Art, Technology, and Culture public lecture series in 1997 serves on the Advisory Board of the RoboGlobal Exchange Traded Fund. Ken is Chief Scientist at Ambidextrous Robotics and on the Editorial Board of the journal Science Robotics. He served as Chair of the Industrial Engineering and Operations Research Department and co-founded the IEEE Transactions on Automation Science and Engineering. Short documentary films he co-wrote were selected for Sundance and one was nominated for an Emmy Award. He lives in the Bay Area and is madly in love with his wife, filmmaker and Webby Awards founder Tiffany Shlain, and their two daughters.

1 0

CS & ECE Seminar: Sarita Adve, Wed. Feb 1 at 4:30
by Emily C. Lawrence 30 Jan '23

30 Jan '23

Speaker: Sarita Adve, University of Illinois at Urbana-Champaign Title: Enabling the Immersive Era of Computing Day: Wednesday, February 1, 2023 Time: 4:30 pm EST Location: CS 105 Host: Kai Li & Sharad Malik Abstract: Computing is on the brink of a new immersive era. Recent innovations in virtual/augmented/mixed reality (extended reality or XR) show the potential for a new immersive modality of computing that will transform most human activities and change how we design, program, and use computers. There is, however, an orders of magnitude gap between the power/performance/quality-of-experience (QoE) attributes of current and desirable immersive systems. Bridging this gap requires an ambitious systems research agenda that is end-to-end QoE driven, is based on hardware-software-algorithm co-design, and spans the end-user device, edge, and cloud. To enable this agenda, my group has built ILLIXR (Illinois eXtended Reality testbed), an open source XR system and testbed for end-to-end immersive systems research. I will describe ILLIXR and the research it has enabled across the computing stack. I will also discuss the industry supported ILLIXR consortium launched to democratize XR systems research, development, and benchmarking, and the Center for Immersive Computing at Illinois spanning immersive hardware to applications research and education. Bio: Sarita Adve is the Richard T. Cheng Professor of Computer Science at the University of Illinois at Urbana-Champaign where she directs the Center for Immersive Computing. Her research interests span the system stack, ranging from hardware to applications. Her work on the data-race-free, Java, and C++ memory models forms the foundation for memory models used in most hardware and software systems today. Recently, her group released ILLIXR (Illinois Extended Reality testbed), an open-source extended reality system and research testbed, and launched the ILLIXR consortium to democratize XR research, development, and benchmarking. She is also known for her work on heterogeneous systems and software-driven approaches for hardware resiliency. She is a member of the American Academy of Arts and Sciences, a fellow of the ACM and IEEE, and a recipient of the ACM/IEEE-CS Ken Kennedy award. As ACM SIGARCH chair, she co-founded the CARES movement, winner of the CRA distinguished service award, to address discrimination and harassment in Computer Science research events. She received her PhD from the University of Wisconsin-Madison and her B.Tech. from the Indian Institute of Technology, Bombay. *This talk will NOT be recorded or live streamed.

1 0

Zhuqi Li will present his FPO "Cross-layer Optimization for Video Delivery on Wireless Networks" on Thursday, January 26, 2023 at 3pm in CS 402
by Nicki Mahler 20 Jan '23

20 Jan '23

Zhuqi Li will present his FPO "Cross-layer Optimization for Video Delivery on Wireless Networks" on Thursday, January 26, 2023 at 3pm in CS 402 The members of his committee are as follows: Examiners: Kyle Jamieson (adviser), Ravi Netravali, and Jennifer Rexford Readers: Amit Levy and Victor Bahl (Microsoft) A copy of his thesis will be available before the FPO upon request. Please email gradinfo(a)cs.princeton.edu if you would like a copy of the thesis. Everyone is invited to attend his talk. Abstract follows below: Mobile video applications have gained increasing popularity and become part of everyone’s daily experience. The quality of video has a significant impact on both the quality of users’ experience for video streaming and the accuracy of video analytic systems, which further impact the application revenue. The challenge to building a consistently high-quality video delivery system lies in two aspects. On the application side, the emerging new video applications are evolving to become more user-interactive, where existing prefetch and buffering algorithms cannot work properly. On the network side, the wireless network itself is fundamentally dynamic and unreliable due to the multipath effect and interference on the wireless channel. In this thesis, we present cross-layer optimizations from the application layer, network layer, and physical layer to improve the quality of video streaming over wireless network with the design and implementation of the following systems: Dashlet, a short video streaming system tailored for a high quality of experience by adapting to dynamic user actions. Dashlet proposes a novel out-of-order video chunk pre-buffering mechanism that leverages a simple, non machine learning-based model of users’ swipe statistics to determine the pre-buffering order and bitrate. Spider, a multi-hop, millimeter-wave (mmWave) wireless relay network design to maximize the video analytic accuracy for the delivered video. Spider integrates a low-latency Wi-Fi control plane with a mmWave relay data plane, allowing agile re-routing around blockages. Spider proposes a novel video bit-rate allocation algorithm coupled with a scalable routing algorithm that maximizes application-layer video analytics accuracy. LAIA, a system to programmable control the wireless channel so that the wireless network can achieve consistently high throughput for robust video delivery. With the programmable interface to control the wireless channel, LAIA can improve wireless channels on the fly for single- and multi-antenna links, as well as nearby networks operating on adjacent frequency bands. Putting it together, this thesis demonstrates a set of optimizations in different layers in through network stack for building a high quality and robustness wireless video delivery system. The extensive evaluation demonstrates a significant improvement on both quality of experience for video streaming and accuracy for video analytics.

1 1

Xingyuan Sun will present his FPO "Gradient-Based Shape Optimization for Engineering Using Machine Learning" on Friday, February 17, 2023 at 1:00 PM in CS 402 and via Zoom.
by CS Grad Department 20 Jan '23

20 Jan '23

Xingyuan Sun will present his FPO "Gradient-Based Shape Optimization for Engineering Using Machine Learning" on Friday, February 17, 2023 at 1:00 PM in CS 402 and via Zoom. Location: CS 402, Zoom Link: [ https://princeton.zoom.us/j/98360015495 | https://princeton.zoom.us/j/98360015495 ] . The members of Xingyuan’s committee are as follows: Examiners: Ryan Adams (Co-Adviser), Szymon Rusinkiewicz (Co-Adviser), Olga Russakovsky Readers: Sigrid Adriaenssens, Felix Heide A copy of his thesis is available upon request. Please email gradinfo(a)cs.princeton.edu if you would like a copy of the thesis. Everyone is invited to attend his talk. Abstract follows below: Shape design problems are important in engineering, e.g., trajectory planning for robot arms, material distribution optimization, etc. However, existing works usually solve these tasks without the help of gradients, whose efficiency can be limited. We formalize design problems as constrained optimization tasks and propose to use gradient-based optimizers with automatic differentiation to solve them. Specifically, we use the adjoint method when the underlying physical process can be characterized by PDEs. In Chapter 2, we solve for extruder paths of 3D printing that can compensate for the deformation caused by the fiber printing process. As the printing process is complex and difficult to model, we create a synthetic dataset and fit it using a neural network to get a differentiable surrogate of the printing simulator. We further speed up the optimization process by using a neural network to amortize it, sacrificing a bit of accuracy but getting much faster, real-time inferences. In Chapter 3, we study the task of fiber path planning, figuring out where to lay reinforcing fibers in plastic for 3D printing, maximizing stiffness of the composite. We build a simulator by solving the linear elastic equations and use the adjoint method for gradient calculation and BFGS for fiber path optimization. In Chapter 4, we investigate the problem of dovetail joint shape optimization for stiffness. To model the contact between two parts of a joint, we build a simulator by alternatively solving one side of the joint while fixing the other side. We use the adjoint method for gradient computation and gradient descent for optimization. All methods across the projects are tested both in simulation and real-world experiments, showing our approach produces high-quality designs, and also the amortized approach provides real-time inference while achieving a comparable design quality.

1 0

William Yang will present his General Exam "Rethinking Out-of-Distribution Detection" on Tuesday, January 24, 2023 at 12:00 PM in CS 401.
by CS Grad Department 20 Jan '23

20 Jan '23

William Yang will present his General Exam "Rethinking Out-of-Distribution Detection" on Tuesday, January 24, 2023 at 12:00 PM in CS 401. Committee Members: Olga Russakovsky (advisor), Ryan Adams, Tom Griffiths Abstract: Out-of-Distribution (OOD) detection is the task of identifying inputs on which a given model should not be trusted to make a prediction. OOD detection is a well-studied problem, commonly formulated for image classifiers as detecting test images from classes that did not appear in the model's training data (“semantic distribution shift''). However, this formulation is limited, as it ignores test images which may come from the familiar classes but may look different from the training examples (“covariate distribution shift''). Incorporating covariate shift into OOD detection poses a challenge because detecting all such examples may undervalue the model's robustness and ability to generalize beyond its training distribution. In this work we propose a new formulation of OOD detection, approaching the problem with respect to the learned model distribution rather than the training data distribution. This formulation seamlessly incorporates both semantic and covariate shift, and synergizes the perspective of OOD detection with model generalization. Our empirical analysis reveals a number of interesting findings, the most striking being that the simplest OOD detection baseline, Maximum Softmax Probability (MSP), appears to outperform all prior state-of-the-art OOD detection methods under the new formulation. This suggests that the time is ripe to rethink the task of OOD detection, and to adopt the more generalized benchmark which encompasses a broader range of real-world situations. Reading List: https://docs.google.com/document/d/122o_aUoyzUEYjNESxGoog4VhonZpLaPmw1HffB-… Everyone is invited to attend the talk, and those faculty wishing to remain for the oral exam following are welcome to do so.

1 0

Re: [talks] Abhishek Panigrahi will present his General Exam "Demystifying Gradient Descent in modern Deep Learning: Implicit training biases and Modular Generalization" on Wednesday, January 18, 2023 at 3:00 PM over Zoom.
by gradinfo＠cs.princeton.edu 18 Jan '23

18 Jan '23

Please note, the time listed in the original subject line of the announcement was incorrect. This general exam will take place at 3:00 PM today, per the body text. From: talks-bounces(a)lists.cs.princeton.edu <talks-bounces(a)lists.cs.princeton.edu> On Behalf Of Louis W. Riehl Sent: Tuesday, January 17, 2023 3:00 PM To: talks <talks(a)lists.cs.princeton.edu> Subject: [talks] Abhishek Panigrahi will present his General Exam "Demystifying Gradient Descent in modern Deep Learning: Implicit training biases and Modular Generalization" on Wednesday, January 18, 2023 at 10:00 AM 3:00 PM over Zoom. Abhishek Panigrahi will present his General Exam "Demystifying Gradient Descent in modern Deep Learning: Implicit training biases and Modular Generalization" on Wednesday, January 18, 2023 at 3:00 PM over Zoom. Zoom link: https://princeton.zoom.us/j/94052692845 Committee Members: Sanjeev Arora (advisor), Elad Hazan, Danqi Chen Abstract: Modern deep learning involves training large scale neural networks, which comes at the cost of deciding the best training recipe. Traditional machine learning fails to explain the hidden mechanisms of such models, owing to the high non-convexity of the model landscape. My research focuses on the training-time interplay between different training algorithms and the architecture that drives the generalizability of these models. In this talk, I will focus on Gradient Descent and its two novel mechanisms: (a) the Edge of Stability in Deep Learning, where the interplay between learning rate and the model landscape leads to implicit regularization of hessian flatness during training, and (b) modular skill acquisition that drives generalization during language-model fine-tuning. Reading List: https://docs.google.com/document/d/1ZWwBodrZi86Jlli3r2CUlsfELlGwRAL1KfsFmpl… Everyone is invited to attend the talk, and those faculty wishing to remain for the oral exam following are welcome to do so. Louis Riehl Graduate Administrator Computer Science Department, CS213 Princeton University (609) 258-8014

1 0

Abhishek Panigrahi will present his General Exam "Demystifying Gradient Descent in modern Deep Learning: Implicit training biases and Modular Generalization" on Wednesday, January 18, 2023 at 10:00 AM over Zoom.
by Louis W. Riehl 17 Jan '23

17 Jan '23

Abhishek Panigrahi will present his General Exam "Demystifying Gradient Descent in modern Deep Learning: Implicit training biases and Modular Generalization" on Wednesday, January 18, 2023 at 3:00 PM over Zoom. Zoom link: [ https://princeton.zoom.us/j/94052692845 | https://princeton.zoom.us/j/94052692845 ] Committee Members: Sanjeev Arora (advisor), Elad Hazan, Danqi Chen Abstract: Modern deep learning involves training large scale neural networks, which comes at the cost of deciding the best training recipe. Traditional machine learning fails to explain the hidden mechanisms of such models, owing to the high non-convexity of the model landscape. My research focuses on the training-time interplay between different training algorithms and the architecture that drives the generalizability of these models. In this talk, I will focus on Gradient Descent and its two novel mechanisms: (a) the Edge of Stability in Deep Learning, where the interplay between learning rate and the model landscape leads to implicit regularization of hessian flatness during training, and (b) modular skill acquisition that drives generalization during language-model fine-tuning. Reading List: [ https://docs.google.com/document/d/1ZWwBodrZi86Jlli3r2CUlsfELlGwRAL1KfsFmpl… | https://docs.google.com/document/d/1ZWwBodrZi86Jlli3r2CUlsfELlGwRAL1KfsFmpl… ] Everyone is invited to attend the talk, and those faculty wishing to remain for the oral exam following are welcome to do so. Louis Riehl Graduate Administrator Computer Science Department, CS213 Princeton University (609) 258-8014

1 0

Princeton Quantum Colloquium: Unifying Quantum Algorithms, Professor Nathan Wiebe, Monday, January 23
by Emily C. Lawrence 17 Jan '23

17 Jan '23

Princeton Quantum Colloquium Please join us on Monday, January 23rd for our Princeton Quantum Colloquium seminar series. There will be a light lunch in the Bowen Atrium at noon before the seminar. Speaker: Professor Nathan Wiebe (University of Toronto) Date: Monday, January 23, 2023 Time: 12:30 pm Room: Bowen Hall Auditorium, Light lunch served at 12:00 pm in Bowen Atrium Host: Professor Sarang Gopalakrishnan Title: Unifying Quantum Algorithms Abstract: Quantum computing has made great strides within the last few years with experimental demonstrations of quantum computers with in excess of one hundred quantum bits, elementary quantum error correction demonstrations and ever more sophisticated simulation experiments. Caught up in this discussion though is the question of what these quantum computers will be useful for and more importantly why they will be useful. This point is often articulated by the claim made by many scientists that there are only a handful of quantum tricks that we know of that we can use to leverage a quantum advantage. The aim of my talk is to provide an intuitive understanding of where the power of quantum computing comes from and to explain several of the core tricks that are used to provide exponential speedups for problems ranging from factoring to simulation. From there I will discuss recent work by my collaborators and myself that actually aims to further reduce this set by devising a unified framework that allows us to think of many of these tricks as a manifestation of a larger concept called a block encoding and show a new family of techniques that can be used to modify these block encodings to solve computational problems. These approaches promise not only to give simpler ways of thinking about quantum algorithms, but also provide a new host of tools that can be used to provide quantum advantage for a wide range of problems that far exceeds the initial scope believed to be possible for quantum computers.

1 0

Re: [talks] UPDATE Kaifeng Lyu will present his General Exam "Theoretical Analysis of Modern Neural Networks with Normalization Layers: Auto Rate-Tuning and Sharpness Reduction" on Tuesday, January 17, 2023 at 11:00 AM over Zoom.
by Grad Info 17 Jan '23

17 Jan '23

This General Exam had to be moved up to TODAY (Tuesday), January 17, 2023 at 11:00 AM. Thank you. From: talks-bounces(a)lists.cs.princeton.edu <talks-bounces(a)lists.cs.princeton.edu> On Behalf Of CS Grad Department Sent: Friday, January 13, 2023 1:00 PM To: talks <talks(a)lists.cs.princeton.edu> Subject: [talks] Kaifeng Lyu will present his General Exam "Theoretical Analysis of Modern Neural Networks with Normalization Layers: Auto Rate-Tuning and Sharpness Reduction" on Tuesday, January 17, 2023 at 1:00 PM over Zoom. Kaifeng Lyu will present his General Exam "Theoretical Analysis of Modern Neural Networks with Normalization Layers: Auto Rate-Tuning and Sharpness Reduction" on Tuesday, January 17, 2023 at 1:00 PM 11:00 AM over Zoom. Zoom link: https://princeton.zoom.us/j/97584410651 Committee Members: Sanjeev Arora(advisor), Jason D. Lee, Elad Hazan Abstract: In the past, traditional machine learning techniques relied on convex optimization for training. However, the deep learning revolution of the last decade has led to significant progress in many machine learning tasks, even though the loss landscape is highly non-convex. My research focuses on theoretical analysis of the training dynamics of deep neural nets, including convergence guarantees to low-loss solutions and implicit regularization effects of gradient-based methods. I will present our work on understanding the training dynamics in presence of normalization layers (e.g., Batch Normalization, Layer Normalization), which are important building blocks of modern neural nets. Our work shows the following: (1) normalization layers make the training process adaptive to different loss landscapes, even when using non-adaptive methods such as SGD; (2) normalization layers prevent full-batch gradient descent from converging to sharp minima and introduce an implicit bias towards flatter minima with better generalization, following a continuous sharpness-reduction flow. Reading List: https://docs.google.com/document/d/1BwImXphwwK5lk1NKDC0LApys0soGY7TVpptTZZz… Everyone is invited to attend the talk, and those facultywishing to remain for the oral exam following are welcome to do so.

1 0