TBD
Wednesday, June 8th, 2022
11:00 am–12:00 pm
Speaker: Laurie Weingart (CMU)
Abstract: In their recent book, Babcock, Peyser, Vesterlund, and Weingart share evidence that across industries, jobs, and levels of seniority, women carry a heavier load of tasks that support the organization but do not advance their careers. These non-promotable tasks range from taking notes at meetings and organizing team-building events to serving on organizational governance committees. When women are overloaded with NPTS they have less time to do the work that matters and end up having to work more hours to avoid falling behind their male colleagues. This is bad for women's careers and bad for their organizations.
In this session, we'll explore why this happens and what you can do to manage your NPT workload. The common thinking is that women just need to learn how to say no, and this will solve the problem. But it's not that simple. We will discuss tried and true strategies that you can use to identify NPTs in your work portfolio and begin the process of bringing your workload into balance, as well as what managers can do to change the distribution of NPTs in the first place.
Bio: Laurie R. Weingart is the Richard M. and Margaret S. Cyert Professor of Organizational Behavior and Theory at the Tepper School of Business, Carnegie Mellon University. She served as CMU's Interim Provost and Chief Academic Officer, and as Senior Associate Dean – Education and Director of the Accelerate Leadership Center within the Tepper School.
Coauthor of The No Club: Putting a Stop to Women's Dead-End Work(©Simon & Schuster, 2022), her research examines collaboration, conflict, and negotiation, with a focus on how differences across people both help and hinder effective problem solving and innovation. Prof. Weingart has published over 70 articles and book chapters in the fields of management, social psychology, industrial psychology, cognitive psychology, and economics. An elected Fellow of the Academy of Management and recipient of the Joseph E. McGrath Award for Lifetime Achievement in the Study of Groups, Dr. Weingart served as President of the International Association for Conflict Management, founding President of the Interdisciplinary Network for Group Research, and as co-editor of the Academy of Management Annals. Dr. Weingart earned her Ph.D. in Organizational Behavior from the Kellogg School of Management at Northwestern University.
Thursday, June 9th, 2022
9:30 am–10:30 am
Speaker: Raluca Ada Popa (UC Berkeley)
Abstract: Existing oblivious storage systems provide strong security by hiding access patterns, but do not scale to sustain high throughput as they rely on a central point of coordination. To overcome this scalability bottleneck, I will present Snoopy, an object store that is both oblivious and scalable such that adding more machines increases system throughput. Snoopy contributes techniques tailored to the high-throughput regime to securely distribute and efficiently parallelize every system component without prohibitive coordination costs. These techniques enable Snoopy to scale similarly to a plaintext storage system. Snoopy achieves 13.7× higher throughput than Obladi, a state-of-the-art oblivious storage system. Specifically, Obladi reaches a throughput of 6.7K requests/s for two million 160-byte objects and cannot scale beyond a proxy and server machine. For the same data size, Snoopy uses 18 machines to scale to 92K requests/s with average latency under 500ms.
Bio: Raluca Ada Popa is the Robert E. and Beverly A. Brooks associate professor of computer science at UC Berkeley working in computer security, systems, and applied cryptography. She is a co-founder and co-director of the RISELab and SkyLab at UC Berkeley, as well as a co-founder & President of Opaque Systems, and a co-founder of PreVeil, two cybersecurity startups. Raluca has received her PhD in computer science as well as her Masters and two BS degrees, in computer science and in mathematics, from MIT. She is the recipient of the 2021 ACM Grace Murray Hopper Award, a Sloan Foundation Fellowship award, Jay Lepreau Best Paper Award at OSDI, Jim and Donna Gray Excellence in Undergraduate Teaching Award, NSF Career, Technology Review 35 Innovators under 35, Microsoft Faculty Fellowship, and a George M. Sprowls Award for best MIT CS doctoral thesis.
11:00 am–12:00 pm
Speaker: Kamalika Chaudhuri (UCSC and Meta AI)
Abstract: The fundamental assumption of learning theory is the statistical learning framework, which assumes that training and test data are drawn from the same distribution.
As machine learning is increasingly deployed in practice, this assumption no longer holds in practice. In this talk, we will take a closer theoretical look at two cases where this assumption is violated. The first is robustness to adversarial examples, where test data consists of perturbed versions of legitimate inputs drawn from the underlying data distribution. The second problem is imbalanced classification — where the training data is highly imbalanced, while test data is less so.
Bio: Kamalika Chaudhuri is a Research Scientist at Meta AI and a Professor at the University of California, San Diego. She received a Bachelor of Technology degree in Computer Science and Engineering in 2002 from Indian Institute of Technology, Kanpur, and a PhD in Computer Science from University of California at Berkeley in 2007. She received an NSF CAREER Award in 2013 and a Hellman Faculty Fellowship in 2012. She has served as the program co-chair for AISTATS 2019 and ICML 2019, and as the General Chair for ICML 2022.
Kamalika’s research interests lie in the foundations of trustworthy machine learning – or machine learning beyond accuracy, which includes problems such as learning from sensitive data while preserving privacy, learning under sampling bias, in the presence of an adversary, and from off-distribution data.
2:00 pm–3:00 pm
Speaker: Barna Saha (UCSD)
Abstract: Edit distance is a fundamental measure of sequence similarity. However, for two strings of length n, exact computation of edit distance requires time quadratic in n which is prohibitive for large data sets. This has resulted in a sequence of recent results that try to design faster algorithms for edit distance allowing for an approximate solution. Can one design any reasonable approximation algorithms for edit distance in time that is sublinear in n? This is of course very challenging. However, recently some progress has been made on this front. In this presentation, I will cover some of our recent works on sublinear time algorithm design for edit distance.
Based on joint works with Elazar Godenberg, Tomasz Kociumaka, Robert Krauthgamer, and Aviad Rubinstein.
Bio: Barna Saha is a Jacob Faculty Scholar and an Associate Professor of Computer Science and Engineering & Halıcıoğlu Data Science Institute at UCSD. Before coming to UCSD, she has held faculty positions at UC Berkeley, and University of Massachusetts Amherst as well as senior research scientist position at the AT&T Shannon Research Laboratory. Her research interest spans algorithm design and analysis, probabilistic methods, and databases. She is a recipient of the
Presidential Early Career Awards for Scientists and Engineers (PECASE), Sloan Fellowship, NSF CAREER Award, Young Alumnus Award (IIT Kanpur), Google and Yahoo faculty awards, and has been recognized as a Kavli fellow from the National Academy of Sciences.
Friday, June 10th, 2022
9:00 am–10:00 am
Speaker: Elette Boyle (Reichman University (IDC) and NTT Research)
Abstract: Zero-knowledge proofs on distributed data (D-ZK) give a means for proving statements on data held distributedly in pieces or secret shared across multiple parties. Concretely, a prover who holds input x wishes to convince a collection of verifiers, each holding a piece of x, that x is indeed contained in some language. The zero knowledge property requires that by participating in the protocol, even subsets of the verifiers do not learn anything about x beyond their original knowledge.
In this talk, we will introduce and discuss the notion of D-ZK proofs, briefly survey known constructions, and present recent applications to cryptographically secure computation protocols.
Bio: Elette Boyle is an Associate Professor and Director of the Foundations & Applications of Cryptography (FACT) Research Center in Computer Science at Reichman University, Israel, and a Senior Scientist at NTT Research. She received her PhD at MIT and BS at Caltech, both in Mathematics, and served as a postdoctoral fellow at Technion Israel and Cornell University.
Elette’s research centers in cryptographic solutions for safely maintaining and processing sensitive data. In particular, her recent focus has been on protocols for secure multi-party computation, as well as underlying primitives such as function/homomorphic secret sharing. Her work has been recognized by awards from the European Research Council (ERC), Israeli Science Foundation (ISF), United States Air Force Office of Scientific Research (AFOSR), Google Research Scholar Award program, and the International Association of Cryptologic Research (IACR).
10:30 am–12:30 pm
Speaker: Gagan Aggarwal (Google Research), Naama Ben-David (VMware), and Kshipra Bhawalkar (Google Research)
Title: Autobidding and Auctions
Abstract: Over the past few years, more and more Internet advertisers have started using automated bidding for optimizing their advertising campaigns. Advertisers using automated bidding have an optimization goal (e.g. to maximize conversions), and some constraints (e.g. a budget or an upper bound on average cost per conversion), and the automated bidding system optimizes their auction bids on their behalf. In this talk, we discuss several fundamental questions around autobidding and auctions. How should an autobidder bid to optimize their goals? Do optimal bidding algorithms depend on the underlying auction? What happens when all advertisers adopt optimal autobidding? Does an equilibrium exist, and is it efficient? How do we define efficiency in the presence of autobidding? What happens to the equilibrium when the auctioneer uses reserve prices to optimize its revenue? We will attempt to answer some of these questions in this talk. In particular, we will present a practical algorithm for autobidding that is optimal if and only if the underlying auction is truthful. We will show that an equilibrium always exists under certain smoothness conditions and that the price of anarchy is no more than 2. We will also show that the use of reserve prices can increase the price of anarchy, and bound the resulting increase.
This is based on joint work with Ashwinkumar Badanidiyuru, Aranyak Mehta, Andres Perlroth and Junyao Zhao.
Bio: Gagan Aggarwal is a Research Scientist at Google, where she co-leads the Market Algorithms research team. Her research interests are in Algorithmic Game Theory and Approximation Algorithms, as well as their application to online marketplaces. She received her PhD in Computer Science in 2005 from Stanford University, and her BTech in Computer Science in 2000 from IIT Delhi.
Title: Fast and Fair Lock-Free Locks
Abstract: Locks are frequently used in concurrent systems to simplify code and ensure safe access to contended parts of memory. However, they are also known to cause bottlenecks in concurrent code, leading practitioners and theoreticians to sometimes opt for more intricate lock-free implementations. In this talk, I’ll show that, despite the seeming contradiction, it is possible to design practically and theoretically efficient lock-free locks. I’ll discuss how we model and reason about concurrent systems theoretically, and present a lock-free lock algorithm with good bounds on running time and fairness.
Bio: Naama Ben-David is a postdoctoral researcher at VMware. Her primary research interests are in the intersection of theory and practice in distributed and concurrent computing. More specifically, Naama strives to theoretically explain phenomena seen in modern machines, and to use obtained insights to design and analyze practical algorithms for multiprocessor settings. Naama’s work has been recognized with several awards, including an honorable mention for the CMU School of Computer Science Dissertation Award, a PPoPP 2022 best paper award, an NSERC postgraduate scholarship and a Microsoft Research PhD Fellowship. After her postdoc at VMware, she will be joining the Technion as an Assistant Professor.
Title: Simple Mechanisms for Rich Advertising Auctions
Abstract: The study of internet advertising auctions has fueled a lot of growth in the area of Algorithmic Game Theory. The traditional model of sponsored Search Auctions is well studied, but more recently we face a more challenging problem of Rich Advertising Auctions, where advertisers provide multiple ads of different sizes and a bid per click. This gives rise to a challenging computational problem of finding the optimal set of ads to fit within a fixed amount of space and a more challenging auction design problem of soliciting true preferences from the advertisers. In this talk, I will describe our recent result that provides a truthful auction that obtains a 3-approximation to the optimal allocation. I will also discuss an extension where with a non-truthful mechanism we prove a 6-approximation in the worst Nash equilibrium.
Bio: Kshipra Bhawalkar is a Research Scientist at Google Research in Mountain View, CA. Her research interest is in Algorithmic Game Theory with applications to online ad auctions and related optimization questions. Kshipra obtained her PhD from Stanford University in 2013 and obtained her bachelors from Duke University.
