Events

Hide Sidebar Show Sidebar
Hide Sidebar Show Sidebar

Current Seminar Series

November 27, 2025

No events scheduled

December 01, 2025

Don’t shout “Bingo!” Understanding (and Addressing) the Shortcomings of Enterprise Threat Detection Products

Adam Bates
University of Illinois at Urbana-Champaign

Part Of

12:00P
- 1:00P

Location

32-G449
Kiva
Add to Calendar 2025-12-01 12:00:00 2025-12-01 13:00:00 America/New_York Don’t shout “Bingo!” Understanding (and Addressing) the Shortcomings of Enterprise Threat Detection Products Abstract: Update -- We are still awful at preventing data breaches and other cybersecurity incidents. Why are these sophisticated (and costly) commercial threat detection products continuing to fail? In this talk, I'll describe our efforts to better understand, and even address, these failure points. First, I'll provide evidence that the extraordinarily high false alarm rates observed in Endpoint Detection & Response (EDR) products can be eliminated by examining the history of alert-triggering processes. Second, I'll explain how the metrics used to evaluate threat detection products often paint a deeply misleading picture of organizations' security readiness. I will conclude by discussing how our ongoing work seeks to resolve industry shortcomings by providing more principled foundations for threat detection and assessment. Bio: Adam Bates is an Associate Professor at the University of Illinois at Urbana-Champaign, where he studies a broad range of topics in computer security. He is best known for his work on data provenance, the practice of examining suspicious activities on computing systems based on their historical context. Fittingly, Adam also appreciates the historical context of computer security research, regularly forcing students in his courses to read James Anderson's 1972 Computer Security Technology and Planning Study… both volumes. Adam is the recipient of two distinguished paper awards (S&P'23, ESORICS'22) and was the runner-up for the ACM SIGSAC Dissertation Award. His research has been recognized and supported by an NSF SaTC FRONTIER, NSF CISE Research Initiation Initiative (CRII), and NSF CAREER Awards, as well as a gift from the VMWare University Research Fund.    TBD

Toward Provable Privacy for Black-Box Algorithms via Algorithmic Stability

Mayuri Sridhar
MIT CSAIL

Part Of

1:00P
- 2:30P

Location

32-G882
Add to Calendar 2025-12-01 13:00:00 2025-12-01 14:30:00 America/New_York Toward Provable Privacy for Black-Box Algorithms via Algorithmic Stability This thesis focuses on enabling the design of algorithms with provable privacy as a first-order goal. We focus on the following setting: an algorithm is trained on sensitive data and the result is then exposed for public use -- how can we quantify the privacy risk of this exposure? Prior work typically focuses on providing privacy through privatizing specific algorithms. These techniques have two main drawbacks: (1) they require significant white-boxing per algorithm and (2) the privacy-utility tradeoffs may be hard to quantify.We first leverage the PAC privacy framework to mitigate the white-boxing requirements. In this talk, we show how we can privatize a wide range of database queries in a black-box manner. We discuss how to build a simple privatization layer, PAC-DB, that can provide provable privacy guarantees for general SQL queries. This allows us to expand the capabilities of private database analytics, enabling complex queries without the use of a trusted curator.We then focus on understanding the utility impacts of privatization. We focus on designing privacy-conscious algorithms. That is, rather than first constructing an algorithm then computing the noise required to privatize it --- a paradigm we refer to as post-hoc privatization --- we optimize the algorithm's hyperparameters given the privacy budget. We instantiate this via regularized linear regression. In particular, we derive the theoretically-optimal regularization weight to maximize utility under a provided privacy budget. We provide experimental results showing the benefits of privacy-conscious design over post-hoc privatization. TBD

December 02, 2025

Enhancing Die-Stacked DRAM Resilience at Scale: The Journey from Research to Industry Standard

Sudhanva Gurumurthi
AMD
11:00A
- 12:00P

Location

32-G882
Add to Calendar 2025-12-02 11:00:00 2025-12-02 12:00:00 America/New_York Enhancing Die-Stacked DRAM Resilience at Scale: The Journey from Research to Industry Standard Talk AbstractJEDEC High Bandwidth Memory (HBM)™ is a widely used DRAM technology in AI and HPC SoCs due to its performance and energy efficiency benefits. Reliability, Availability, and Serviceability (RAS) are additional requirements for SoCs deployed at scale to ensure that computing is reliable and to minimize disruptions for long-running workloads such as AI training and scientific computing.However, improving RAS for HBM is challenging due to a combination of architectural limitations and practical considerations for a standardized solution. In this talk, I shall first provide a primer on DRAM RAS and define the problem statement for HBM RAS. I shall then present the research that was carried out in collaboration with the memory industry to develop an improved HBM resilience architecture. I shall present the data and analyses that drove this work, specific decisions made as we navigated a space of various design options, and the rationale for each decision. The outcome of this effort was a new HBM RAS architecture that was adopted in the third generation HBM standard (HBM3), has been commercialized by DRAM manufacturers, and is now used in GPUs and AI accelerators across industry. This RAS architecture is also included in the fourth generation HBM standard (HBM4) that was announced by JEDEC earlier this year.  Speaker BioSudhanva Gurumurthi is a Fellow at AMD, where he is responsible for research and advanced development in RAS. His work has impacted numerous AMD products, multiple industry standards, and external research in the field. Before joining industry, Sudhanva was an Associate Professor in the Computer Science Department at the University of Virginia. He currently serves as the Editor-in-Chief of IEEE Computer Architecture Letters and on the College of Science and Engineering advisory board at Texas State University.  Sudhanva is the recipient of an NSF CAREER Award, a Google Focused Research Award, and is named to the ISCA Hall of Fame. He received his PhD in Computer Science and Engineering from Penn State in 2005. TBD

Visual Computing Seminar: One String to Pull Them All: Fast Assembly of Curved Structures from Flat Auxetic Linkages

Akib Zaman
CSAIL
12:00P
- 1:00P

Location

32-D463
Star Room
Add to Calendar 2025-12-02 12:00:00 2025-12-02 13:00:00 America/New_York Visual Computing Seminar: One String to Pull Them All: Fast Assembly of Curved Structures from Flat Auxetic Linkages Abstract: We present a computational approach for designing freeform structures that can be rapidly assembled from initially flat configurations by a single string pull. The target structures are decomposed into rigid spatially varied quad tiles that are optimized to approximate the user-provided surface, forming a flat mechanical linkage. Our algorithm then uses a two-step method to find a physically realizable string path that controls only a subset of tiles to smoothly actuate the structure from flat to assembled configuration. We initially compute the minimal subset of tiles that are required to be controlled with the string considering the geometry of the structure and interaction among the tiles. We then find a valid string path through these tiles that minimizes friction, which will assemble the flat linkage into the target 3D structure upon tightening a single string. The resulting designs can be easily manufactured with computational fabrication techniques such as 3D printing, CNC milling, molding, etc. in flat configuration that, in addition to manufacturing, facilitates storage and transportation. We validate our approach by developing a series of physical prototypes and showcasing various application case studies, ranging from medical devices, space shelters, to architectural designs. TBD

CSAIL Forum with Andrew Lo: Quantamental Investing and Generative AI

Andrew Lo
CSAIL, Sloan

Part Of

12:00P
- 1:00P

Location

TBD
via Zoom, registration required
Add to Calendar 2025-12-02 12:00:00 2025-12-02 13:00:00 America/New_York CSAIL Forum with Andrew Lo: Quantamental Investing and Generative AI The convergence of quantitative and fundamental investment styles has been a pipe dream of many asset management companies and hedge funds for decades, but with few if any industrial examples. The rise of generative AI and large language models (LLMs) has dramatically lowered the barriers between these two disparate methods of investing, but several remaining challenges must be overcome before a true rapprochement is realizable. In this talk, Prof. Lo will describe those challenges and map out a process by which quantamental investing may be realized within the next few years.Registration required: https://mit.zoom.us/meeting/register/33v2S1mjSLOQm5HEpCbQBQAndrew Lo biography: https://www.csail.mit.edu/person/andrew-lo TBD

HCI Seminar - Lace Padilla - Seeing the Unknown: Advanced Techniques for Communicating Uncertainty in Data

Lace Padilla
Northeastern University

Part Of

4:00P
- 5:00P

Location

32-D463
Add to Calendar 2025-12-02 16:00:00 2025-12-02 17:00:00 America/New_York HCI Seminar - Lace Padilla - Seeing the Unknown: Advanced Techniques for Communicating Uncertainty in Data Abstract:We live in an uncertain world. From extreme weather hazards to pandemics, forecasters present uncertainty daily. Unfortunately, uncertainty is challenging for both the public and experts to understand, making the effective communication of scientific findings essential. Visualizations can help us interpret complex data by leveraging our visual system’s advanced pattern recognition, allowing us to process vast datasets efficiently. This talk covers cutting-edge uncertainty visualization techniques and cognitive processes that lead to misunderstandings of uncertain forecasts. We'll explore best practices in information visualization to help communicators understand how their choices shape audience perceptions, promoting accessible and ethical communication of future projections.Bio:Lace Padilla joined Northeastern University in 2023 as an Assistant Professor of Computer Science and Psychology. Her work sits at the intersection of information visualization, behavioral decision-making, and HCI.  Her research on uncertainty communication explores how to align data visualizations of future events with human decision-making capabilities. She has received numerous honors, including a Best Paper Award at IEEE VIS, the APA Early Career Award, the NSF CAREER Award, and the IEEE VGTC Significant New Researcher Award. She is also PI or Co-PI on multiple grants funded by NSF (#2122174, #2028374, #1810498, #2400471), NIH (#1R01AI188576-01), and the U.S. Department of Energy.This talk will also be streamed over Zoom: https://mit.zoom.us/j/91341611426. TBD

December 03, 2025

Hybrid Search & Analytic Processing: Building Databases for the AI Era

Mingyu (Rayner) Chen
VeloDB / Apache Doris
1:00P
- 2:00P

Location

32-G882
Add to Calendar 2025-12-03 13:00:00 2025-12-03 14:00:00 America/New_York Hybrid Search & Analytic Processing: Building Databases for the AI Era Abstract: Modern AI applications demand more than traditional analytics—they require real-time, multi-modal retrieval that unifies structured data, text search, and vector semantics. In this talk, we explore the evolution from customer-facing analytics to agent-facing intelligence, and how Hybrid Search & Analytic Processing (HSAP) enables AI systems to reason and act on enterprise data. Using Apache Doris as a case study, we will dive into core capabilities such as real-time analytics, hybrid retrieval, and lakehouse workloads, and discuss how a unified database architecture can power the next generation of AI agents and data-driven applications.Bio: Rayner Chen, Apache Doris PMC Chair & VPE@VeloDB, 10 years of experience in distributed system, focusing on distributed scalable analytical databases. Now primarily overseeing Lakehouse-related development.----Please reach out to markakis@mit.edu for the Zoom password. TBD

December 04, 2025

Will Artificial Intelligence Be the End of Civilization, or the Beginning?

Henry Liebermand and Christopher Fry

Part Of

6:30P
- 8:00P

Location

32-G449
Add to Calendar 2025-12-04 18:30:00 2025-12-04 20:00:00 America/New_York Will Artificial Intelligence Be the End of Civilization, or the Beginning? Boston Chapter of the IEEE Computer Society and GBC/ACM7:00 PM, Thursday, 4 December 2025MIT Room 32-G449 (Kiva) and online via ZoomWill Artificial Intelligence Be the End of Civilization, or the Beginning?Henry Lieberman, MIT Computer Science and Artificial Intelligence Lab + Christopher Fry, MIT Media Lab, Sloan, IBM, startups (Retired)            https://www.whycantwe.orgPlease register in advance for this seminar even if you plan to attend in person athttps://acm-org.zoom.us/webinar/register/8917630641635/WN_FKvNEH5NQAO5nzIM_jWxxw After registering, you will receive a confirmation email containing information about joining the webinar.Indicate on the registration form if you plan to attend in person. This will help us determine whether the room is close to reaching capacity. We plan to serve light refreshments (probably pizza) before the talk starting at around 6:30 pm. Letting us know you will come in person will help us determine how much pizza to order.We may make some auxiliary material such as slides and access to the recording available after the seminar to people who have registered.Abstract:Popular press articles whipsaw the public between two starkly different views of Artificial Intelligence.  On one hand, AI is presented as a magic genie that can solve all of our problems with superhuman intelligence. On the other hand, it's presented as an unprecedented threat to humanity, with the danger of loss of jobs, loss of privacy, automated discrimination, even some kind of "robot rebellion". No wonder the public is confused. Which is it?We present a view that is different from both the self-interested promotion of the tech companies, and from the pessimism of the social critics. Believe it or not, the biggest value of AI will lie, not insimply improving the operations of today's industry and government, but in making it possible to have a more cooperative, less competitive world.Our view is:- Optimistic. Mitigating possible dangers of AI in today's society is important. But we don't want to let fear cause us to miss the potential for AI to tackle big problems people now think are intractable: war, poverty, climate, etc.- Radical. Many tech boosters imagine simply pouring AI into today's economy and electoral politics. We think these systems need to be redesigned from scratch for the AI era. We have two concrete proposals: Makerism (economics) and  Reasonocracy (governance).    - Original. Not conventionally Left or Right, though our ideas share some design goals with both sides. Not (yet) heard on mainstream or activist media. TBD

December 05, 2025

Multi-Key Homomorphic Secret Sharing: From Theory To Practice

Kevin He (MIT) and Lali Devadas (MIT)

Part Of

10:30A
- 12:00P

Location

32-D463
Add to Calendar 2025-12-05 10:30:00 2025-12-05 12:00:00 America/New_York Multi-Key Homomorphic Secret Sharing: From Theory To Practice Homomorphic secret sharing (HSS) enables efficient, low-communication secure computation without the use of fully homomorphic encryption. In all existing HSS schemes, parties participate in a correlated setup phase or a public-key infrastructure, then exchange shares of their inputs and perform local computations to obtain additive shares of the output.In the first part of the talk, we define multi-key homomorphic secret sharing (MKHSS), which replaces the setup in HSS with only a common reference string, and construct MKHSS for NC1 circuits from the decisional composite residuosity (DCR) assumption. This implies the first realization of succinct two-round secure computation for NC1 circuits without lattice-based assumptions.In the second part of the talk, we present optimizations to speed up the MKHSS construction by 45x. Crucial to this speedup is an insight that reduces the largest modulus from N^4 to N^2. As a bonus, we discover a structural simplification that is of independent interest to other HSS schemes.A practical application of MKHSS is a non-interactive conditional key exchange protocol, where two parties obtain the same key only if their inputs satisfy some predicate, which can be an arbitrary branching program. We give practical instantiations for two concrete predicates—geolocation proximity and fuzzy password matching—and achieve a total running time in a few seconds for realistic parameters.Joint works with Geoffroy Couteau (Université Paris Cité, CNRS, IRIF), Srini Devadas (MIT), Aditya Hedge (Johns Hopkins University), Abhishek Jain (Johns Hopkins University and NTT Research), and Sacha Servan-Schreiber (Tinfoil). TBD

December 09, 2025

Visual Computing Seminar: Addressing the Unexpected - Anomaly Detection and AI Safety

Niv Cohen
NYU
12:00P
- 1:00P

Location

TBD
Add to Calendar 2025-12-09 12:00:00 2025-12-09 13:00:00 America/New_York Visual Computing Seminar: Addressing the Unexpected - Anomaly Detection and AI Safety Abstract:While AI models are becoming an ever-increasing part of our lives, our understanding of their behavior in unexpected situations is drifting even further out of reach. This gap poses significant risks to users, model owners, and society at large.In the first part of the talk, I will overview my research on detecting unexpected phenomena with and within deep learning models. Specifically, detecting (i) anomalous samples, (ii) unexpected model behavior, and (iii) unexpected security threats. In the second part of the talk, I will dive into my recent research on a specific type of unexpected security threat: attacks on image watermarks. I will review such attacks and present my recent work toward addressing them. I will conclude with a discussion of future research directions.Bio: Niv Cohen is a postdoctoral researcher at the school of Computer Science & Engineering at New York University. He received his Ph.D. in Computer Science from the Hebrew University in 2024. His research interests include representation learning, computer vision, and AI safety. He is a recipient of the VATAT Scholarship for Outstanding Postdoctoral Fellows in Data Science and the 2024 Blavatnik Prize for Outstanding Israeli Doctoral Students in Computer Science. TBD

[Thesis Defense] Yung-Sung Chuang: "Towards Factual and Trustworthy Large Language Models"

Yung-Sung Chuang
MIT CSAIL

Part Of

3:00P
- 4:00P

Location

45-792
(the big glass room on the 7th floor of building 45)
Add to Calendar 2025-12-09 15:00:00 2025-12-09 16:00:00 America/New_York [Thesis Defense] Yung-Sung Chuang: "Towards Factual and Trustworthy Large Language Models" Thesis Advisor: James GlassThesis Committee: Yoon Kim, Jacob AndreasCalendar Invitation: http://people.csail.mit.edu/yungsung/defense.icsSpeaker's Website: https://yung-sung.github.ioAbstract: Large Language Models (LLMs) have transformed how we interact with information, yet hallucinations, e.g., plausible but factually incorrect outputs, remain a critical barrier to their deployment in high-stakes applications. This thesis presents a comprehensive approach to understanding and mitigating hallucinations across several fundamental dimensions of knowledge in AI systems: parametric, contextual, and attribution knowledge.We identify that hallucinations arise from different failure modes requiring distinct solutions. First, models may fail to leverage parametric knowledge already encoded in their weights. We introduce DoLa (Decoding by Contrasting Layers), which amplifies factual knowledge by dynamically contrasting predictions across transformer layers, improving factuality without training or external knowledge. Second, in retrieval-augmented generation settings, models often fail to properly use provided context. We develop Lookback Lens, which analyzes attention patterns to detect and reduce hallucinations. Third, even when models generate correct content, users need verifiable evidence. We present SelfCite, a self-supervised alignment method that enables LLMs to provide accurate sentence-level citations through a reward design of context ablation. Together, these methods form a roadmap towards better AI systems, working towards systems that are not only capable but also reliable, transparent, and trustworthy. TBD

Can we speed safely?

Ronitt Rubinfeld
CSAIL, EECS

Part Of

4:15P
- 5:15P

Location

32-G449
Refreshments at 4:00 PM
Add to Calendar 2025-12-09 16:15:00 2025-12-09 17:15:00 America/New_York Can we speed safely? Often, algorithmic tasks can be greatly sped up for inputs that are promised to have certain structural properties, such as inputs that are assumed to be random, or to come from restricted classes of graphs. However, in practice, we rarely know if these promises hold, and verifying them can cost more than using a worst case algorithm. This talk surveys emerging lines of work that build trustworthy fast algorithms, by pairing speedups with weaker notions of testability. Our new algorithms, which either give an answer that is certified to be correct, or flag the input as one that does not satisfy the promised conditions, have complexities that are significantly faster than that of algorithms that do not rely on promises. This talk will discuss works that are joint with Arsen Vasilyan, Talya Eden, Cassandra Marcussen, and Madhu Sudan. TBD

December 10, 2025

Which Algorithms Have Tight Generalization Bounds?

Thomas Weinberger
EPFL

Part Of

4:00P
- 5:00P

Location

32-G575
Add to Calendar 2025-12-10 16:00:00 2025-12-10 17:00:00 America/New_York Which Algorithms Have Tight Generalization Bounds? Generalization measures (GMs) are a central tool for providing performance guarantees and informing algorithmic design. Yet, most such bounds are known to be loose (or even vacuous) in practise. In a series of works [1, 2], we focus on GMs in the overparametrized supervised learning setting, where we show that this looseness is unavoidable due to fundamental lower bounds. Notably, these lower bounds hold on average over finite collections of distributions, and with numerically appreciable values.For GMs that are computed solely from the training sample but depend on neither the algorithm nor distribution, we show non-tightness across a large fraction of (algorithm, distribution) combinations.For GMs that can also depend on the algorithm, we show that there can be a trade-off between the algorithm’s ability to learn and the ability to verify the algorithm’s learning success with a GM.Next, we study algorithm-dependent GMs for algorithms that admit a natural notion of algorithmic implicit bias. There, non-tightness of GMs provably occurs whenever the underlying distribution class is rich enough, which is the case for example when learning VC-classes.Lastly, we show that a certain notion of algorithmic stability is sufficient for the existence of tight GMs.Joint work with Ido Nachum (University of Haifa), Jonathan Shafer (MIT), and Michael Gastpar (EPFL).[1]: ICLR 2024, see https://arxiv.org/abs/2309.13658[2]: Neurips 2025 (Spotlight), see https://arxiv.org/abs/2410.01969 TBD

Dertouzos Distinguished Lecture: Yossi Matias, Vice President, Google

Yossi Matias
Google

Part Of

4:00P
- 5:00P

Location

32-G449
Patil/Kiva Seminar Room
Add to Calendar 2025-12-10 16:00:00 2025-12-10 17:00:00 America/New_York Dertouzos Distinguished Lecture: Yossi Matias, Vice President, Google Yossi Matias is Vice President, Google, and the Head of Google Research.Under Yossi’s leadership, world-class global teams are leading breakthrough research on Foundational Machine Learning & Algorithms, Computing Systems & Quantum Computing, Science, AI for Societal Impact in Health, Climate, Sustainability, Education and Cultural and Behavioral Intelligence. Yossi also leads teams advancing foundational research advancements in Generative AI, driving real-world impact and shaping the future of technology.Yossi was previously on Google Search leadership for over a decade, driving strategic features and technologies, and pioneered Conversational AI innovations to help transform the phone experience and help remove barriers of modality and languages. He was also the founding lead of Google center in Israel and supported other global sites. During his tenure at Google Yossi founded and spearheaded initiatives such as Google's AI for Social Good, Crisis Response, Google for Startups Accelerator, social and cultural initiatives seeding Google Arts & Culture, and programs fostering startups, sustainability, and STEM and AI literacy for youth.Prior to Google Yossi was on the Computer Science faculty at Tel Aviv University, a visiting professor at Stanford, and a Research Scientist at Bell Labs. He’s published over 200 papers and is the inventor of over 80 patents. He pioneered some of the early technologies for internet privacy, contextual search, and the effective analysis of Big Data. He is a recipient of the Gödel Prize, an ACM Fellow, and a recipient of the ACM Kanellakis Theory and Practice Award for seminal work on streaming algorithms, data sketches, and large-scale data analyticsYossi has a track record of impact-driven breakthrough research and innovation, and extensive product leadership, transforming products and advancing AI to help address global challenges.  TBD

December 15, 2025

Spectral Graph Neural Networks are Incomplete on Graphs with a Simple Spectrum

Snir Hordan
Technion – Israel Institute of Technology
1:00P
- 2:00P

Location

32-D463
Star
Add to Calendar 2025-12-15 13:00:00 2025-12-15 14:00:00 America/New_York Spectral Graph Neural Networks are Incomplete on Graphs with a Simple Spectrum Spectral features are widely incorporated within Graph Neural Networks (GNNs) to improve their expressive power, or their ability to distinguish among non-isomorphic graphs. One popular example is the usage of graph Laplacian eigenvectors for positional encoding in MPNNs and Graph Transformers. We leverage a well-studied paradigm of classifying graphs by their largest eigenvalue multiplicity to introduce an expressivity hierarchy for SGNNs. We then prove that many SGNNs are incomplete even on graphs with distinct eigenvalues. To mitigate this deficiency, we adapt rotation equivariant neural networks to the graph spectra setting to propose a method to provably improve SGNNs’ expressivity on simple spectrum graphs. TBD

February 10, 2026

February 17, 2026

Visual Computing Seminar: Learning a distance measure from the information-estimation geometry of data

Guy Ohayon
Flatiron Institute
12:00P
- 1:00P

Location

32-D463
Add to Calendar 2026-02-17 12:00:00 2026-02-17 13:00:00 America/New_York Visual Computing Seminar: Learning a distance measure from the information-estimation geometry of data Abstract:The perceptual distance between images is widely believed to be related to the distribution of natural images. But how can a probability distribution give rise to a distance measure—let alone one that aligns with human perception? What properties should such a distance satisfy, and how can it be learned from an image database in an unsupervised manner? In this talk, I will address these questions by presenting the Information–Estimation Metric (IEM), a novel form of distance function derived from a given probability density over a domain of signals. The IEM is rooted in a fundamental relationship between information theory and estimation theory, which links the log-probability of a signal with the errors of an optimal denoiser, applied to noisy observations of the signal. For Gaussian-distributed signals, the IEM coincides with the Mahalanobis distance. But for more complex distributions, it adapts, both locally and globally, to the geometry of the distribution. I will discuss and illustrate the theoretical properties of the IEM—including its global and local behavior. Finally, I will demonstrate that the IEM effectively predicts human perceptual judgments when trained (unsupervised) on natural images.Bio:Guy is a postdoctoral researcher working with Eero Simoncelli at the Flatiron Institute. His research focuses on developing computational models of human perception that are grounded in principles from information theory. He received his PhD in Computer Science from the Technion—Israel Institute of Technology, where he worked with Michael Elad and Tomer Michaeli on the design and theoretical analysis of image restoration and compression methods that rely on generative models. TBD

February 24, 2026

March 03, 2026