Technion
Location: Zisapel 506
Abstract:
Verifying the local robustness of neural networks is crucial for understanding their safety level.
However, the complexity of a complete analysis is exponential in the number of unstable neurons, which introduce nonlinearity.
To scale, many complete verifiers split the verification task into smaller subtasks and select a split by relying on heuristics or learning.
We study the problem of finding optimal splits and phrase it as information gain maximization whose goal is to reduce the number of unstable neurons.
The challenge is that the information gain is defined over probabilities that are intractable to compute.
We present SIGNAL which efficiently computes a split by relying on a differentiable estimate of this information gain.
Our key idea is to extend moment propagation to the setting of local robustness verification.
We prove that our differentiable estimate converges to the true value as the neurons’ input dimensionality grows.
We integrate SIGNAL in the alpha-beta-CROWN verifier.
For fully-connected networks, whose neurons have high input dimensionality, SIGNAL scales prior approaches by at least 1.6x on average.
For the task of computing the largest robust $\epsilon$-ball, within 2 hours, SIGNAL
computes 1.25x larger radius for fully-connected and convolutional networks.
MSc seminar. Supervisor: Dr. Dana Drachsler Cohen
Raissa Nataf
Technion
4/3/2026, 11:30
Location: Taub 601
Distributed asynchronous systems often require explicit synchronization to ensure the correct implementation of shared objects.
Dr. Aviv Yaish
Yale University
On 11/02/2026 at 10:30
Location: 506, Zisapel Building and Zoom
Title: Deconstructing and Rebuilding Trust in Decentralized Economies
Abstract:
Financial systems are becoming increasingly digital and decentralized, demanding a practical fusion of distributed systems security and economic theory. A key enabler of this change, blockchain technology, promises more private and egalitarian economic mechanisms, built by facilitating consensus between pseudonymous actors. However, the theoretical security of these systems may mask significant real-world risks. In this talk, I will present recent advances in bridging this gap between theory and practice. First, I will discuss the resolution of a decade-old puzzle: the lack of observed attacks on major consensus mechanisms. I will then distill the lessons learnt into a holistic approach to designing robust systems and demonstrate its adoption in practice using several lines of work on the economics and security of modern digital markets, tackling problems including denial-of-service resistance in distributed systems and pseudonymous markets where consumers may cheaply create new identities.
Bio:
Aviv is a postdoc at Yale University, where he makes and breaks distributed systems by bridging economic theory and practice. His approach is driven by a philosophy of constructive deconstruction: pushing systems to their limits is key to making them robust. Aviv’s work has been recognized across several fields: security (CCS Distinguished Paper award), economics (CBER Best Paper award), and industry (three prizes from the Ethereum Foundation and Flashbots). He earned his Ph.D. in Computer Science from the Hebrew University (HUJI), where he was the sole lecturer for large-scale courses and won a teaching award. During his studies, he served as a research consultant at Matter Labs. His honors include the AIANI and Jabotinsky fellowships, and inclusion in HUJI’s top 10 CS teaching staff of ‘20, CBER’s Top PhD Graduates of ‘23-‘24, CBER’s Rising Stars of ’25, and HUJI’s 40 Under 40 of ‘25 lists.
Tel-Aviv University
Abstract:
Recent advances in natural language processing (NLP) have given rise to transformative models, including large language models (LLMs) and text retrievers. Still, critical concerns remain regarding the security of these models: chiefly, LLMs can be jailbroken and misused (e.g., to launch cyberattacks), and text retrievers in search applications can be manipulated to prioritize adversary-chosen content. In this talk, I will present our recent efforts toward making LLMs and text retrievers more secure. In particular, I will show how potent attacks can provide explanations for models’ vulnerabilities, which, in turn, enable us to enhance security. Crucially, I will also demonstrate how our insights can inform the design of even stronger attacks, establishing a cycle that guides continuous model improvements.Based on joint work with Matan Ben-Tov and Mor Geva.
Title: ToggleCCI: Dynamic Cost Optimization for Multi-Cloud Transfers
Abstract: Cloud computing is now extremely common, and many organizations operate workloads across multiple providers. As a result, cross-cloud data transfer has become a critical requirement in modern distributed systems. New services such as Google Cross-Cloud Interconnect (CCI) offer dedicated, high-throughput links with low per-GB transfer costs. Yet these benefits come with high fixed leasing fees and a provisioning delay of several days, creating a difficult cost-performance trade-off under unknown traffic patterns.
In this talk, we study the problem of online cross-cloud routing, where decisions must be made without future knowledge. We show that no online algorithm can guarantee constant-factor optimality. To address this, we introduce ToggleCCI, a simple online algorithm that dynamically switches between VPN and CCI based on recent cost observations while accounting for leasing constraints and delays. Using both synthetic workloads and real-world traffic traces, we demonstrate that ToggleCCI consistently tracks the best static strategy and achieves substantial cost savings across a wide range of scenarios.
M.Sc. student under the supervision of Prof. Isaac Keslassy.
Title: Efficient LLM Systems: From Algorithm Design to Deployment
Abstract:
Large Language Models (LLMs) have transformed what machines can do and how systems are designed to serve them. These models are both computationally and memory demanding, revealing the limits of traditional optimization methods that once sufficed for conventional systems. A central challenge in building LLM systems is improving system metrics while ensuring response quality.
This talk presents approaches for reducing latency in LLM systems to support interactive applications, from scheduling algorithm design to deployment. It introduces scheduling frameworks that use lightweight predictions of request behavior to make informed decisions about prioritization and memory management across two core settings: standalone LLM inference and API-augmented LLMs that interact with external tools. Across both settings, prediction-guided scheduling delivers substantial latency reductions while remaining practical for deployment.
Bio:
Rana Shahout is a Postdoctoral Fellow at Harvard University, working with Michael Mitzenmacher and Minlan Yu. She received her Ph.D. in Computer Science from the Technion and previously worked as a Senior Software Engineer at Mellanox (now NVIDIA). Her research combines machine learning, systems, and algorithmic theory to design efficient and scalable AI systems. Rana is a recipient of the Eric and Wendy Schmidt Postdoctoral Award, the Zuckerman Postdoctoral Fellowship, the Weizmann Institute Women’s Postdoctoral Career Development Award, the VATAT Postdoctoral Fellowship, and first place in the ACC Feder Family Award for Best Student Work in Communications.
Columbia University12/1, 12:30
Location: TBDTitle: cache_ext: Customizing the Page Cache with eBPFAbstract:
The OS page cache is central to the performance of many applications, by reducing excessive accesses to storage. However, its one-size-fits-all eviction policy performs poorly in many workloads. While the systems community has experimented with new and adaptive eviction policies in non-OS settings (e.g., key-value stores, CDNs), it is very difficult to implement such policies in the kernel. To address these shortcomings, we design a flexible eBPF-based framework for the Linux page cache, called cache_ext, that allows developers to customize the page cache without modifying the kernel. cache_ext enables applications to customize the page cache policy for their specific needs, while also ensuring that different applications’ policies do not interfere with each other and preserving the page cache’s ability to share memory across different processes. We demonstrate the flexibility of cache_ext’s interface by using it to implement eight different policies, including sophisticated eviction algorithms. Our evaluation shows that it is indeed beneficial for applications to customize the page cache to match their workloads’ unique properties, and that they can achieve up to 70% higher throughput and 58% lower tail latency.
Bio:
Tal Zussman is a PhD student in Computer Science at Columbia University, advised by Prof. Asaf Cidon. He works on operating systems and eBPF, with a focus on accelerating, customizing, and modernizing memory management and storage systems. He received his MS and BS degrees at Columbia, and is an NSF Graduate Research Fellow.