We hope you enjoyed spring break.  Here are next week's CS Department Colloquium Series talks.  Just a reminder, you can find the full calendar of events here: https://www.cs.princeton.edu/general/newsevents/events  

Speaker: Colin Raffel, University of North Carolina at Chapel Hill
Date: Monday, March 20
Time: 12:30pm EST
Location: CS 105
Host: Danqi Chen
Event page: https://www.cs.princeton.edu/events/26354 

Title: Collaborative, Communal, & Continual Machine Learning

Abstract:  Pre-trained models have become a cornerstone of machine learning thanks to the fact that they can provide improved performance with less labeled data on downstream tasks. However, these models are typically created by resource-rich research groups that unilaterally decide how a given model should be built, trained, and released, after which point it is never updated. In contrast, open-source development has demonstrated that it is possible for a community of contributors to work together to iteratively build complex and widely used software. This kind of large-scale distributed collaboration is made possible through a mature set of tools including version control and package management. In this talk, I will discuss a research focus in my group that aims to make it possible to build machine learning models in the way that open-source software is developed. Specifically, I will discuss our preliminary work on merging multiple models while retaining their individual capabilities, patching models with cheaply-communicable updates, designing modular model architectures, and tracking changes through a version control system for model parameters. I will conclude with an outlook on how the field will change once truly collaborative, communal, and continual machine learning is possible.

Bio: Colin Raffel is an Assistant Professor at UNC Chapel Hill and a Faculty Researcher at Hugging Face. His work aims to make it easy to get computers to do new things. Consequently, he works mainly on machine learning (enabling computers to learn from examples) and natural language processing (enabling computers to communicate in natural language). He received his Ph.D. from Columbia University in 2016 and spent five years as a research scientist at Google Brain.

Speaker: Benjamin Eysenbach, Carnegie Mellon University
Date: Tuesday, March 21
Time: 12:30pm EST
Location: CS 105
Host: Karthik Narasim
Event page: https://www.cs.princeton.edu/events/26345

Title: Self-Supervised Reinforcement Learning

Abstract:  Reinforcement learning (RL) promises to harness the power of machine learning to solve sequential decision making problems, with the potential to enable applications ranging from robotics to chemistry. However, what makes the RL paradigm broadly applicable is also what makes it challenging: only limited feedback is provided for learning to select good actions. In this talk, I will discuss how we have made headway of this challenge by designing a class of self-supervised RL methods, ones that can learn skills for acting using unsupervised (reward-free) experience. These skill learning methods are practically-appealing and have since sparked a vibrant area of research. I will also share how we have answered some open theoretical questions in this area.

Bio: Benjamin Eysenbach a final-year PhD student at Carnegie Mellon University. His research has developed machine learning algorithms for sequential decision making. His algorithms not only achieve a high degree of performance, but also carry theoretical guarantees, are typically simpler than prior methods, and draw connections between many areas of ML and CS. Ben is the recipient of the NSF and Hertz graduate fellowships. Prior to the PhD, he was a resident at Google Research and studied math as an undergraduate at MIT.

Speaker: Ari Holtzman, University of Washington
Date: Thursday, March 23
Time: 12:30pm EST
Location: CS 105
Host: Danqi Chen
Event page: https://www.cs.princeton.edu/events/26350

Title:  Controlling Large Language Models: Generating (Useful) Text from Models We Don’t Fully Understand

Abstract: Generative language models have recently exploded in popularity, with services such as ChatGPT deployed to millions of users. These neural models are fascinating, useful, and incredibly mysterious: rather than designing what we want them to do, we nudge them in the right direction and must discover what they are capable of. But how can we rely on such inscrutable systems?

This talk will describe a number of key characteristics we want from generative models of text, such as coherence and correctness, and show how we can design algorithms to more reliably generate text with these properties. We will also highlight some of the challenges of using such models, including the need to discover and name new and often unexpected emergent behavior. Finally, we will discuss the implications this has for the grand challenge of understanding models at a level where we can safely control their behavior.

Bio: Ari Holtzman is a PhD student at the University of Washington. His research has focused broadly on generative models of text: how we can use them and how can we understand them better. His research interests have spanned everything from dialogue, including winning the first Amazon Alexa Prize in 2017, to fundamental research on text generation, such as proposing Nucleus Sampling, a decoding algorithm used broadly in deployed systems such as the GPT-3 API and academic research. Ari completed an interdisciplinary degree at NYU combining Computer Science and the Philosophy of Language.

Quantum Seminar Speaker
Speaker: Kaitlin (Kate) Smith, Super.tech
Date: Thursday, March 23
Time: 4:30pm EST
Location: B205 EQuad
Host: Andrew Houck
Event page: https://www.cs.princeton.edu/events/26356

Title: An Architect’s Perspective on Quantum Computer Scaling:  Why, What, and How?

Abstract:  Quantum computation has potential to solve problems that are out of reach for today’s classical computers. Many of the proposed applications for quantum computers (QCs), such as those in chemistry, material science, and optimization, are capable of substantial human impact. However, the full promise of quantum will only be realized if better qubits and QCs emerge that are capable of large-scale computation. The roadmap to QC scaling does not only contain a single path but many that run in parallel. In addition to pursuing devices with more qubits, quantum researchers must 1) co-design software that pushes the frontier of existing machines and 2) build models that guide future QC design toward optimized performance. In this talk, I discuss the why, what, and how involved with scaling today’s QCs. First, I motivate the pursuit of quantum computing and introduce fundamental concepts. Next, I present a case study that explores optimized quantum circuit compilation, reducing decoherence via circuit slack. I show how quantum algorithms can adapt to the unique characteristics of today’s QCs through optimized gate scheduling, leading to significant improvements in success during runtime. In the third part of this talk, hardware challenges that restrict the number qubits on-chip are highlighted. With a focus on fixed-frequency transmon QCs, I explore the viability of modular architectures to scale quantum devices, presenting promising results in terms of yield, gate performance, and application-based analysis. Finally, an outlook is given on future directions in QC software and hardware co-design that aim to accelerate progress toward achieving practical quantum machines.

Bio: Kaitlin is a quantum software manager at Super.tech, a software division of Infleqtion. >From 2020-2022, she was an IBM and Chicago Quantum Exchange Postdoctoral Scholar in the University of Chicago’s Department of Computer Science, advised by Prof. Fred Chong. Kaitlin is a co-author of the 2022 IEEE International Symposium on High-Performance Computer Architecture (HPCA) Best Paper, named a 2021 MIT EECS Rising Star, and the recipient of the 2021 IEEE Computer Society Technical Committee on Multiple Valued Logic (TC-MVL) Kenneth C. Smith Early Career Award in Microelectronics.