- talks - lists.cs.princeton.edu

Christina Shatford will present her General Exam "Simultaneous Pattern and Detector Optimization for Visual Tags" on Thursday, May 18, 2023 at 3:00 PM in CS 402.
by Louis W. Riehl 15 May '23

15 May '23

Christina Shatford will present her General Exam "Simultaneous Pattern and Detector Optimization for Visual Tags" on Thursday, May 18, 2023 at 3:00 PM in CS 402. Committee Members: Szymon Rusinkiewicz (advisor), Ryan Adams, Olga Russakovsky Abstract: Accurate and robust object localization has been a long studied problem in computer vision. In this work, I present a new approach for localizing visual tags in images. Instead of optimizing a detector to locate a tag with a static pattern, this approach involves the simultaneous optimization of both the detector and the pattern of the visual tag. I will show how this joint learning leads to increased localization accuracy when compared to only training the detector and using a QR code or AprilTag, two standard static tags. In addition, I will discuss how pivoting away from tags in general, to an approach that uses steganography inspired techniques to hide a payload in the spatial domain of images, has its own host of useful applications. Reading List: [ https://docs.google.com/document/d/1_h9vNeeefjGkkD1FueUpRXBuLvTfGZrR2r6eX46… | https://docs.google.com/document/d/1_h9vNeeefjGkkD1FueUpRXBuLvTfGZrR2r6eX46… ] 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

Alexander Wettig will present his General Exam "Towards a Better Understanding of Language Pre-training and Fine-tuning" on Thursday, May 18, 2023 at 3:00 PM in FC 114 and via zoom.
by Louis W. Riehl 15 May '23

15 May '23

Alexander Wettig will present his General Exam "Towards a Better Understanding of Language Pre-training and Fine-tuning" on Thursday, May 18, 2023 at 3:00 PM in FC 114 and via zoom. Zoom link: [ https://princeton.zoom.us/j/93571660087 | https://princeton.zoom.us/j/93571660087 ] Committee Members: Danqi Chen (advisor), Karthik Narasimhan, Sanjeev Arora Abstract: The paradigm of pre-training and fine-tuning has become a cornerstone in Natural Language Processing (NLP) and has achieves state-of-the-art performance on a wide range of tasks. In this talk, we will delve into key aspects of masked language model (MLM) pre-training and fine-tuning dynamics. In the first part, we will challenge the conventional wisdom surrounding the masking rate in MLM pre-training. Traditionally, a 15% masking rate has been widely adopted, but we show that this is not universally optimal, and 40% can outperform 15% under an efficient pre-training recipe. Surprisingly, even an 80% masking rate can still maintain 95% fine-tuning performance and acquire non-trivial linguistic features. We explore how the masking rate depends on the model size and the masking strategies, and shed light on the distinct effects of increasing the masking rate on task difficulty and optimization. In the second part, we will investigate whether kernel-based dynamics can describe the fine-tuning dynamics of a pre-trained language models, particularly in the few-shot settings. Inspired by the Neural Tangent Kernel (NTK), our analysis characterizes the conditions under which gradient features can solve the fine-tuning updates. We conduct experiments across 14 NLP tasks and demonstrate that formulating the downstream task as a masked word prediction problem through prompting often induces kernel-based dynamics during fine-tuning. The kernel view also provides an explanation for the success of parameter-efficient subspace-based fine-tuning methods. Reading List: [ https://docs.google.com/document/d/1HCyyShO_2ilodDbv9ibG2_84m3tkvatKHOlavLo… | https://docs.google.com/document/d/1HCyyShO_2ilodDbv9ibG2_84m3tkvatKHOlavLo… ] 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

Alexander Raistrick will present his General Exam "Procedural Synthetic Data for Computer Vision" on Wednesday, May 17, 2023 at 3:00 PM in CS 301.
by Louis W. Riehl 12 May '23

12 May '23

Alexander Raistrick will present his General Exam "Procedural Synthetic Data for Computer Vision" on Wednesday, May 17, 2023 at 3:00 PM in CS 301. Committee Members: Jia Deng (advisor), Szymon Rusinkiewicz, Felix Heide Abstract: Data, especially large scale labeled data, has been a critical driver of progress in computer vision. At the same time, many tasks remain starved of high-quality data. Synthetic data from computer graphics is a promising solution to this data challenge, but sees limited use, which we hypothesize is due to limited diversity of assets. We propose a synthetic data generator which can produce infinite photorealistic 3D scenes of the natural world. Our system is entirely procedural: every asset, from shape to texture, is generated from scratch via randomized mathematical rules. This results in a limitless variety of 3D assets, with rich compositionality and pixel-level 3D detail. Moreover, it means that all assets are original, relying on no external source, and can be released for free as open source code. We provide such procedural rules for a wide range of natural objects and scenes, including plants, animals, terrain and lighting. Procedural modeling requires significant programming effort, and generating full scenes poses a large computational challenge. To this end, we provide tools and frameworks to accelerate procedural rule creation, and enable large-scale rendering of complex procedurally generated scenes. In all, anyone can freely use our system to generate unlimited high-quality training data for a wide range of computer vision tasks, including object detection, semantic segmentation, optical flow and 3D reconstruction. We expect this to be a useful tool for computer vision research and beyond. Reading List: [ https://docs.google.com/document/d/1N5spwmoIOqaErKy4X0B4l3TvM3m8S6rhGSRHFYz… | https://docs.google.com/document/d/1N5spwmoIOqaErKy4X0B4l3TvM3m8S6rhGSRHFYz… ] 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

Jihoon Chung will present his General Exam "Enabling Detailed Action Recognition Evaluation Through Video Dataset Augmentation" on Wednesday, May 17, 2023 at 10:15 AM in CS 302 .
by Louis W. Riehl 12 May '23

12 May '23

Jihoon Chung will present his General Exam "Enabling Detailed Action Recognition Evaluation Through Video Dataset Augmentation" on Wednesday, May 17, 2023 at 10:15 AM in CS 302 . Committee Members: Olga Russakovsky (advisor), Jia Deng, Felix Heide Abstract: Human action recognition is the task of automatically classifying a video according to the action of the person depicted in the video. Despite the overall high accuracy on current benchmarks, it is well-known in the video understanding community that human action recognition models suffer from background bias, i.e., over-relying on scene cues in making their predictions. However, it is difficult to quantify this effect using existing evaluation frameworks. We introduce the Human-centric Analysis Toolkit (HAT), which enables the evaluation of learned background bias without the need for new manual video annotation. It does so by automatically generating synthetically manipulated videos and leveraging the recent advances in image segmentation and video inpainting. Using HAT we perform an extensive analysis of 74 action recognition models trained on the Kinetics dataset. We confirm that all these models focus more on the scene background than on the human motion; further, we demonstrate that certain model design decisions (such as training with fewer frames per video or using dense as opposed to uniform temporal sampling) appear to worsen the background bias. We open-source HAT to enable the community to design more robust and generalizable human action recognition models. Reading List: [ https://docs.google.com/document/d/1FQEQrX3HcM0E9bBPEcoVPSy19o7IV77KvyirxVd… | https://docs.google.com/document/d/1FQEQrX3HcM0E9bBPEcoVPSy19o7IV77KvyirxVd… ] 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

Madelyne Xiao will present her General Exam "Machine Learning Approaches for Automated Misinformation Detection" on Tuesday, May 16, 2023 at 3:00 PM in Sherrerd 008 and via zoom.
by Louis W. Riehl 11 May '23

11 May '23

Madelyne Xiao will present her General Exam "Machine Learning Approaches for Automated Misinformation Detection" on Tuesday, May 16, 2023 at 3:00 PM in Sherrerd 008 and via zoom. Zoom link: [ https://princeton.zoom.us/my/madelyne.x | https://princeton.zoom.us/my/madelyne.x ] Committee Members: Jonathan Mayer (advisor), Arvind Narayanan, Andrew Guess Abstract: The spread of misinformation, disinformation, and false news on social media platforms in recent years has triggered a corresponding proliferation of automated misinformation detection methods. We present a systematic review and analysis of existing machine learning-driven approaches to the automated detection of misinformation. Across 248 well-cited papers in the field, we develop a taxonomy of five different information “scopes” for which numerous automated methods have been developed: claims, news articles, social media posts, users/authors, and whole websites. Having established this framework, we identify errors in ML-driven misinformation detection methods, with a focus on errors in corpus curation and model development. In particular, we find that methods that rely solely on textual features suffer from a lack of information context, such as reader response and source reputation. As a result, these methods are especially susceptible to text-specific dependencies (e.g., narrowly-defined subject domains and stylistic idiosyncrasies) that interfere with actual signals for misinformative content. Meanwhile, methods that consider extra-textual features (user behavior, social media platform-specific UI elements, knowledge networks) suffer from errors common to the “curse of dimensionality”: poor feature selection and mismanagement of data proxies. We support these findings with three replication studies, whose results highlight serious issues of data availability, reproducibility, and method generalizability within the existing misinformation detection literature. In view of the results of our literature analysis and replication studies, we conclude that existing ML-based methods for automated misinformation detection fall far short of what might be reasonably expected of methods deployable in a real world setting. As such, we make recommendations for proactive misinformation interventions beyond automated ML detection methods. Reading List: [ https://docs.google.com/document/d/1J80hwND6Jiw29ohiuEizjT7AchNNOtZFLH2ySTE… | https://docs.google.com/document/d/1J80hwND6Jiw29ohiuEizjT7AchNNOtZFLH2ySTE… ] 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

Special Quantum Seminar: Quantum Technologies, Thursday, May 11 at 4:00pm
by Emily C. Lawrence 11 May '23

11 May '23

QUANTUM SEMINAR Speakers: Dr. Marianna Bonanome, Dr. Kayla Quinnies, Dr. Constantijn van der Poel, SandboxAQ Day: Thursday, May 11, 2023 Time: 4:00 PM EST Location: CS 104 Reception: 3:30pm Convocation Room Hosts: Jennifer Rexford and Andrew Houck Title: Quantum Technologies Abstract: Starting in 1900 with Max Plank, some of the world’s brightest talents developed a new way of thinking about how matter and energy work on the atomic and subatomic scale. This gave rise to quantum theory and kicked off a chain reaction of scientific revolutions that are still going on today. We are now in the midst of the 3rd quantum revolution. We’re able to use quantum particles as sensors, to perform computations, and to create applications. Quantum sensing devices are beginning to revolutionize the fields of medical diagnostics and navigation. Quantum key distribution protocols are already being used to create secret keys in order to protect data. We're just now getting to the point where we're starting to understand the real promise of what quantum science can provide to humanity. At SandboxAQ we are a technology company that leverages the power of AI and quantum technology to develop commercial products for various industries, including financial services, healthcare, and telecommunications. Our team at SandboxAQ comes from diverse backgrounds, including physics, computer science, neuroscience, mathematics, and cryptography brought together in a culture of experimental thinking and collaboration. We operate with a mission to create breakthrough solutions that push the boundaries of physics and AI. Join us as we discuss these cutting edge technologies and their potential! Marianna Bonanome, PhD, Head of Education Outreach, SandboxAQ Bio: Before joining SandboxAQ as the Head of Education Outreach, Marianna was an Associate Professor of Applied Mathematics and Computer Science at the New York City College of Technology, CUNY. She is also the co-founder, and formerly CEO, of Rationarium, a blockchain edtech startup. She is an ardent open source advocate and is an executive board member and Community Manager of the non-profit: The WeBWorK Project (TWP). While at CUNY, she was a co-director of a Title V Department of Education collaborative grant and founded and led an open source university-wide initiative. She received a B.S. in mathematics with a minor in physics from NYU Tandon School of Engineering and both her master’s and doctoral degrees from the Graduate School and University Center of CUNY in mathematics. Marianna has been teaching since 2003 and has a passion for mathematics, physics and teaching pedagogy and so actively researches and publishes in the areas of combinatorial group theory, quantum computation and education. Kayla Quinnies, PhD, Residency Program Manager, SandboxAQ Bio: Prior to joining the education team at SandboxAQ as the Residency Program Manager, Kayla received her PhD in neuroscience from the University of Virginia, did postdoctoral work in biomedical informatics at Columbia University, and was the Outreach Program Manager for Columbia University Engineering. While at Columbia Engineering, she ran an education program for high school students, working closely with undergraduates and faculty. She has been working in STEM outreach and science community management for over ten years, and is passionate about creating a meaningful experience for students in the Residency Program at SandboxAQ. Constantijn van der Poel, PhD, AI Resident, SandboxAQ Bio: As an AI Resident at SandboxAQ, Constantijn’s doctoral research lies at the intersection of quantum physics and artificial intelligence, specifically Natural Language Processing (NLP). Constantijn utilizes tensor networks, which comes from the world of quantum, to model Motzkin spin-chains, whose entanglement captures long-range correlations also found in language. Constantijn is always eager to learn about new technologies and their applications, and has a passion for sharing what he learns with others. He also has a deep love for delving into the cultures and languages of the world, past and present.

1 1

CS Special Seminar: Shimon Schocken, Thursday, May 11 at 12:30pm
by Emily C. Lawrence 11 May '23

11 May '23

Speaker: Shimon Schocken, Reichman University, Israel Date: Thursday, May 11 Time: 12:30pm EST Location: CS 105 Event page: [ https://www.cs.princeton.edu/events/26410 | https://www.cs.princeton.edu/events/26410 ] Title: Nand to Tetris: Applied Computer Science from the Ground Up Abstract: I'll present COS205, a course that I taught at Princeton during the Spring 2023 semester. The course synthesizes many abstractions, algorithms, and data structures learned in core CS courses, and makes them concrete by building a complete computer system from the ground up. The methodology is based on guiding students through a set of 12 hands-on projects that gradually construct and unit-test a basic hardware platform and a modern software hierarchy, yielding a simple yet surprisingly powerful computer system. The hardware projects are done in a hardware description language and a hardware simulator supplied by us. The software projects (assembler, virtual machine, and a compiler for an object-based, Java-like language) can be developed in any language, using API's and test programs supplied by us. We also build a basic OS. The result is a general-purpose computer system, simulated on the student's PC. Students who wish to do so can build the computer physically, using FPGA. All the course materials are freely available in open source at www.nand2tetris.org (joint work with Noam Nisan, Hebrew University). Bio: Shimon Schocken, William R. Kenan Jr. Visiting Professor for Distinguished Teaching School of Engineering, Princeton University (on leave from Reichman University, Israel) Shimon Schocken is founding-dean of the Efi Arazi School of Computer Science at Reichman University. He is also co-founder of the Google-Reichman School of Technology, an academic bootcamp dedicated to training young men and women from underprivileged backgrounds, and Matific, a company that develops computer games that teach elementary school mathematics. He was a tenured professor at NYU, and chairman of the computer science committee at Israel's Ministry of Education.

1 1

Andrew Johnson will present his General Exam "Flexible, Line-rate Flow Monitoring of Packet Sequences" on May 17, 2023 at 10am in CS 301 and Zoom.
by Nicki Mahler 10 May '23

10 May '23

Andrew Johnson will present his General Exam "Flexible, Line-rate Flow Monitoring of Packet Sequences" on May 17, 2023 at 10am in CS 301 and Zoom. Zoom Link: [ https://princeton.zoom.us/j/5374916014 | https://princeton.zoom.us/j/5374916014 ] Committee members are as follows: David Walker(adviser), Zachary Kincaid, and Aarti Gupta Abstract and Reading list can be found here: [ https://docs.google.com/document/d/15FOVJU3GKyUwiyupVqNo9BA60eCSVJma-lVuwmn… | https://docs.google.com/document/d/15FOVJU3GKyUwiyupVqNo9BA60eCSVJma-lVuwmn… ] [ https://docs.google.com/document/d/1dGTPt_8kcmmlemUmvDFB57pNsUV_AM9VItxzvf1… | https://docs.google.com/document/d/1dGTPt_8kcmmlemUmvDFB57pNsUV_AM9VItxzvf1… ]

1 0

Daniel Suo will present his FPO "Scaling Machine Learning in Practice" on Wednesday, May 10, 2023 at 3pm in CS 402.
by Nicki Mahler 10 May '23

10 May '23

Daniel Suo will present his FPO "Scaling Machine Learning in Practice" on Wednesday, May 10, 2023 at 3pm in CS 402. The members of his committee are as follows: Examiners: Kai Li (Adviser), Olga Troyanskaya, and Ryan Adams; Readers: Elad Hazan (Co-Adviser) and Naman Agarwal (Google LLC) All are welcome to attend. In recent years, machine learning has become pervasive, powering algorithmic clinicians, translators, and world-beating go masters. As practitioners build on this success, they repeatedly observe that scale–data, model size, compute–is critical. However, scale is now a challenge in and of itself; simple tasks such as gathering data become formidable, even prohibitive. In this talk, we discuss techniques for addressing scale in three areas: 1. Differential reinforcement learning for physical devices: reinforcement learning has emerged as a potential strategy for machines to make decisions in complex, dynamic environments. However, successful demonstrations have required vast experience to learn an optimal policy, making real-world physical applications particularly challenging. We present a method that uses limited experience to learn a differentiable simulator of a physical system (medical ventilator) and then uses gradient methods on the simulator to learn a state-of-the-art policy for controlling that system. 2. Practical optimization for deep learning: optimization is an essential aspect of deep learning. However, while a constellation of optimization algorithms dot the literature, the low burden of proof and empirical nature of deep learning has led practitioners to rely on defaults (i.e., Adagrad, Adam) rather than view optimization as a lever for progress. To rigorously test ideas in optimization, we introduce a comprehensive benchmark that currently includes 8 deep learning workloads and rules for training procedures, computational budget, and evaluation. 3. Scaling computer systems via thread scheduling: large global-scale applications are expensive and complex to operate let alone optimize. As a result, many simple parameters that govern important behaviors of these systems are simply set once and never touched again. However, we show that these parameters present low-hanging fruit for significant eciency improvements

1 1

Nicolaas Kaashoek will present his General Exam "RADICAL Serverless" on May 16, 2023 at 1pm in CS 301
by Nicki Mahler 10 May '23

10 May '23

Nicolaas Kaashoek will present his General Exam "RADICAL Serverless" on May 16, 2023 at 1pm in CS 301. The members of his committee are as follows: Wyatt Lloyd (adviser), Amit Levy, and Mike Freedman Links to Abstract and Reading list below: https://docs.google.com/document/d/1axoXIJtRm70Og4sHIWANJFKzqk1xOYsAkxW8X8Z… https://docs.google.com/document/d/1-xgMy9BvoJ8Rwb7tLwGH_92X_EfUk0vRZifPUlJ… Everyone is invited to attend the talk, and those faculty wishing to remain for the oral exam following are welcome to do so. Abstract: The serverless, or functions-as-a-service (FAAS), cloud computing execution model has proven a useful tool for running computations efficiently in the cloud. Serverless functions scale on demand, can be run either in a data center or at the edge, and use available resources more efficiently than a virtual machine. However, serverless cannot support many workloads that could take advantage of these properties because the functions are stateless. Simply adding storage to serverless to make it stateful is not sufficient either, as it can impose a heavy performance cost. In this talk, I will present my work towards building a more flexible, stateful serverless execution environment: RADICAL serverless. RADICAL serverless functions can run anywhere with durability, consistency, and low latency. RADICAL is able to deliver on its promise by simultaneously executing functions in both a local (closer to the user) context to provide low latency, and a remote context that has access to consistent, durable storage. This combination allows RADICAL to, through the use of read-set extraction and speculative execution, get the best of both worlds.

1 1