STUDENTS

Postdocs

Neophytos Charalambides

Neophytos’s research interests include coding theory/coded computing, cryptography, machine learning, optimization, randomized numerical linear algebra, and spectral graph theory.

 

Gregory Kehne

Gregory Kehne is a final-year Ph.D. student in the Harvard Computer Science Department, where Ariel Procaccia advises him. He is interested in approximation algorithms and computational social choice. Currently, he is working on problems in social choice, max-min allocation, and online algorithms.

Marina Knittel

Marina is pursuing her Ph.D. in Computer Science at the University of Maryland, College Park, and is expected to commence her postdoctoral research in January 2024.

Her academic interests revolve around developing large-scale graph algorithms, exploring fairness in machine learning, studying computational geometry, and investigating graphical models for economic markets.

Neophytos Charalambides
Gregory Kehne
Marina Knittel


Vasilis Kontonis

Vasilis is a Ph.D. student at the Computer Science Department of the University of Wisconsin-Madison, where he is advised by Professor Christos Tzamos. He is particularly interested in algorithms and theoretical machine learning. Currently, Vasilis is working on problems related to efficient robust inference in high dimensions.

Before coming to UW-Madison, Vasilis studied Electrical and Computer Engineering at the National Technical University of Athens where he was advised by Professor Dimitris Fotakis.

Zehua Lai

Zehua Lai is a Ph.D. student at the Committee on Computational and Applied Mathematics at the University of Chicago. Prior to coming to Chicago, he received his bachelor’s degree in Economics and Mathematics at Tsinghua University. His research interest is to employ techniques in pure and applied mathematics to deepen our understanding of computational and statistical problems in modern machine learning. He is also interested in statistical physics and applied algebraic geometry.

Jessica Sorrell

Jessica is a postdoc at UPenn, working with Aaron Roth and Michael Kearns. She completed her Ph.D. at UCSD, advised by Daniele Micciancio and Russell Impagliazzo. She is broadly interested in the theoretical foundations of responsible computing, particularly questions related to fairness, privacy, robustness, and reproducibility. Jessica also works on problems in lattice-based cryptography, focusing on secure computation.

Vasilis Kontonis
Zehua Lai
Jessica Sorrell

Zhengchao Wan

Zhengchao’s research interests include applications of geometry and probability to problems arising from data analysis with a focus on topological data analysis (TDA) and clustering.

Ingvar Ziemann

Ingvar is interested in statistical learning theory, control theory, and probability. More specifically, his research primarily focuses on how learning algorithms are affected by dynamics and control through phenomena such as slow mixing or distribution shift. Current projects include among other things trying to characterize the instance-specific complexity of specific learning problems in the context of slowly mixing time series and specific applications to meta-learning.
 
Zhengchao Wan
Ingvar Ziemann
Akbar Rafiey
Jonathan Shi
Zhengchao Wan

Graduate Students

Tristan Brugère

Tristan is a graduate student working with Pr. Yusu Wang. His main research interests lie in machine learning theory, and my work is in machine learning on graph data, optimal transport, and applications to chip design.

Hadley Black

Hadley is broadly interested in Theoretical Computer Science, Combinatorics, and Probability Theory. His research focuses on property testing and learning theory.

Samantha Chen

Samantha is a graduate student working with Prof. Yusu Wang. Her research interests include Wasserstein distances, metric learning, approximation algorithms, and neural networks.

Tristan Brugère
Hadley Black
Samantha Chen

Dimitrios Christou

Dimitrios (aka Dimitris) Christou is a Computer Science Ph.D. student at UT Austin, where Prof. Shuchi Chawla advised him. Prior to joining UT Austin, he received a Diploma in Electrical and Computer Engineering from the National Technical University of Athens. He has completed two research internships offered by the LIP6 Research Institute at Sorbonne University, France.

His research interests lie in the intersection of Online Algorithms and Data-Driven algorithmic design. His previous work resolves open questions for online decision-making, server problems, and multistage optimization. His current work focuses on data-driven algorithmic design and learning-augmented algorithms, aiming to breach the gap between traditional theory and modern data-oriented methods.

 

Ryan Ellin

Broadly, Rudrajit is interested in developing provably better optimization algorithms and generalization-improving techniques for machine learning. More specifically, his research interests include large-scale optimization, federated learning, differentially private training, and methods for improving learning in the presence of weak/imperfect supervision.

Ryan Ellin

Ryan’s research interests include differential privacy and more broadly, investigating generalization properties of randomized algorithms.

Dimitrios Christou
Rudrajit Das
Ryan Ellin

Harsh Harshvardhan

Harsh is interested in improving the theoretical understanding of optimization algorithms, with a focus on those encountered in large-scale, heterogeneous, and distributed learning.

 

Akhil Jalan

Akhil’s research interests include statistics, system identification, controls, and cellular agriculture.

Syamantak Kumar

Syamantak is a graduate student in the Computer Science Department at UT Austin, working with Prof. Purnamrita Sarkar.
 
His research interests are broadly at the intersection of statistics and machine learning. He is currently working on streaming algorithms for dependent data streams. 
 
Harsh Harshvardhan
Akhil Jalan
Syamantak Kumar

Donghwan Lee

Donghwan’s research includes uncertainty quantification and high-dimensional analysis in machine learning.

Georgy Noarov

Namiko is a second-year Ph.D. student advised by Prof. Arya Mazumdar. Her research interests are signal reconstruction, high-dimensional statistical inference, and machine learning, emphasizing statistical-computational tradeoffs, phase transitions, sparsity, and nonconvex optimization. Namiko has done work in one-bit compressed sensing.

Georgy Noarov

Georgy is broadly interested in theoretical computer science and machine learning. His current research focuses on fair/robust/trustworthy ML and (distribution-free) uncertainty quantification.

He is especially interested in developing and studying algorithmic fairness and robustness notions that enhance the performance of the underlying ML model, rather than detract from it. He also works on algorithmic game theory and online learning.

Donghwan Lee
Namiko Matsumoto
Georgy Noarov

Rojin Rezvan

Rojin is a fourth-year Ph.D. student at the University of Texas at Austin, advised by Shuchi Chawla. She received her master’s degree from the University of Wisconsin-Madison

She is broadly interested in algorithmic game theory and mechanism design. More specifically, she has done research in multi-dimensional mechanism design in the paradigm of simple vs. optimal, fairness in auctions and fair allocation. Rojin is generally interested in the intersection of mechanism design and other fields such as fairness and decentralized systems.

Geelon So

Geelon So is a Ph.D. Student in Computer Science at UC San Diego, advised by Sanjoy Dasgupta and Yian Ma. His interests are in algorithmic statistics, unsupervised learning, and continual learning.

Yimeng (Kobe) Wang

Yimeng is a Ph.D. student from UCLA advised by Professor Raghu Meka. He is broadly interested in theoretical computer science, especially in Algorithmic Machine Learning and Computational Complexity. 

Rojin Rezvan
Geelon So
Yimeng Wang

Xinyue Xia

Xinyue’s research interest lies in learning with graph-structured data. In particular, she is interested in developing a well-grounded graph machine-learning framework for graph representation learning and graph generative modeling. 
 

Chris Ye

Chris’s research interests include fine-grained complexity, complexity theory, and learning theory.

Heng Zhu

Heng Zhu is a second-year Ph.D. student working with Prof. Arya Mazumdar. His research interests include federated learning, distributed optimization, and statistical learning. 
 
Xinyue Xia
Chris Ye
Heng Zhu

EnCORE Alumni

Sainyam Galhotra

Sainyam is joining Cornell University as an Assistant Professor in the Department of Computer Science. His research aims to develop data discovery and integration tools for responsible and effective analytics. Sainyam received his PH.D. from the University of Massachusetts Amherst under the supervision of Barna Saha.
 
He completed his undergraduate studies at the Indian Institute of Technology Delhi (IIT Delhi) in May 2014 under the guidance of Prof. Amitabha Bagchi. Prior to joining UMass, Sainyam worked as a budding scientist at Xerox Research Centre India, Bangalore for a year.
 

Venkata Gandhikota

Ventaka is an Assistant Professor at the Electrical Engineering and Computer Science Department at Syracuse University. In his work, Venkata explores the theoretical aspects of computing with a focus on coding theory and lattices. Venkata is interested in investigating combinatorial designs to develop provably efficient algorithms for various computational tasks. A part of his recent research focuses on applying the developed techniques to devise robust algorithms for various distributed learning tasks. 

His research interests include the foundations of machine learning, algorithms for big data, and error-correcting codes and lattices.

Abolfazl Hashemi

Abolfazl is an Assistant Professor at The Elmore Family School of Electrical and Computer Engineering at Purdue University and the director of MINDS Group. Previously, he was a Postdoctoral scholar at the Oden Institute for Computational Engineering and Sciences at the University of Texas at Austin. Abolfazl received his Ph.D. and MSE degrees in Electrical and Computer Engineering department at UT Austin in 2020 and 2016, and his BS degree in Electrical Engineering from Sharif University of Technology in 2014.

Abolfazl’s research interests include the study of Machine Learning, Online Learning, Learning in Games, and Networked/Distributed Decision-Making from the perspective of optimization. In doing so, algorithms with efficient designs and mathematical guarantees to render practical deployment of learning-based systems possible under a variety of considerations such as limited resources, and adversarial attacks

 
Sainyam Galhotra
Venkata Gandhikota
Abolfazl Hashemi

Dominik Kempa

Dominik is an assistant professor in the Department of Computer Science at Stony Brook University. Prior to joining Stony Brook, he was a postdoc at Johns Hopkins University hosted by Ben Langmead, and before that at UC Berkeley (Theory group), advised by Barna Saha.

During his Ph.D. at the University of Helsinki, where he was advised by Juha Kärkkäinen and Esko Ukkonen, Dominik implemented a collection of parallel and external-memory algorithms on strings. Earlier, Dominik was an intern at Google, Zürich. He is a recipient of the Junior Researcher Award and Outstanding Doctoral Dissertation Award.

Tomasz Kociumaka

Tomasz Kociumaka is currently a postdoc at Max Planck University. Previously he was a postdoc at UC Berkeley mentored by Barna Saha. Tomasz received his Ph.D. from the University of Warsaw. He is an expert in String algorithms.

Andrea Lincoln

Andrea Lincoln is an incoming Assistant Professor at Boston University. Previously she was a postdoc at UC Berkeley mentored by Barna Saha, and even before that, she finished her Ph.D. at MIT.

She hopes to build and refine theoretical models that crystallize concerns about complex systems. Andrea strives to build networks of reductions that give shared explanations for the hardness of problems. Her research interests include Fine-Grained Complexity, Average-Case Complexity, Lower Bounds, and Algorithms.

Dominik Kempa
Tomasz Kociumaka
Andrea Lincoln

Seth Neel

Seth is an Assistant Professor at Harvard housed in the Department of Technology and Operations Management (TOM) at HBS. He is the Co-Principal Investigator of the Secure and Fair Machine Learning Lab (SAFR ML) in Harvard’s new D^3 Institute, and part of the Theory of Computation research group, and the AI@Harvard Initiative. His primary academic interest is in machine learning, particularly as it concerns ethical notions like fairness and (differential) privacy. More generally, Seth is interested in all mathematical aspects of artificial intelligence. Recently, he lead the integration of algorithms he co-developed during my Ph.D. into the open-source efforts of IBM AI Research

Seth Neel

Product Manifold Learning with Independent Coordinate Selection

Jesse He (UCSD)

Abstract: In many dimensionality reduction tasks, we wish to identify the constituent components that explain our observations. For manifold learning, this can be formalized as factoring a Riemannian product manifold. Recovering this factorization, however, may suffer from certain difficulties in practice, especially when data is sparse or noisy, or when one factor is distorted by the other. To address these limitations, we propose identifying non-redundant coordinates on the product manifold before applying product manifold learning to identify which coordinates correspond to different factor manifolds.

Tear and Repulsion Enabled Registration of Point Clouds for Manifold Learning

Dhruv Kohli (UCSD)

Abstract: We present a framework for aligning the local views of a possibly closed/non-orientable data manifold to produce an embedding in its intrinsic dimension through tearing. Through a spectral coloring scheme, we render the embeddings of the points across the tear with matching colors, enabling a visual recovery of the topology of the data manifold. The embedding is further equipped with a tear-aware metric that enables computation of shortest paths while accounting for the tear. To measure the quality of an embedding, we propose two Lipschitz-type notions of global distortion—a stronger and a weaker one—along with their pointwise counterparts for a finer assessment of the embedding. Subsequently, we bound them using the distortion of the local views and the alignment error between them. We show that our theoretical result on strong distortion leads to a new perspective on the need for a repulsion term in manifold learning objectives. As a result, we enhance our alignment approach by incorporating repulsion. Finally, we compare various strategies for the tear and repulsion enabled alignment, with regard to their speed of convergence and the quality of the embeddings produced.


This is joint work with my advisors Gal Mishne and Alex Cloninger at UCSD.

Efficient Online Clustering with Moving Costs

Dimitrios Christou (UT Austin)

Abstract: In this talk, I will consider an online-learning problem, called Online Clustering with Moving Costs, at which a learner maintains a set of facilities over rounds so as to minimize the connection cost of an adversarially selected sequence of clients. The learner is informed on the positions of the clients at each round only after its facility-selection and can use this information to update its decision in the next round. However, updating the facility positions comes with an additional moving cost based on the moving distance of the facilities. I will be presenting the first (polynomial-time) approximate-regret algorithm for this setting through a combination of different algorithmic techniques such as HST embeddings, the FTRL framework with a dilated entropic regulariser as well as a novel rounding scheme.

Encoding Structural Symmetry is Key for Length Generalization in Arithmetic Tasks

Mahdi Sabbaghi (UPenn)

Abstract: Despite the success of Transformers on language understanding, code generation, and logical reasoning, they still fail to (length) generalize on basic arithmetic tasks such as addition and multiplication. A major reason behind this failure is the vast difference in structure of numbers versus text; for example, numbers are associated with a specific significance order that plays a role in calculating the answer. In contrast, for text, such symmetries are quite unnatural. In this work, we propose to encode these semantics explicitly into the model via appropriate data formatting and custom positional encodings. To further elucidate the importance of explicitly encoding structure, in a simplified linear setting, we prove that standard positional encodings even when trained with augmentations to implicitly induce structure fail at such generalization, whereas enforcing structure via positional encodings succeeds.

Bio: Mahdi Sabbaghi is a second year PhD student at UPenn, department of Electrical and System Engineering, supervised by Professors Hamed Hassani and George Pappas. Previously he obtained a B. Sc. degree in Electrical Engineering as well as a B. Sc. degree in Physics from the Sharif University of Technology, in Tehran.

Private Estimation of U Statistics

Shourya Pandey (UT Austin)

Abstract: We consider the problem of private estimation of U statistics. U statistics are widely used estimators that naturally arise in a broad class of problems, from nonparametric signed rank tests to subgraph counts in random networks. Despite the recent outpouring of interest in private mean estimation, private algorithms for more general U statistics have received little attention.  We propose a framework where, for a broad class of U statistics, one can use existing tools in private mean estimation to obtain confidence intervals where the private error does not overwhelm the irreducible error resulting from the variance of the U statistics. However, in specific cases that arise when the U statistics degenerate or have vanishing moments, the private error may be of a larger order than the non-private error. To remedy this, we propose a new thresholding-based approach that uses Hajek projections to re-weight different subsets. As we show,  this leads to more accurate inference in certain settings.

Finite-Time Logarithmic Bayes Regret Upper Bounds

Alexia Atsidakou (UT Austin)

Abstract: We derive the first finite-time logarithmic Bayes regret upper bounds for Bayesian bandits, for BayesUCB and Thompson Sampling. In Gaussian and Bernoulli multi-armed bandits, we obtain $O(c_\Delta \log n)$ and $O(c_h \log^2 n)$ upper bounds for an upper confidence bound algorithm, where $c_h$ and $c_\Delta$ are constants depending on the prior distribution and the gaps of bandit instances sampled from it, respectively. The latter bound asymptotically matches the lower bound of Lai (1987). Our proofs are a major technical departure from prior works, while being simple and general. The key idea in our proofs is to conduct a frequentist per-instance analysis with Bayesian confidence intervals, and then integrate it over the prior.

Our results provide insights on the value of prior in the Bayesian setting, both in the objective and as a side information given to the learner. They significantly improve upon existing $\tilde{O}(\sqrt{n})$ bounds, which have become standard in the literature despite the logarithmic lower bound of Lai (1987).

Pareto-Optimal Algorithms for Learning in Games

Natalie Collina and Eshwar Ram Arunachaleswaran

Abstract: We study the problem of characterizing optimal learning algorithms for playing repeated games against an adversary with unknown payoffs. In this problem, the first player (called the learner) commits to a learning algorithm against a second player (called the optimizer), and the optimizer best-responds by choosing the optimal dynamic strategy for their (unknown but well-defined) payoff. Classic learning algorithms (such as no-regret algorithms) provide some counterfactual guarantees for the learner, but might perform much more poorly than other learning algorithms against particular optimizer payoffs.

In this paper, we introduce the notion of asymptotically Pareto-optimal learning algorithms. Intuitively, if a learning algorithm is Pareto-optimal, then there is no other algorithm which performs asymptotically at least as well against all optimizers and performs strictly better (by at least $\Omega(T)$) against some optimizer. We show that well-known no-regret algorithms such as Multiplicative Weights and Follow The Regularized Leader are Pareto-dominated. However, while no-regret is not enough to ensure Pareto-optimality, we show that a strictly stronger property, no-swap-regret, is a sufficient condition for Pareto-optimality.

Proving these results requires us to address various technical challenges specific to repeated play, including the fact that there is no simple characterization of how optimizers who are rational in the long-term best-respond against a learning algorithm over multiple rounds of play. To address this, we introduce the idea of the asymptotic menu of a learning algorithm: the convex closure of all correlated distributions over strategy profiles that are asymptotically implementable by an adversary. Interestingly, we show that all no-swap-regret algorithms share the same asymptotic menu, implying that all no-swap-regret algorithms are “strategically equivalent”.

This talk is based on work with Jon Schneider.

Metric Learning from Lazy Crowds

Geelon So

Abstract: We consider crowd-based metric learning from preference comparisons, where given two items, a user prefers the item that is closer to their latent ideal item. Here, “closeness” is measured with respect to a shared but unknown Mahalanobis distance. Can we recover this distance when we can only obtain very few responses per user?

In this very low-budget regime, we show that generally, nothing at all about the metric is revealed, even with infinitely many users. But when the items have subspace cluster structure, we present a divide-and-conquer approach for metric recovery, and provide theoretical recovery guarantees and empirical validation.

This is joint work with Zhi Wang (UCSD) and Ramya Korlakai Vinayak (UW–Madison).

Random Walks, Conductance, and Resistance for the Connection Graph Laplacian

Zhengchao Wan

Abstract: We investigate the concept of effective resistance in connection graphs, expanding its traditional application from undirected graphs. We propose a robust definition of effective resistance in connection graphs by focusing on the duality of Dirichlet-type and Poisson-type problems on connection graphs. Additionally, we delve into random walks, taking into account both node transitions and vector rotations. This approach introduces novel concepts of effective conductance and resistance matrices for connection graphs, capturing mean rotation matrices corresponding to random walk transitions. Thereby, it provides new theoretical insights for network analysis and optimization.

This is based on a joint work with Alexander Cloninger, Gal Mishne, Andreas Oslandsbotn, Sawyer Jack Robertson and Yusu Wang.

Approximability and Inapproximability of Strict-CSPs

Akbar Rafiey

Abstract: We study the approximability and inapproximability of Strict-CSPs. An instance of the Strict-CSPs consists of a set of constraints over a set of variables and a cost function over the assignments. The goal is to find an assignment to the variables of minimum cost which satisfies all the constraints. Some prominent problems that this framework captures are (Hypergraph) Vertex Cover, Min Sum k-Coloring, Multiway Cut, Min Ones, and others.

We focus on a systematic study of Strict-CSPs of the form Strict-CSPs(H), that is, Strict-CSPs where the type of constraints is limited to predicates from a set H. Our first result is a dichotomy for approximation of Strict-CSPs(H), where H is a binary predicate, i.e., a digraph. We prove that if digraph H has bounded width, then Strict-CSPs(H) is approximable within a constant factor (depending on H); otherwise, there is no approximation for Strict-CSPs(H) unless P=NP.

Second, we study the inapproximability of Strict-CSP and present the first general hardness of approximation for Strict-CSP. More precisely, we prove a dichotomy theorem that states every instance of Strict-CSP(H) (H being a digraph) is either polynomial-time solvable or APX-complete. Moreover, we show the existence of a universal constant 0<\delta<1 such that it is NP-hard to approximate Strict-CSP(H) within a factor of (2-\delta) for all digraphs H where Strict-CSP(H) is NP-complete.

Buy-many Mechanisms for Many Unit-demand Buyers

Rojin Rezvan

Abstract: A recent line of research has established a novel desideratum for designing approximatelyrevenue-optimal multi-item mechanisms, namely the buy-many constraint. Under this constraint, prices for different allocations made by the mechanism must be subadditive implying that the price of a bundle cannot exceed the sum of prices of individual items it contains. This natural constraint has enabled several positive results in multi-item mechanism design bypassing well-established impossibility results. Our work addresses a main open question from this literature involving the design of buymany mechanisms for multiple buyers. Our main result is that a simple sequential item pricing mechanism with buyer-specific prices can achieve an O(log m) approximation to the revenue of any buy-many mechanism when all buyers have unit-demand preferences over m items. This is the best possible as it directly matches the previous results for the single-buyer setting where no simple mechanism can obtain a better approximation. Our result applies in full generality: even though there are many alternative ways buy-many mechanisms can be defined for multibuyer settings, our result captures all of them at the same time. We achieve this by directly competing with a more permissive upper-bound on the buy-many revenue, obtained via an ex-ante relaxation.

Streaming PCA for Markovian Data

Syamantak Kumar

Abstract: Since its inception in 1982, Oja’s algorithm has become an established method for streaming principle component analysis (PCA). We study the problem of streaming PCA, where the data-points are sampled from an irreducible, aperiodic, and reversible Markov chain. Our goal is to estimate the top eigenvector of the unknown covariance matrix of the stationary distribution. This setting has implications in scenarios where data can solely be sampled from a Markov Chain Monte Carlo (MCMC) type algorithm, and the objective is to perform inference on parameters of the stationary distribution. Most convergence guarantees for Oja’s algorithm in the literature assume that the data-points are sampled IID. For data streams with Markovian dependence, one typically downsamples the data to get a “nearly” independent data stream. In this paper, we obtain the first sharp rate for Oja’s algorithm on the entire data, where we remove the logarithmic dependence on the sample size, resulting from throwing data away in downsampling strategies.

A d^{1/2 + o(1)} Monotonicity Tester for Boolean Functions on d-Dimensional Hypergrids

Hadley Black

Abstract: Monotonicity testing of Boolean functions over the n-width, d-dimensional hypergrid is a classic problem in property testing, where the goal is to design a randomized algorithm which can distinguish monotone functions from those which are far from any monotone function while making as few queries as possible. The special case of n = 2 corresponds to the hypercube domain. Here a long line of works exploiting a very interesting connection with isoperimetric inequalities for Boolean functions culminated in a non-adaptive tester making ~O(d^{1/2}) queries in a celebrated paper by Khot, Minzer, and Safra (SICOMP 2018). This is known to be optimal for non-adaptive testers. However, the general case of hypergrids for n > 2 remained open. Very recently, two papers (Black-Chakrabarty-Seshadhri STOC 2023 and Braverman-Khot-Kindler-Minzer ITCS 2023) independently obtained ~O(poly(n) d^{1/2}) query testers for hypergrids. These results are essentially optimal for n < polylog(d), but are far from optimal for n >> polylog(d).

This talk covers our most recent result (appearing at FOCS 2023) which obtains a non-adaptive d^{1/2+o(1)} query tester for all n, resolving the non-adaptive monotonicity testing problem for hypergrids, up to a factor of d^{o(1)}. Our proof relies on many new techniques as well as two key theorems which we proved in earlier works from SODA 2020 and STOC 2023.

SmoothLLMs: Defending LLMs against Adversarial Attacks

Alex Robey

Abstract: Despite efforts to align large language models (LLMs) with human values, widely-used LLMs such as GPT, Llama, Claude, and PaLM are susceptible to jailbreaking attacks, wherein an adversary fools a targeted LLM into generating objectionable content.  To address this vulnerability, we propose SmoothLLM, the first algorithm designed to mitigate jailbreaking attacks on LLMs.  Based on our finding that adversarially-generated prompts are brittle to character-level changes, our defense first randomly perturbs multiple copies of a given input prompt, and then aggregates the corresponding predictions to detect adversarial inputs.  SmoothLLM reduces the attack success rate on numerous popular LLMs to below one percentage point, avoids unnecessary conservatism, and admits provable guarantees on attack mitigation.  Moreover, our defense uses exponentially fewer queries than existing attacks and is compatible with any LLM.

Composition of Nested Embeddings with an Application to Outlier Removal

Kristin Sheridan

Abstract: We study the design of embeddings into Euclidean space with outliers. Given a metric space $(X,d)$ and an integer $k$, the goal is to embed all but $k$ points in $X$ (called the “”outliers””) into $\ell_2$ with the smallest possible distortion $c$. Finding the optimal distortion $c$ for a given outlier set size $k$, or alternately the smallest $k$ for a given target distortion $c$ are both NP-hard problems. In fact, it is UGC-hard to approximate $k$ to within a factor smaller than $2$ even when the metric sans outliers is isometrically embeddable into $\ell_2$. We consider bi-criteria approximations. Our main result is a polynomial time algorithm that approximates the outlier set size to within an $O(\log^2 k)$ factor and the distortion to within a constant factor.

The main technical component in our result is an approach for constructing a composition of two given embeddings from subsets of $X$ into $\ell_2$ which inherits the distortions of each to within small multiplicative factors. Specifically, given a low $c_S$ distortion embedding from $S\subset X$ into $\ell_2$ and a high(er) $c_X$ distortion embedding from the entire set $X$ into $\ell_2$, we construct a single embedding that achieves the same  distortion $c_S$ over pairs of points in $S$ and an expansion of at most $O(\log k)\cdot c_X$ over the remaining pairs of points, where $k=|X\setminus S|$. Our composition theorem extends to embeddings into arbitrary $\ell_p$ metrics for $p\ge 1$, and may be of independent interest. While unions of embeddings over disjoint sets have been studied previously, to our knowledge, this is the first work to consider compositions of {\em nested} embeddings.

Graph Sparsification by Approximate Matrix Multiplication

Neo Charalambides

Abstract:  Graphs arising in statistical problems, signal processing, large networks, combinatorial optimization, and data analysis are often dense, which causes both computational and storage bottlenecks. One way of sparsifying a weighted graph, while sharing the same vertices as the original graph but reducing the number of edges, is through spectral sparsification. We study this problem through the perspective of RandNLA. Specifically, we utilize randomized matrix multiplication to give a clean and simple analysis of how sampling according to edge weights gives a spectral approximation to graph Laplacians, without requiring spectral information. Through the CR−MM algorithm, we attain a simple and computationally efficient sparsifier whose resulting Laplacian estimate is unbiased and of minimum variance. Furthermore, we define a new notion of additive spectral sparsifiers, which has not been considered in the literature.

Talk Title

Student Name

Abstract