CS RESEARCH DAY 2022

Wednesday, 28.12.2022, 12:30

CS Taub Lobby

The 10th CS Research Day for graduate studies will be held on Wednesday, December 28, 2022 between 12:30-14:30, at the lobby of the CS Taub Building.
Research Day events are opportunity for our graduate students to expose their researches using posters and presentations to CS faculty and all degrees students, Technion distinguished representatives and to high-ranking delegates from the hi-tech leading industry companies in Israel and abroad.
The participating researches wil...

[ Full version ]

[ Full version ]

CEClub: Finding darkness in the Light: Recovering Speech and Cryptographic keys from Light Emitted from Power LEDs and Light Bulbs

Ben Nassi (Ben-Gurion University)

Wednesday, 28.12.2022, 11:30

Room 861, EE Meyer Building & Zoom Lecture: 94673013539

In this talk, I will present a journey that started three years ago at the intersection between light leakage and information confidentiality. In the first part of the talk, I will present the topic of electro-optical speech eavesdropping which is based on three methods we developed to recover speech from light emitted from light bulbs (Lamphone USENIX Security 22), power LEDs of speakers (Glowworm Attack CCS 21), and from light reflected from shiny ornaments and objects (The Litt...

[ Full version ]

[ Full version ]

CS Colloquia: Constructing and deConstructing Trust in ML: A new Role for Cryptography Today

Shafi Goldwasser (UC Berkeley)

Tuesday, 27.12.2022, 11:30

Room 337 Taub Bld.

For decades now cryptographic tools and models have at their essence transformed technology platforms controlled by worst case adversaries to trustworthy platforms. In this talk I will describe how to use cryptographic tools and cryptographic modeling to build trust in various phases of the machine learning pipelines. We will touch on privacy in the training and inference stage, verification protocols for the quality of machine learning models, and robustness in presence of advers...

[ Full version ]

[ Full version ]

CS Lecture: Reinforcement Learning in the Presence of Irrelevant Information

Yonathan Efroni (Meta, New York)

Tuesday, 20.12.2022, 10:30

Taub 601

Reinforcement Learning (RL) is a field concerned with designing general purpose learning algorithms that solve sequential-decision tasks. In recent years, by using deep neural networks, RL algorithms were applied on high-dimensional and challenging domains, witnessing unprecedented success. Yet, despite recent advancements, the theoretical foundations of high-dimensional RL are not fully understood.
A recurring theme in high-dimensional RL is the presence of irrelevant informa...

[ Full version ]

[ Full version ]

Coding Theory: Association Schemes and Injection Codes

Nathan Lindzey (CS, Technion)

Sunday, 18.12.2022, 14:30

Taub 601

The first half of the talk will overview Delsarte's association scheme-theoretic approach to coding theory, which has been used extensively over the past few decades to obtain bounds on many different classes of codes. For example, the Hamming scheme and the Johnson scheme both have classical roles in the theory of linear codes and constant-weight codes respectively, whereas the permutation scheme has been used recently in the study of permutation codes, a less well-known class o...

[ Full version ]

[ Full version ]

Reasons for the Superiority of Stochastic Estimators over Deterministic Ones: Robustness, Consistency and Perceptual Quality

Stochastic restoration algorithms allow to explore the space of solutions that correspond to the degraded input. In this paper we reveal additional fundamental advantages of stochastic methods over deterministic ones, which further motivate their use. First, we prove that any restoration algorithm that attains perfect perceptual quality and whose outputs are consistent with the input must be a posterior sampler, and is thus required to be stochastic. Second, we illustrate that whi...

[ Full version ]

[ Full version ]

Pixel Club: Improving Computational Imaging Systems through Deep Learning and Optimization

Wolfgang Heidrich (KAUST Visual Computing Center)

Wednesday, 14.12.2022, 13:30

Room 815, EE Meyer Building

Computational imaging systems are based on the joint design of optics and associated image reconstruction algorithms. Historically, many such systems have employed simple transform-based reconstruction methods. Modern optimization methods and priors can drastically improve the reconstruction quality in computational imaging systems. Furthermore, learning-based methods can be used to design the optics along with the reconstruction method, yielding truly end-to-end optimized imaging...

[ Full version ]

[ Full version ]

Theory Seminar: Error Correcting codes that Achieve BSC Capacity Against Channels that are Poly-Size Circuits

Ronen Shaltiel (Haifa university)

Wednesday, 14.12.2022, 12:30

Taub 201

Guruswami and Smith (J. ACM 2016) considered codes for channels that are computationally bounded and flip at most a p-fraction of the bits of the codeword. This class of channels is significantly stronger than Shannon’s binary symmetric channel (which flips each bit independently with probability p) but weaker than Hamming’s channels (which may flip at most a p-fraction of the bits, and are computationally unbounded).
The goal of this area is to construct codes against chan...

[ Full version ]

[ Full version ]

Direct Access to Answers of Conjunctive Queries with Aggregation

We study the fine-grained complexity of conjunctive queries with grouping and aggregation. For some common aggregate functions (e.g., min, max, sum), such a query can be phrased as an ordinary conjunctive query over a database annotated with a suitable commutative semiring. Specifically, we study the ability to evaluate such queries by constructing in log-linear time a data structure that provides logarithmic-time direct access to the answers ordered by a given lexicographic orde...

[ Full version ]

[ Full version ]

CS Colloquia: Notions of Simplicity In Deep Learning: From Time Series to Images

Amir Globerson (Tel-Aviv university)

Tuesday, 13.12.2022, 14:30

Room 337 Taub Bld.

It is standard practice in deep learning to train large models on relatively small datasets. This can potentially lead to severe overfitting, but more often than not, test error is actually good. This phenomenon has prompted research on the so-called "Implicit Bias of Deep Learning Algorithms". Here I will discuss our recent works on multiple novel facets of this bias, and present theoretical and empirical results in different settings. In particular, I will discuss analysis of im...

[ Full version ]

[ Full version ]

Patient-level Microsatellite Status Assessment from Whole Slide Images By Combining Momentum Contrast Learning and Group Patch Embeddings

Assessing microsatellite stability status of a patient's colorectal cancer is crucial in personalizing treatment regime. Recently, convolutional-neural-networks (CNN) combined with transfer-learning approaches were proposed to circumvent traditional laboratory testing for determining microsatellite status from hematoxylin and eosin stained biopsy whole slide images (WSI). However, the high resolution of WSI practically prevent direct classification of the entire WSI. Current appro...

[ Full version ]

[ Full version ]

CS Colloquia: Formal Methods for a Robust Network Ecosystem

George Varghese (UCLA)

Monday, 12.12.2022, 14:30

Room 1003, EE Meyer Building

Network verification, applying formal methods to verify properties of router configurations, is already mainstream with startups like Forward and Veriflow Networks, and divisions in established companies such as Amazon’s ARG and Cisco’s Candid. In this talk, I will survey what remains to be done including: extending formal methods to implementations, and to other parts of the network ecosystem besides routing. I will illustrate these points with recent work we have done on i...

[ Full version ]

[ Full version ]

Coding Theory: Reed-Solomon Codes Against Adversarial Insertions and Deletions

Roni Con (Tel-Aviv University)

Sunday, 11.12.2022, 14:30

Taub 601

The task of constructing codes against adversarial insertions and deletions (insdel) has recently received much attention.
In this work, we study the performance of Reed-Solomon codes against insdel errors. We prove that there are Reed-Solomon codes that achieve the half-Singleton bound. In other words, there are optimal Reed-Solomon codes also against insdel errors. We also give a set of evaluation points that define a Reed-Solomon code that achieves this bound. As the field s...

[ Full version ]

[ Full version ]

Personal Branding Workshop: How to do it Right?

Wednesday, 07.12.2022, 18:30

Room 337 Taub Bld.

You are invited to a workshop by Raviv Gortenstein (Director of People Brand & Experience at Riskified) who will present practical tools for building a personal brand and how it can help in building a career, expanding the networkng, finding opportunities and positioning expertise in the personal field.
The workshop will be held on Wednesday, December 7, 2022, 18:30, in Taub 337.
Please ...

[ Full version ]

[ Full version ]

Theory Seminar: Streaming Algorithms for Submodular Maximizatio with a Cardinality Constraint

Moran Feldman (Haifa university)

Wednesday, 07.12.2022, 13:30

Taub 201

Motivated by machine learning applications, much research over the last decade was devoted to solving submodular maximization problems under Big Data computational models. Perhaps the most basic such problem is the problem of maximizing a submodular function subject to a cardinality constraint. A recent series of papers has studied this problem in the data stream model, and in particular, fully determined the approximation ratio that can be obtained for it by (semi-)-streaming alg...

[ Full version ]

[ Full version ]

Recruitment Day by Biosense Webster

Wednesday, 07.12.2022, 12:30

CS Taub Lobby & Taub 3

Engineers and recruitment teams from Biosense Webster will visit CS to offer employment options and open positions, and to lecture on the connection between medicine and computer science.
Wednesday, December 7, 2022, in the lobby and Taub 3 at CS Taub Building:
12:30 in Taub Lobby - meeting with engineers and the recruitment teams
13:30 in Taub 3 - Lecture on Geometrical Challenges in Treating Arrythmia by Dr. Fadi Matzareva, CS graduate and software engineer in the 3D fiel...

[ Full version ]

[ Full version ]

Sketching Streaming Data

Stream monitoring is fundamental in many data stream applications, such as financial data trackers, security, anomaly detection, and load balancing. To cope with high-speed data streams, these applications require algorithms that are both time and space efficient to cope with high-speed data streams. Space efficiency is needed due to the memory hierarchy structure, to enable cache residency and to avoid page swapping. Even if the potential computing cost is low, this residency is ...

[ Full version ]

[ Full version ]

CEClub: Accelerating Big Data Analytics with the Speedata Analytics Processing Unit (APU

Yoav Etsion (CS & EE, Technion)

Wednesday, 07.12.2022, 11:30

Room 861, EE Meyer Building

At a time when Moore’s Law is reaching its end, the volume of data created, curated, and consumed worldwide grows exponentially. As a result, general-purpose processors are struggling to keep up with the throughput demand of big data analytics workloads.
In this talk I will present Speedata and the Analytics Processing Unit (APU), the first hardware accelerator for big data analytics. Speedata’s APU accelerates existing software frameworks (e.g., Apache Spark, Apache Presto...

[ Full version ]

[ Full version ]

Theory Seminar: Random Walkson Rotating Expanders

Irit Dinur (Weizmann Institute of Science) & Gil Cohen (Tel-Aviv university)

Wednesday, 30.11.2022, 12:30

Taub 201

Random walks on expanders are extremely useful in TOC. Unfortunately though, they have an inherent cost. E.g., the spectral expansion of a Ramanujan graph deteriorates exponentially with the length of the walk (when compared to a Ramanujan graph of the same degree). In this talk, we will see how this exponential cost can be reduced to linear by applying a permutation after each random step. These permutations are tailor-made to the graph at hand, requiring no randomness to generat...

[ Full version ]

[ Full version ]

Coding Theory: Generalized Unique Reconstruction from Substrings

Daniella Bar-Lev (CS, Technion)

Sunday, 27.11.2022, 15:45

Taub 601

New families of reconstruction codes motivated by DNA data storage and sequencing applications will be discussed. In such applications, DNA strands are sequenced by reading some subset of their substrings. The discussion will start with the extreme case of the torn-paper channel in which substrings are read with no overlap. Our model extends the previously researched probabilistic setting to the worst-case. We will construct asymptotically optimal codes, with efficient encoding an...

[ Full version ]

[ Full version ]

"Research on the Bar" Evening - TED Lectures

Wednesday, 23.11.2022, 18:30

CS Taub Build. Auditorium 1

You are invited to the "Research on the Bar" evening - TED lectures and a meeting with three faculty members and their research groups on Wednesday, November 23, 2022 at 18:30 pm in Taub 337:
Dr. Sarah Kern: Does it pay for my robot to be nice?
Dr. Yaniv Romano: Polygraph for learning systems: even machines make mistakes sometimes
Dr. Nir Rosenfeld: Systems learn in a human environment, or: Who needs a phone cradle...

[ Full version ]

[ Full version ]

Open Day for Graduate Studies at CS

Wednesday, 23.11.2022, 12:30

Room 337 Taub Bld.

Hello to outstanding undergraduate students at CS!
CS invites undergraduate outstanding students to continue the tradition of excellence and participate in a meeting on postgraduate studies at the faculty.
The meeting will present the multitude of options and opportunities offered to advanced degrees students, and will be held on Wednesday, November 23, 2022, 12:30-14::00 Taub 337, 3rd floor, Taub Computer Science Building. CS Acting Dean and the Vice Dean for graduate stu...

[ Full version ]

[ Full version ]

Theory Seminar: APMF < APSP? Gomory-Hu Tree in Subcubic Time

Amir Abboud (Weizmann Institute of Science)

Wednesday, 23.11.2022, 12:30

Taub 201

The All-Pairs Max-Flow problem (APMF) asks to compute the maximum flow (or equivalently, the minimum cut) between all pairs of nodes in a graph. The naive solution of making n^2 calls to a (single-pair) max-flow algorithm was beaten in 1961 by a remarkable algorithm of Gomory and Hu that only makes n-1 calls. Within the same time bound, their algorithm also produces a cut-equivalent tree (a.k.a. GH-Tree) that preserves all pairwise minimum cuts exactly. This gives a cubic upper bo...

[ Full version ]

[ Full version ]

Null Messages, Information and Coordination

This work investigates the transfer of information in fault-prone synchronous systems using null messages.
The notion of an {\em $f$-resilient message block} is defined to capture the fundamental communication pattern for knowledge transfer. This pattern may involve null messages in addition to explicit messages, and hence, it provides a fault-tolerant extension of the classic notion of a message-chain.
Based on the above, we provide tight necessary and sufficient characteriza...

[ Full version ]

[ Full version ]

CS Colloquia: Dynamic Matching with Better-than-2 Approx in Polylog Update Time

David Wajc (Google Research → Technion)

Tuesday, 22.11.2022, 14:30

Room 337 Taub Bld.

We present dynamic algorithms with polylogarithmic update time for estimating the size of the maximum matching of a graph undergoing edge insertions and deletions with approximation ratio strictly better than 2. This answers in the affirmative the value version of a major open question, repeatedly asked in the dynamic graph algorithms literature.
Based on an upcoming SODA 2023 best paper, joint with Sayan Bhattacharya, Peter Kiss and Thatchaphol Saranurak.
...

[ Full version ]

[ Full version ]

Pikoya Lecture: Building the Monstera Games Platform

Monday, 21.11.2022, 18:30

Room 337 Taub Bld.

You are invited to the lecture: Building the Monstera Games Platform by David Yanai, CTO at Pikoya and CS magister graduate, about the world and the gaming industry used by billions of users worldwide, and about the implementation aspects, the architecture and the many challenges that Pikoya faced in building the platform.
The lecture will take place on Monday, November 21, 2022, at 18:30 pm in Taub 337.
Please ...

[ Full version ]

[ Full version ]

Learning to Efficiently Compute Accurate Geodesic Distances on Surfaces

A high order accurate deep learning method for computing geodesic distances on surfaces is introduced. We consider two main components for computing distances on surfaces; A numerical solver that locally approximates the distance function and an efficient causal ordering scheme by which surface points are updated. The proposed method exploits a dynamic programming principle which lends itself to a scheme with quasi-linear computational complexity. The quality of the distance appro...

[ Full version ]

[ Full version ]

CGGC Seminar: Isosurfaces: Fast Generation of Large Geometric Models from Scalar Data

Rephael Wenger (The Ohio State University)

Sunday, 20.11.2022, 13:30

Taub 301

Isosurfaces are surface meshes representing ”object” boundaries in scalar field data, such as MRI or CT data. More precisely, isosurfaces are surface meshes representing level sets {f−1(σ) : σ ∈ R}, constructed from a sampling of a function f : R → R3. Isosurface construction, particularly the classical Marching Cubes algorithm (1988), is a standard tool in scientific visualization and geometric modeling, We will discuss Marching Cubes and an alternative, Dual Contouri...

[ Full version ]

[ Full version ]

Thermodynamic Models for the Homeostasis And Development of the Epidermis and Spherical Organoids

Epithelial tissues formed of layered cell surfaces are prevalent in multiple tissues and play an essential role in key biological processes such as development, organ homeostasis and cancer.
In this work we study the mechanical aspects of layered multi-cell systems using pseudo-thermodynamic models.
First, by use of stochastic simulations we were able to characterize the proliferation properties of a new type of stem cell in the mouse interfollicular epidermis.
Furthe...

[ Full version ]

[ Full version ]

Theory Seminar: Dimension-free Relations in Communication Complexity

Lianna Hambardzumyan (The Hebrew University of Jerusalem)

Wednesday, 16.11.2022, 12:30

Taub 201

In this talk we will discuss dimension-free relations between basic communication and query complexity measures and various matrix norms. Dimension-free relations are inequalities that bound a parameter as a function of another parameter without dependency on the number of input bits. This is in contrast to the more common framework in communication complexity where polylogarithmic dependencies are tolerated. Dimension-free bounds are closely related to structural results, where o...

[ Full version ]

[ Full version ]

All You Wanted to Know about a Master's Degree

Monday, 14.11.2022, 17:30

Graduates Club (Taub Build., Floor2)

You are invited to a question and answer session about all that interests you in your master's degree, with a panel of CS graduate students, who will also give you tips and share with you their experience and challenges, with the participation of CS Ph.D. students: Rana Shahout, Hadas Orgad, Eden Seig and Omar Sabary, On Monday, November 14, 2022, at 17:30 at the Graduates Club (at the end of the corridor on the 2nd floor).
The meeting is intended for master's degree students ...

[ Full version ]

[ Full version ]

CGGC Seminar: Shape Matching of Discrete Shells

Florine Hartwig (University of Bonn)

Sunday, 13.11.2022, 13:30

Room 337 Taub Bld.

Finding correspondences between shapes is a central task in geometry processing with many applications such as texture or deformation transfer and shape interpolation. We focus on developing a method to find correspondences between non-isometric geometric shapes. Our method follows the functional map approach. However, unlike existing classical functional map approaches our method is able to match extrinsic surface features by design. To achieve this we consider eigenfunctions of ...

[ Full version ]

[ Full version ]

Theory Seminar: Optimal Weak to Strong Learning

Kasper Green Larsen (Aarhus university)

Wednesday, 09.11.2022, 12:30

Taub 201

The classic algorithm AdaBoost allows to convert a weak learner, that is an algorithm that produces a hypothesis which is slightly better than chance, into a strong learner, achieving arbitrarily high accuracy when given enough training data. We present a new algorithm that constructs a strong learner from a weak learner but uses less training data than AdaBoost and all other weak to strong learners to achieve the same generalization bounds. A sample complexity lower bound shows t...

[ Full version ]

[ Full version ]

Recruitment Day by DELL Technologies

Wednesday, 09.11.2022, 12:30

CS Taub Lobby

You are invited to a Recruitment Day by DELL Technologies to meet engineers and recruit teams who will present programs for graduates and students and their open jobs, on Wednesday, November 9, 20022, 12:30-14:30 at CS Lobby....

[ Full version ]

[ Full version ]

The Real Life behind Entrepreneurship

Tuesday, 08.11.2022, 18:30

Room 337 Taub Bld.

You are invited to a lecture by Iris Shor about the real life behind entrepreneurship at Oribi, who will tell about her path as an entrepreneur and CEO and about life as it is for entrepreneurs, on Tuesday, November 8, 2022, at 18:30 on Taub Terrace.
The lecture is open to all CS students.
Please pre-register
...

[ Full version ]

[ Full version ]

Strategic Classification with Inter-Dependent Strategic Responses

Strategic classification studies learning in settings where agents can modify their features to obtain favourable outcome. Most current works focus on simple decision rules that trigger independent agent responses. Here we examine the implications of learning with more elaborate models that break the independence assumption. We present two works, each studying different (but related) models and tasks, and in which dependencies are introduces through space (using graphs) and time. ...

[ Full version ]

[ Full version ]

Coding Theory: Codes for Constrained Periodicity

Orian Leitersdorf (Technion)

Sunday, 06.11.2022, 14:30

Taub 601

Reliability is an inherent challenge for the emerging nonvolatile technology of racetrack memories, and there exists a fundamental relationship between codes designed for racetrack memories and codes with constrained periodicity. Previous works have sought to construct codes that avoid periodicity in windows, yet have either only provided existence proofs or required high redundancy. This paper provides the first constructions for avoiding periodicity that are both efficient (aver...

[ Full version ]

[ Full version ]

Theory Seminar: The power of the Binary Value Principlein Proof Complexity

Yaroslav Alekseev (St. Petersburg university)

Wednesday, 02.11.2022, 12:30

Taub 201

The (extended) Binary Value Principle (eBVP: $k + x_0 + 2x_1 + … + 2^n x_n $ for $k>0$ and $x^2_i=x_i$) has received a lot of attention recently: several lower bounds have been proved for it (Alekseev et al 2020, Alekseev 2021, Part and Tzameret 2021),
and a polynomial simulation of a strong semialgebraic proof system in IPS+eBVP has been shown (Alekseev et al 2020).
In this talk, we consider Polynomial Calculus with the algebraic version of Tseitin’s extension rule. We sh...

[ Full version ]

[ Full version ]

Learning Subclasses of Junta from Membership Queries

In this work, we address the problem of learning subclasses of $\mathbb{JUNTA}$ under the {\it exact learning model from membership queries} or {\it black box queries}.
First, we address the problem of learning subclasses of decision trees from membership queries. For adaptive non-proper learning of decision trees of depth at most $d$, we give a randomized polynomial time algorithm that asks $\tilde O(2^{2d}) + 2^{d}\log n$ membership queries and a deterministic polynomial tim...

[ Full version ]

[ Full version ]

Coding Theory: DNA storage: Capacity and Error Probability Bounds

Nir Weinberger (Technion)

Sunday, 30.10.2022, 14:30

Taub 601

We will discuss results on the capacity and error probability bounds of the DNA storage channel. First, we consider the case in which the sequencing channel is memoryless and the coverage depth is constant. We will describe lower (achievability) and upper (converse) bounds on the capacity of the channel, as well as a lower (achievability) bound on the reliability function of the channel. Second, we will consider general sequencing channels, and coverage depth scaling, and focus on...

[ Full version ]

[ Full version ]

Using Transformers to Model Electronic Health Records of ICU and for Prediction of Blood Stream Infection

Machine learning made many recent advances in science and technology, specifically in healthcare information technology. Electronic Health Records (EHR) data store the healthcare information. EHR data consists of many features. It is highly complicated, noisy and includes many outliers and missing values. It also contains time-dependent information such as vital sign measurements, diagnosis, treatment, etc.
Therefore, basic machine learning models have poor performance on this ...

[ Full version ]

[ Full version ]

Similarity-based Regularization for Mitigating Artifacts

Common methods for mitigating spurious correlations in natural language understanding (NLU) usually operate in the output space, encouraging a main model to behave differently from a bias model by down-weighing examples where the bias model is confident.
While improving out of distribution (OOD) performance, it was recently observed that the internal representations of the presumably debiased models are actually more, rather than less biased.
We propose SimgReg, a new method for ...

[ Full version ]

[ Full version ]

Theory Seminar: Learning Dimensions

Amir Yehudayoff (Technion)

Wednesday, 26.10.2022, 12:30

Taub 201

We shall survey and discuss several dimensions in the theory of machine learning....

[ Full version ]

[ Full version ]

CGGC Seminar: Deep Learning for Surface Geometries and Its Applications

Prof. Tatsuya Yatagawa (School of Engineering, The University of Tokyo) - CANCELLED!

Monday, 24.10.2022, 13:30

Taub 401

Deep learning has been one of the most common techniques in many industrial and research fields, and that for processing 3D geometries has also been investigated intensely in the last several years.
However, compared to the techniques for 3D volumetric images and point sets, those for surface geometries, e.g., represented by triangular meshes, have not been well established due to the difficulty in handling spatial and topological features simultaneously using a neural network....

[ Full version ]

[ Full version ]

ceClub: Challenges with Analog In-Memory Computing Arrays for Efficient Deep Learning Acceleration

Jongeun Lee (UNIST, Ulsan, Korea)

Thursday, 20.10.2022, 11:30

Zoom Lecture: 94673013539 and Meyer 861

ReRAM (Resistive Random-Access Memory) crossbar arrays have the potential to provide extremely efficient matrix-vector multiplication (MVM) operations, which are the cornerstone of many DNN (Deep Neural Network) applications. However, there are several challenges in order for ReRAM crossbar arrays (RCAs) to be useful for accelerating large-scale DNN applications. In this talk we discuss two of those challenges. The first one is the distortion in the MVM output of RCAs due to nonid...

[ Full version ]

[ Full version ]

Neural Sequence Models: A Formal Lens

Neural sequence models (NSMs) - neural networks adapted specifically for the task of processing input sequences - have emerged as powerful tools in sequence processing, with the current most popular architectures being transformers and RNN variants. But what is a trained network really doing? In this talk we will approach this question, starting from the question of what a network *can* do, and progressing to the question of what a trained network *has* learned in practice.
Sp...

[ Full version ]

[ Full version ]

ceClub: Starlight: Fast Container Provisioning on the Edge and over the WAN

Moshe (Mickey) Gabel (York University)

Wednesday, 19.10.2022, 11:30

Zoom Lecture: 94673013539 and Meyer 861

Containers, originally designed for cloud environments, are increasingly popular for provisioning workers outside the cloud, for example in mobile and edge computing. These settings, however, bring new challenges: high latency links, limited bandwidth, and resource-constrained workers. The result is longer provisioning times when deploying new workers or updating existing ones, much of it due to network traffic.
Our analysis shows that current piecemeal approaches to reducing p...

[ Full version ]

[ Full version ]

Terraforming-Environment Manipulation During Disruption Events for Lifelong Multi-Agent Path Finding

Lifelong Multi-Agent Pathfinding (L-MAPF) is concerned with planning collision-free paths for a team of agents as they continuously handle tasks involving pick-up and delivery. When modeling autonomous warehouses, typical approaches for L-MAPF consider the environment as populated with static obstacles in the form of inventory pods that the agents must avoid. These obstacles impose narrow passageways, forcing agents to resort to detours and suffer delays on account of bottlenecks....

[ Full version ]

[ Full version ]

Multi A(ge)nt Systems on Graphs

Multi-agent systems are a fascinating, multidisciplinary field with applications to robotics, distributed systems, biology, and social dynamics. A multi-agent system is a distributed system composed of several interacting, autonomous agents that cooperate to achieve some desired behavior. The topic of this talk is the way in which local interactions between extremely simple agents can result in desirable global states. We study this topic from two perspectives: the perspective of ...

[ Full version ]

[ Full version ]

Geometry-based Dynamic Connectivity Analysis of Biological Neural Networks

Learning in organisms is one of their most fundamental but intricate processes. Understanding learning is a longstanding problem in neuroscience as well as in artificial neural networks. In this work, we focus on studying in biological networks during motor learning through the lens of the connectivity of neurons in the primary motor cortex (M1). For this purpose, we analyze the neural activity recorded from awake and behaving mice using two-photon calcium imaging. These imaging m...

[ Full version ]

[ Full version ]

Generalizing Reinforcement Learning Agents with Abstract Contextual Embeddings

Classic reinforcement learning methods such as Q-learning and policy gradient methods have seen great success in learning to perform a plethora of common tasks in AI, from playing video games to controlling self-driving vehicles and beyond. As a result, these methods and their extensions have become standard in most reinforcement learning settings. However, they have trouble adapting to changes in the environment. We believe this issue can be solved by giving the agent awareness o...

[ Full version ]

[ Full version ]

Succinct Interactive Oracle Proofs

\textit{Interactive Oracle Proofs} (IOPs) are a new type of proof-system that combines key properties of interactive proofs and PCPs: IOPs enable a verifier to be convinced of the correctness of a statement by interacting with an untrusted prover while reading just a few bits of the messages sent by the prover. IOPs have become very prominent in the design of efficient proof-systems in recent years.
In this work we study \textit{succinct IOPs}, which are IOPs in which the comm...

[ Full version ]

[ Full version ]

On The Structure Of Heilbronn’s Configurations

Heilbronn's triangle problem asks how to place n points in the unit square, such that the smallest of the $\binom{n]{ 3}$ triangles is maximized. This problem, which was opened in the 1950s by the mathematician Heilbronn, has not yet been answered for n>8. Calling H)n( the area of the smallest triangle of the best set of n points, no tight bounds on it have been found so far. The best upper and lower bounds are $O(n^{-\mu+\epsilon})$, where $\mu=\frac{8}{7}$ , and $\Omega(\frac{\...

[ Full version ]

[ Full version ]

Interpreting Embedding Spaces by Conceptualization

One of the main methods for semantic interpretation of text is mapping it into a vector in some embedding space. Such vectors can then be used for a variety of textual processing tasks. Recently, most embedding spaces are a product of training large language models. One major drawback of this type of representation is their incomprehensibility to humans. Understanding the embedding space is crucial for several important needs, including the need to explain the decision of a sys...

[ Full version ]

[ Full version ]

A Framework for Clinical Classification of Multivariate Time Series using Koopman Operators

Clinical multivariate time series (MTS) arising from sensor data, such as EEG and ECG, is used in a variety of tasks.
The sensors are composed of multiple leads connected to the body, where each lead generates a time series of data.
Combining information from the different leads allows inference of cardiac activity from ECG (arrhythmia, acute coronary syndrome) or brain dysfunction from EEG (brain tumors, strokes, epilepsy).
We present a framework for clinical classific...

[ Full version ]

[ Full version ]

Graph Neural Networks Pretraining Through Inherent Supervision for Molecular Property Prediction

Recent global events have emphasized the importance of accelerating the drug discovery process. This process may take more than a decade and its overall cost might exceed one billion dollars. A way to deal with these issues is to use machine learning to increase the rate at which drugs are made available to the public while reducing the cost of the entire process. However, chemical labeled data for real-world applications is extremely scarce making traditional approaches less effe...

[ Full version ]

[ Full version ]

Fair Correlation Clustering In General Graphs

We consider the family of Correlation Clustering optimization problems under fairness constraints.
In Correlation Clustering we are given a graph whose every edge is labeled either with a $+$ or a $-$, and the goal is to find a clustering that agrees the most with the labels: $+$ edges within clusters and $-$ edges across clusters.
The notion of fairness implies that there is no over, or under, representation of vertices in the clustering: every vertex has a color and the distr...

[ Full version ]

[ Full version ]

On the Orchestration of Advanced Cellular Networks

In recent years network operators are experiencing changes in clients needs. Self-driving cars, augmented reality games and large scale data streaming are simple examples of new applications that require faster service with higher bandwidth availability to the clients. These changes force the network operators shift their business model and new network paradigms arise.
The ongoing transition into 5G networks (and 6G networks that will soon arrive) is enabled in part by the comb...

[ Full version ]

[ Full version ]

Extremal Properties of Polyominoes

Lattice animals are edge-connected sets of cells over various lattices. Some famous examples are polyominoes, polyhexes, polyiamonds and polycubes which are in the square, hexagonal triangular and cubical lattices respectively. Lattice animals have been studied extensively both as a combinatorial object, and as modeling tool in statistical physics.
The two most significant problems related to lattice animals are the counting problem and the growth rate problem. The first is sim...

[ Full version ]

[ Full version ]

Theory Seminar: Hardness of Approximation in P via Short Cycle Removal: Cycle Detection, Distance Oracles, and Beyond

Seri Khoury (UC Berkeley)

Wednesday, 10.08.2022, 12:30

Amado 814

Triangle finding is at the base of many conditional lower bounds in P, mainly for distance computation problems, and the existence of many $4$- or $5$-cycles in a worst-case instance had been the obstacle towards resolving major open questions.
We present a new technique for efficiently removing almost all short cycles in a graph without unintentionally removing its triangles. Consequently, triangle finding problems do not become easy even in almost $k$-cycle free graphs, for a...

[ Full version ]

[ Full version ]

FLOC 2022: The Eighth Federated Logic Conference

Sunday, 31.07.2022, 09:00

Technion, Haifa

FLOC 2022: The Eighth Federated Logic Conference (FLoC 2022,July 31-August 12, 2022, Haifa, Israel)
Hosted by the Henry and Marilyn Taub Faculty of Computer Science at the Technion
ABOUT FLOC
During the past forty years there has been extensive, continuous, and growing interaction between logic and computer science. In many respects, logic provides computer science with both a unifying foundational framework and a tool for modeling. In fact, logic has been called “the cal...

[ Full version ]

[ Full version ]

A Machine Learning Exploration of Relations between Protein Structures and their Genetic Coding

Synonymous codons translate into the same amino acid. Although the identity of synonymous codons is often considered inconsequential to the final protein structure there is mounting evidence for an association between the two. Protein structure plays an important role in understanding the biological function and mechanism of a protein therefore understanding the relations between protein structures and their genetic coding is crucial. Our study examined the association between the...

[ Full version ]

[ Full version ]

CGGC Seminar: 3D Content Creation Made Fast and Easy

Hsueh-Ti Derek Liu (University of Toronto)

Sunday, 24.07.2022, 11:00

Room 337 Taub Bld.

Creating digital 3D objects has been a central task across different disciplines and the key towards democratizing the metaverse. However, 3D content creation is still a privilege reserved for professional modelers because existing content creation tools are difficult to use by the general public. My research aims to lower the difficulty of 3D content creation to the point where everyone can manipulate digital 3D shapes. In this talk, I will first discuss how to build easy-to-use ...

[ Full version ]

[ Full version ]

Topology-Controlled Reconstruction from Partial Planar Cross-Sections

The problem of three-dimensional reconstruction from planar cross-sections arises in many fields, such as biomedical image analysis, and geographical information systems. The problem has been studied extensively in the past~40 years.
Each cross-section of the input contains multiple contours, where each contour divides the plane into different material types.
The reconstructed three-dimensional object is a valid volume (surrounded by a closed surface) that interpolates the ...

[ Full version ]

[ Full version ]

On Synthesis and Reconstruction of human Facial Photometry and Corresponding Geometry

In this thesis, we study the modeling of human faces. As all structured data is believed to reside on some low-dimensional manifold in a high-dimensional space, we wish to study and model the so-called manifold of human faces. By uncovering the latent manifold of faces one can project onto the manifold (facial reconstruction) as well as sample from the manifold (facial synthesis), two tasks with a wide range of applications such as gaming, animation, and AR/VR to name a few.
In...

[ Full version ]

[ Full version ]

Concurrent Games with Multiple Topologies

Concurrent multi-player games with omega-regular objectives are a standard model for systems that consist of several interacting components, each with its own objective. The standard solution concept for such games is Nash Equilibrium (NE), which is a ``stable'' strategy profile for the players. In many settings, the system is not fully observable by the interacting components, e.g., due to internal variables. Then, the interaction is modelled by a partial information game. Unfort...

[ Full version ]

[ Full version ]

Reliable Concurrent Computing

The rapid deployment of multi-core architectures has resulted in a dire need for scalable and reliable concurrent algorithms. This dissertation focuses on the design of concurrent data structures, which constitute building blocks for concurrent algorithms. Two major design goals in this domain are reliability and efficiency. This talk will concentrate on the hardness of reclaiming concurrent data-structures' memory efficiently, while taking care to preserve reliability. It will in...

[ Full version ]

[ Full version ]

Theory Seminar: APMF < APSP? Gomory-Hu Tree in Subcubic Time

Amir Abboud (Weizmann Institute of Science) - CANCELLED!

Wednesday, 29.06.2022, 12:30

Taub 201

The All-Pairs Max-Flow problem (APMF) asks to compute the maximum flow (or equivalently, the minimum cut) between all pairs of nodes in a graph. The naive solution of making n^2 calls to a (single-pair) max-flow algorithm was beaten in 1961 by a remarkable algorithm of Gomory and Hu that only makes n-1 calls. Within the same time bound, their algorithm also produces a cut-equivalent tree (a.k.a. GH-Tree) that preserves all pairwise minimum cuts exactly. This gives a cubic upper bo...

[ Full version ]

[ Full version ]

Pixel Club: Graph Neural Networks through the Lens of Measure Theory

Despite their growing popularity, graph neural networks (GNNs) still suffer from multiple unsolved problems, including lack of embedding expressiveness, propagation of information to distant nodes, and training on large-scale graphs. Understanding the roots of and providing solutions for such problems require developing analytic tools and techniques. In this talk we provide a measure theoretic point of view for the above-mentioned problems, and derive a notion of “recoverabilit...

[ Full version ]

[ Full version ]

Deep Reinforcement Learning for 5G Dense Urban Wireless Routing

We study the problem of routing real-time flows over a multi-hop mmWave mesh. We develop a model-free Deep Reinforcement Learning algorithm that determines which subset of the mmWave links should be activated during each time slot and using what power level. The proposed algorithm, called AARL (Adaptive Activator RL), can handle a variety of network topologies, packet loads and interference models. It does not require prior knowledge of the interdependence of different mmWave link...

[ Full version ]

[ Full version ]

A New Lower Bound on the Growth Constant of Polycubes in Three Dimensions

In this thesis, we deal the approximation problem of the growth rate of polycubes in three dimensions. We consider three-dimensional polycubes, which are finite collections of face-connected 3D-cubes, centered in points of $\mathbb{Z}^3$, where the lexicographically smallest cube is centered in $(0,0,0)$. If we denote the number of 3D polycubes comprised of $n$ cubes by $A(n)$, then we know from prior results that this sequence behaves like an exponential, and so we denote its gro...

[ Full version ]

[ Full version ]

Protocol Infernece from Program Executable Using Symbolic Execution and Automata Learning

Protocol Inference is the process of gaining information about a protocol from a binary code that implements it. This process is useful in cases such as extraction of the command and control protocol of a malware, uncovering security vulnerabilities in a network protocol implementation or verifying conformance to the protocol's standard. Protocol inference usually involves time-consuming work to manually reverse engineer the binary code.
We present a novel method to automatical...

[ Full version ]

[ Full version ]

Projects Fair on IoT, Android, Arduino and Networks

Sunday, 26.06.2022, 14:00

CS Taub Lobby

You are invited to the CS Taub projects fair for the Spring Semester of 2022, where 30 teams of undergraduate students will present and demonstrate projects in various fields in IoT, Android, Arduino and Networks, developed as part of the final project in the software engineering and communication networks track, most of which were carried out in collaboration with various social associations and organizations, and were intended to make a contribution to the community.
The fai...

[ Full version ]

[ Full version ]

Special Seminar: Efficient Detection of High Probability Cryptographic Properties of Large Boolean Functions via Surrogate Differentiation

Prof. Adi Shamir (Weizmann Institute of Science)

Thursday, 23.06.2022, 12:30

CS Taub Build. Auditorium 1

A central problem in cryptanalysis is to find all the significant deviations from randomness in a given $n$-bit cryptographic primitive. When $n$ is large, the only practical way to find such statistical properties was to exploit the internal structure of the primitive and to speed up the search with a variety of heuristic rules of thumb. However, such bottom-up techniques can miss many properties, especially in cryptosystems which are designed to have hidden trapdoors.
In this...

[ Full version ]

[ Full version ]

Theory Seminar: Communication Complexity-based Lower Bounds for Proof Complexity of Natural Formulas

Artur Ryazanov (St. Petersburg University)

Wednesday, 22.06.2022, 12:30

Taub 201

A canonical communication problem Search(ϕ) is defined for every unsatisfiable CNF ϕ: an assignment to the variables of ϕ is distributed among the communicating parties, they are to find a clause of ϕ falsified by this assignment. Lower bounds on the randomized k-party communication complexity of Search(ϕ) in the number-on-forehead (NOF) model imply tree-size lower bounds, rank lower bounds, and size-space tradeoffs for the formula ϕ in the semantic proof system Tcc(k,c) tha...

[ Full version ]

[ Full version ]

Pixel Club: CLIP as a Generative Foundation Model

Amit Bermano (Tel-Aviv University)

Tuesday, 21.06.2022, 11:30

Room 1061, EE Meyer Building

Large scale Mega-models are impossible to train using standard hardware, but encompass a vast semantic understanding of our world. In this talk, I explore three ways to leverage the knowledge encompassed in the Recent Large scale"Contrastive Language-Image Pre-training" (CLIP) model, as a foundation to push the boundaries of generative capabilities:
- InStyleGAN-NADA, we show how to adapt the StyleGAN generator across a multitudeof domains characterized by diverse styles and s...

[ Full version ]

[ Full version ]

CGGC Seminar: Application-Driven Geometric Machine Learning

Prof. Justin Solomon (MIT)

Monday, 20.06.2022, 12:30

Taub 401

From 3D modeling to autonomous driving, a variety of applications can benefit from data-driven reasoning about geometric problems. The available data and preferred shape representation, however, varies widely from one application to the next. Indeed, the one commonality among most of these settings is that they are not easily approached using data-driven methods that have become de rigueur in other branches of computer vision and machine learning. In this talk, I will summarize...

[ Full version ]

[ Full version ]

SHE S Ladies Night Eevent 2022

Sunday, 19.06.2022, 19:30

CS Taub Terrace

You are invited to the the lady students community annual event at Technion CS, on Sunday, June 19, 2022, starting at 19:00 on the Taub terrace.
In the program:
19:00 - Mingling, sushi and beers
20:00 - Stand-up show by Mor Chen
21:00 - Afterparty Karaoke
Please pre-register (registration for the event involves a nominal fee)....

[ Full version ]

[ Full version ]

Theory Seminar: Binary Codes with Resilience Beyond 1/4 via Interaction

Klim Efremenko (Ben-Gurion University)

Wednesday, 15.06.2022, 12:30

Taub 201

In the reliable transmission problem, a sender, Alice, wishes to transmit a bit-string x to a remote receiver, Bob, over a binary channel with adversarial noise. The solution to this problem is to encode x using an error-correcting code. As it is long known that the distance of binary codes is at most 1/2, reliable transmission is possible only if the channel corrupts (flips) at most a 1/4-fraction of the communicated bits.
We revisit the reliable transmission problem in the tw...

[ Full version ]

[ Full version ]

Machine Learning and Real-time AI: Reveal the Mystery

Tuesday, 14.06.2022, 18:30

Zoom Event: Registration

You are invited to a technology enrichment lecture, by Nava Levy, AI / ML Dev Advocate, Redis, on machine learning in the high-tech industry, with an emphasis on real-time machine learning, and the difference between it and deep learning and the challenges in real-time AI, on roles and responsibilities of a typical data science and machine learning team in an industrial company, on the differences between research and practice, and on tools that can be started today.
The lectur...

[ Full version ]

[ Full version ]

TRX'22: Technion Robotics Expo

Sunday, 12.06.2022, 16:30

CS Taub Build. Auditorium 1

You are invited to the TRX'22: Technion Robotics Expo, on Sunday, June 12, 2022, 16:30-20:00 in CS Taub Auditorium 1.
Please pre-register.
...

[ Full version ]

[ Full version ]

Theory Seminar: Nonlinear Repair Schemes of Reed-Solomon Codes

Itzhak Tamo (Tel-Aviv University)

Wednesday, 08.06.2022, 12:30

Taub 201

The problem of repairing linear codes, particularly Reed Solomon (RS) codes, has attracted a lot of attention in recent years due to its importance in distributed storage systems. In this problem, a failed code symbol (node) needs to be repaired by downloading as little information as possible from a subset of the remaining nodes. There are examples of RS codes with efficient repair schemes, and some are even optimal. However, these schemes fall short in several aspects; for examp...

[ Full version ]

[ Full version ]

Recruitment Day by Vast Data

Wednesday, 08.06.2022, 12:30

CS Taub Lobby and Visitors Center Auditorium

Vast Data engineers and recruitment teams will arrive at CS to demonstrate technologies and offer open positions, on Wednesday, June 8, 2022, 12:30, in the Taub Lobby, and at 13:30 for a technological lecture on the challenges of building data structures and distributed algorithms used to establish the largest storage systems in the world, in th...

[ Full version ]

[ Full version ]

ceClub: Conflict and Technology, AI Chip Wars from the Inside Out

Oskar Mencer (Maxeler Technologies)

Wednesday, 08.06.2022, 11:30

Room 861, EE Meyer Building and zoom Lecture: 91011338796

The microprocessor is 50 years old. 50 years ago, a single ALU at kHz speeds had to be shared by multiple tasks, multiple applications, multiple users and multiple organizations. Due to transistor scaling we can now have 1M ALUs on a chip at GHz speeds. In this talk I will swap the cause and effect equation, instead of talking about solutions to problems, I will talk about problems created by solutions. Today data movement dominates compute time. I will describe how by optimally s...

[ Full version ]

[ Full version ]

CGGC Seminar: Nozzle Modification for Efficient FDM 3D Printing

Shir Rorberg (CS, Technion)

Wednesday, 08.06.2022, 10:00

Taub 401

3D printing is based on layered manufacturing, where the layers are printed consecutively in increasing height order. In Fused Depositing Modeling (FDM), the printing head may travel without extruding material between separated “islands” of the sliced layers. These travel movements increase the printing time and reduce the quality of the 3D printed part. We present an extended nozzle modification, which can be applied to off-the-shelf FDM printers, and a corresponding toolpath...

[ Full version ]

[ Full version ]

Online Meeting: How to be "On it" Financially

Tuesday, 07.06.2022, 18:30

Zoom Event: Registration

You are invited to an online meeting led by Yael Marom, Savings Product Manager at RiseUp, on: How to be "on it" financially, which will explain in depth what our monthly financial situation is and the way to know how much we really need to spend, how to prepare ahead, for short term and long term, in order to gradually build our economic growth, on Tuesday, June 7, at 18:30.
For link to the meeting ...

[ Full version ]

[ Full version ]

Robustness and Rotation Equivariance in Geometric Deep Learning

Graph neural networks (GNNs) have shown broad applicability in a variety of domains.
These domains, e.g., social networks and recommendation systems, are fertile ground for malicious users and behavior. In a series of works, we study the robustness of GNNs under different scenarios and present a simple rotation and permutation equivariant point-cloud GNN.
We show that GNNs are vulnerable to the scenario of strategic behavior of multiple users (i.e., Strategic Classification...

[ Full version ]

[ Full version ]

Theory Seminar: From Selection Theorems to Weak Epsilon-Nets in Higher Dimensions (and back?)

Natan Rubin (Ben-Gurion University)

Wednesday, 01.06.2022, 12:30

Taub 201

Given a finite point set $P$ in $R^d$, and $\eps>0$ we say that a point set $N$ in $R^d$ is a weak $\eps$-net if it pierces every convex set $K$ with $|K\cap P|\geq \eps |P|$.
Let $d\geq 3$. We show that for any finite point set in $R^d$, and any $\eps>0$, there exists a weak $\eps$-net of cardinality $o(1/\eps^{d-1/2})$. Here $delta>0$ is an arbitrary small constant.
This is the first improvement of the bound of $O^*(1/\eps^d)$ that was obtained in 1993 by Chazelle, Edelsbrunne...

[ Full version ]

[ Full version ]

Recruitment Day by Vayyar

Wednesday, 01.06.2022, 10:30

Vayyar representatives will visit CS to demonstrate 3D sensor technologies capable of seeing through objects, and for a lecture on 13:30 by Uri Adar, Vayyar system group member, about 3D imaging in RF, on Wednesday, June 1, 2022, between 10:30-14:30, at the CS Taub Lobby. More details in the attached poster.
Please ...

[ Full version ]

[ Full version ]

Pixel Club: Moving Forward with StyleGAN to Real Data and New Domains

Ron Mokady (Tel-Aviv University)

Tuesday, 31.05.2022, 11:30

Zoom Lecture: https://technion.zoom.us/my/chaimbaskin

StyleGAN is already quite famous for its unremarkable image editing capabilities. Although other generative models (e.g. diffusion models) achieve comparable synthesis quality, they cannot reproduce these semantically richmanipulations. In particular, StyleGAN allows the modification of various attributes, such as hair, age, pose, expression, and make-up, while still maintaining a high level of realism.
Yet, it is still challenging to leverage these traits for real data or new ...

[ Full version ]

[ Full version ]

Microsoft Lecture: Improving Productivity Through NLP

Tuesday, 24.05.2022, 17:00

Room 337 Taub Bld.

You are invited to a technological lecture by Dikla Dotan Cohen, Director of Research at Microsoft, on Improving Productivity Through NLP, which will present the methodology of Office 365 products and the challenges it poses, on Tuesday, May 24, 17:00, in Room 337, CS Taub Building.
Please pre-register....

[ Full version ]

[ Full version ]

CS-Hackathon 2022 - Doing Good

Thursday, 19.05.2022, 09:30

CS Taub Building

You are invited to join the CS Hackathon-Doing Good programming competition to be held on Thursday-Friday, May 19-20, 2022, at CS Taub Building, and which this year will deal with social work and community contribution, and will focus on developing engineering-technological solutions to increase accessibility for people with disabilities, to help them improve their quality of life.- an opportunity to implement creative ideas that promote values of involvement and con...

[ Full version ]

[ Full version ]

Moran Samuel: "To Cross the Line"

Thursday, 19.05.2022, 09:00

CS Taub Build. Auditorium 1

You are invited to a lecture by Moran Samuel, Paralympic Medalist who will light a beacon on the upcoming Independence Day, and a visitor of the CS-Hackathon 2022, which this year deals with developing solutions to increase accessibility for people with disabilities, and who will tell about her life as an athlete and a winner, on Thursday, May 19, 2022, at 09:00, CS Taub Building Auditorium 1.
...

[ Full version ]

[ Full version ]

Recruitment Day by CYE

Wednesday, 18.05.2022, 12:30

CS Taub Lobby

You are invited to a Recruitment Day by CYE, that provides security solutions for organizations, on Wednesday, May 18, 2022, starting at 12:30 in of the Taub Building lobby, and between 13: 30-14: 30 for a technology lecture on "Cyber, Programming and all in between" - attacks and strategies using code - by Eyal Greenberg, one of the company's founders and leader of the research team.
Please ...

[ Full version ]

[ Full version ]

Theory Seminar: Temporal Path Finding in the Presence of Delays

Hendrik Molter (Ben-Gurion University)

Wednesday, 18.05.2022, 12:30

Taub 201

Consider planning a trip in a train network. In contrast to, say, a road network, the edges are temporal, i.e., they are only available at certain times. Another important difficulty is that trains, unfortunately, sometimes get delayed. This is especially bad if it causes one to miss subsequent trains. The best way to prepare against this is to have a connection that is robust to some number of (small) delays. An important factor in determining the robustness of a connection is ho...

[ Full version ]

[ Full version ]

ceClub: Non-explicit Information Exchange in Distributed Computing

Guy Goren (EE, Technion)

Wednesday, 18.05.2022, 11:30

Room 861, EE Meyer Building and zoom Lecture: 92697362743

Distributed systems become more and more ubiquitous. These systems entail different instances of distributed computing problems, each with unique characteristics. In this talk, I will discuss my Ph.D. research on theoretical and algorithmic aspects of distributed computing. Specifically, I will focus on two works that represent different aspects of my research on non-explicit information exchange in distributed systems. I will start with a fundamental work on concepts of informati...

[ Full version ]

[ Full version ]

Sound Source Modeling of Drones In Free Space and Indoors towards Acoustic-Based Indoor Localization

To model the self-sound of drones, acoustics analysis and high-fidelity computational fluid dynamic methods can be used. However, these methods require significant computational resources. Therefore, data-driven and analytical methods are commonly used to model the sound source, enabling the generation of a pressure-time history of the moving rotors along a time varying shaft position. We suggest a simple and low computational data-driven method for modeling the sound source of a ...

[ Full version ]

[ Full version ]

Theory Seminar: On the Size of Succinct Non-interactive Arguments in the Random Oracle Model

Eylon Yogev (Bar-Ilan University)

Wednesday, 11.05.2022, 12:30

Taub 201

Are all SNARG constructions in the random oracle model inherently large? The answer to this question is still open, but I will survey recent work that makes significant progress towards a solution.
In particular, we will see a new construction that achieves a smaller argument size. This is the first progress on the Micali construction since it was introduced over 25 years ago.
Our construction relies on a strong soundness notion for PCPs and a weak binding notion for commitm...

[ Full version ]

[ Full version ]

Preparation Workshop for Technical Interviews

Tuesday, 10.05.2022, 17:30

Zoom Event: Registration

You are invited to a preparation workshop for technical Interviews, led by Or Ben-Hayal, CS graduate and software engineer at Google, which will deal with the components of the job interview and review common mistakes made by candidates, and which will include a preparatory lecture for the various aspects of the interview, as well as practice of technical questions from job interviews and their solutions, on Tuesday, May 10, 2022, at 17:30. in Taub 337.
Please ...

[ Full version ]

[ Full version ]

Pixel Club: Fundamental Visual Motion Cues for Autonomous Navigation

Daniel Raviv (Florida Atlantic University (FAU))

Tuesday, 10.05.2022, 11:00

Room 337 Taub Bld.

This talk is about low-level fundamental visual motion cues that can help autonomous vehicles navigate in unknown structured and unstructured environments. Following bio-inspired and behavior-based observations and motivations, the talk focuses on relevant concepts and recent results as obtained from simulated and real data.
Some of the visual cues, e.g., the“visual looming” cue, are environment, scale, and rotation independent, and are measured in time units. Obtaining ...

[ Full version ]

[ Full version ]

Scalable Blockchain Anomaly Detection with Sketches

The growing popularity of Blockchain networks attracts also malicious and hacking users. Effectively detecting inappropriate and malicious activity should thus be a top priority for safeguarding blockchain networks and services. Blockchain behavior analysis can be used to detect unusual account activities or time periods with network-wide irregular properties. Thus, optimized anomaly detection based on historical data is an essential task for securing transactions and services. Ho...

[ Full version ]

[ Full version ]

Recruitment Day by Pikoya

Monday, 09.05.2022, 10:00

CS Taub Lobby

You are invited to Recruitment Day by Pikoya (startup in publication and online games promoting), to meet their engineers and recruitment staff, to be held on Monday, May 9, 2022, starting at 10:00, at the Taub Lobby.
You are all invited!
...

[ Full version ]

[ Full version ]

Practical Workshop in Linkedin

Monday, 02.05.2022, 18:30

Zoom Event: Registration

You are invited for a practical LinkedIn workshop, led by Anat Kegel Taub, director of the Sourcing Team at Meta, who will provide tips to help you brand yourself to future recruiters, executives and co-workers, on Monday, May 2, 2022, at 18:30 in Taub 337.
Please pre-register....

[ Full version ]

[ Full version ]

Towards Predicting Fine Finger Motions from Ultrasound Images via Kinematic Representation

A central challenge in building robotic prostheses is the creation of a sensor-based system able to read physiological signals from the lower limb and instruct a robotic hand to perform various tasks. Existing systems typically perform discrete gestures such as pointing or grasping, by employing electromyography (EMG) or ultrasound (US) technologies to analyze the state of the muscles.
In this research, we study the inference problem of identifying the activation of specific f...

[ Full version ]

[ Full version ]

ceClub: Twilight: A Differentially Private Payment

Yossi Gilad (The Hebrew University of Jerusalem)

Monday, 25.04.2022, 11:30

Room 861, EE Meyer Building and zoom Lecture: 93108695810

Payment channel networks (PCNs) provide a faster and cheaper alternative to transactions recorded on the blockchain. Clients can trustlessly establish payment channels with relays by locking coins and then send signed payments that shift coin balances over the network’s channels. Although payments are never published, anyone can track a client’s payment by monitoring changes in coin balances over the network’s channels. We present Twilight, the first PCN that provides a rigo...

[ Full version ]

[ Full version ]

Distributional Robustness: From Pricing to Auctions

Robust mechanism design is a rising alternative to Bayesian mechanism design, which yields designs that do not rely on assumptions like full distributional knowledge. We apply this approach to mechanisms for selling a single item, assuming that only the mean and range of the distribution of buyer values is known. We seek the mechanism that maximizes revenue over the worst-case distribution compatible with the known parameters. Such a mechanism arises as an equilibrium of a zero-su...

[ Full version ]

[ Full version ]

Theory Seminar: Efficient Multiparty Interactive Coding for Insertions, Deletions and Substitutions

Ran Gelles (Bar-Ilan University)

Wednesday, 13.04.2022, 12:30

Taub 601 Taub Bld.

Interactive coding allows two or more parties to carry out a distributed computation over a communication network that may be noisy. The ultimate goal is to develop efficient coding schemes that tolerate a high level of noise while increasing the communication by only a constant factor (i.e., constant rate). In this work we provide computationally efficient, constant rate schemes that conduct any computation on arbitrary networks, and succeed with high probability in the presence ...

[ Full version ]

[ Full version ]

Only Real Friends Matter - A clusters Based Deep Learning Paradigm For High Dimension Multivariate Forecasting

Multivariate time series forecasting differs from univariate time series forecasting by trying to model the dependencies between the different time series in order to make a more precise forecasting. Despite reaching better results in a multivariate setting, classical and deep learning multivariate models are not scalable, having their total number of parameters growing square of the number of time-series. We present in this paper a novel paradigm in how deep learning models shall...

[ Full version ]

[ Full version ]

Recruitment Day by Philips

Tuesday, 12.04.2022, 11:00

CS Taub Lobby

You are invited to recruitment day by Philips and to meet their engineers and recruitment staff, on Tuesday, April 12, 2022, starting at 11:00, at the CS Taub Lobby.
You are all invited!...

[ Full version ]

[ Full version ]

Intel Preparatory Workshop for Job Interview

Monday, 11.04.2022, 18:00

Zoom Event: Registration

You are invited to Intel preparation workshop for the technical part in job interview, on Monday, April 11, 2022, 18:00, in a zoom session - a link will be sent after pre-registration.
You are all invited!...

[ Full version ]

[ Full version ]

qualitative SLAM

Simultaneous localization and mapping (SLAM) is essential in numerous robotics applications such as autonomous navigation. Traditional SLAM approaches infer the metric state of the robot (position and orientation) along with a metric map of the environment. While existing algorithms exhibit good results, they are still sensitive to measurement noise, sensors quality, data association and are still computationally expensive. Computational load is especially problematic in active pl...

[ Full version ]

[ Full version ]

Research Meetup with a Focus on NLP

Wednesday, 06.04.2022, 18:30

Taub Terrace

You are invited to the first Research Meetup with a focus on NLP, on Wednesday, April 6th, 18:30 in the Taub Terrace.
The Meeting's goal is to bring together Israel’s ML and NLP communities in a setting far more casual than the traditional well known seminar/conference/poster session.
The event format will be as follows:
- The night will open with a quick speech from Professor Yonatan Belinkov
- A meeting of undergraduate and graduate students, researchers and faculty wi...

[ Full version ]

[ Full version ]

Theory Seminar: Public-Key Quantum Money with a Classical Bank

Omri Shmueli (Tel-Aviv University)

Wednesday, 06.04.2022, 12:30

Taub 201

Quantum money is a main primitive in quantum cryptography, that enables a bank to distribute to parties in the network, called wallets, unclonable quantum banknotes that serve as a medium of exchange between wallets. While quantum money suggests a theoretical solution to some of the fundamental problems in currency systems, it still requires a strong model to be implemented; quantum computation and a quantum communication infrastructure. A central open question in this context is ...

[ Full version ]

[ Full version ]

Geometric Sorting of Simple Agents on Grid Environments with applications to autonomous traffic management

We study a geometrically constrained combinatorial problem inspired by the following scenario: autonomous vehicles move on a $m$-lane freeway, where $m \geq 2$. Each vehicle heads to some destination and is allowed to exit the road only through a designated exit lane when approaching its destination. We assume that vehicles have limited memory and sensing capabilities, and cannot directly communicate with their peers.
We present a completely decentralized distributed algorith...

[ Full version ]

[ Full version ]

Temporal Graphs: Embedding with Fairness

In this thesis, we study temporal graphs and how to best represent their nodes and edges for multiple classification tasks. We first study the basics of how to represent nodes and edges in (un)weighted and (un)directed temporal graphs. We then present methods to leverage different aspects of temporal graphs, such as a temporal message passing and multiple attributes over edges. Finally, we study how bias manifests itself in temporal graphs and propose methods to balance accuracy a...

[ Full version ]

[ Full version ]

Learning Multigrid

During the last decade, Neural Networks (NNs) have proved to be extremely effective tools in many fields of engineering, including autonomous vehicles, medical diagnosis and search engines, and even in art creation. Indeed, NNs often decisively outperform traditional algorithms. One area that is only recently attracting significant interest is using NNs for designing numerical solvers, particularly for discretized partial differential equations. Several recent papers have consider...

[ Full version ]

[ Full version ]

Improved Bounds for Online Aggregation and Caching

Ohad Talmon

Wednesday, 30.03.2022, 14:00

Taub 601 Taub Bld.

Uncertainty is a key factor in real-time systems, where decisions must be made over time based on incomplete or partial information. Competitive analysis is the prominent paradigm for the design and analysis of algorithms for such environments, which are called online algorithms. The field of competitive analysis has been studied extensively throughout the last few decades, and is highly useful in the analysis of such real-time systems.
In an online problem data is revealed ove...

[ Full version ]

[ Full version ]

Best Project Contest - The Finals

Wednesday, 30.03.2022, 12:30

CS Taub Lobby

You are invited to the finals event of the Best Project Contest, that will take place in the format of a project fair, on Wednesday, March 30, 2022, starting at 12:30, and to the announcing and awarding the winners at 14:00, at the CS Taub Lobby.
You are all invited to come and meet the best researchers and researches!...

[ Full version ]

[ Full version ]

Theory Seminar: Locality-Preserving Hashing for Shifts with Connections to Cryptography

Ohad Klein (Bar-Ilan University)

Wednesday, 30.03.2022, 12:30

Taub 201 Taub Bld.

Alice receives a non-periodic string (such as ABCDEF), while Bob receives a string (such as CDEFAB), obtained by applying a hidden cyclic shift to Alice’s string.
Alice and Bob query their strings in a small number of positions (sublinear in the amount of shifting) and then exchange a single short message.
How can they detect the shift with minimal error probability?
Based on Joint works with Elette Boyle, Itai Dinur, Niv Gilboa, Yuval Ishai, Nathan Keller.
...

[ Full version ]

[ Full version ]

ceClub: Three Papers Session

Alon Rashelbach, Lior Zeno, Haggai Eran

Wednesday, 30.03.2022, 11:30

Room 861, EE Meyer Building and zoom Lecture: 96624383219

In this CE club session, we will present three papers that will appear next week at NSDI'22:
Scaling Open vSwitch with a Computational Cache
by Alon Rashelbach
Open vSwitch (OVS) is a widely used open-source virtual switch implementation. In this work, we seek to scale up OVS to support hundreds of thousands of OpenFlow rules by accelerating the core component of its data-path – the packet classification mechanism. To do so we use NuevoMatch, a recent algorithm tha...

[ Full version ]

[ Full version ]

Theory Seminar: On the Complexity of Two-Party Differential Privacy

Eliad Tsfadia (Tel-Aviv University)

Wednesday, 23.03.2022, 12:30

Taub 201 Taub Bld.

In distributed differential privacy, the parties perform analysis over their joint data while preserving the privacy for both datasets. Interestingly, for a few fundamental two-party functions such as inner product and Hamming distance, the accuracy of the distributed solution lags way behind what is achievable in the client-server setting. McGregor, Mironov, Pitassi, Reingold, Talwar, and Vadhan [FOCS ’10] proved that this gap is inherent, showing upper bounds on the accuracy o...

[ Full version ]

[ Full version ]

CS Open Day for Graduate Studies

Wednesday, 23.03.2022, 12:30

Room 337 Taub Bld.

Technion CS open day 2022 invites outstanding undergraduates from all universities to learn about the Computer Science Department and register for Winter Semester 2022-23.
The event will be held on Wednesday, March23, 2022. between 12:30-14:00, room 337, Taub Building for Computer Science, Technion.
The program will include review on curriculum, r...

[ Full version ]

[ Full version ]

How Gender Debiasing of NLP Models Affects Internal Model Representations, and Why It Matters

Common studies of gender bias in natural language processing (NLP) focus either on extrinsic bias which is measured by model performance on a specific task or on intrinsic bias which is measured on a models' internal representations. However, the relationship between extrinsic and intrinsic bias is relatively unknown. In this work, we illuminate this relationship by measuring both quantities together: we debias a model during downstream fine-tuning, which reduces extrinsic bias, a...

[ Full version ]

[ Full version ]

Recruitment Day by NVIDIA

Monday, 21.03.2022, 12:00

CS Taub Lobby

CS students are invited to Recruitment Day by NVIDIA to be held on Monday, March 21, 2022, 12:00, in the CS Taub Lobby.
You are all invited....

[ Full version ]

[ Full version ]

Constructions and Bounds of Codes Correcting Combinations of Deletion Errors

Correcting insertions/deletions as well as substitution errors simultaneously plays an important role in DNA-based storage systems as well as in classical communications. However, in DNA data storage as well as in file/symbol synchronization, not only insertions/deletions occur, but also classical substitution errors. Additionally, some cases feature array-like words and as such pose a new type of variance in insertions/deletions errors - column insertions/deletions as opposed to ...

[ Full version ]

[ Full version ]

Semantic Perception under Uncertainty with Viewpoint-Dependent Models

Semantic perception is the process of acquiring and maintaining knowledge of the environment of a robot (or more generally - embodied agent) beyond geometric structure, i.e. capturing meaning - such as classes and other high-level properties of visible scene elements - as opposed to pure geometry. Semantic perception is key to enabling autonomous robots to operate in diverse, low-structured and dynamic environments and alongside humans. In the past decade semantic information has ...

[ Full version ]

[ Full version ]

CS Orientation Day 2022

Tuesday, 15.03.2022, 11:00

CS Taub Build. Auditorium 1

CS 2022 Orientation Day for new students will be held on Tuesday, March 15, 2022, and will begin at 10:00 with a Technion meeting at the Churchill Auditorium where the Senior Vice President, the Dean for Undergraduate Studies and the Students Dean and Chairman of the Technion Student Association will speak to the new students, and between 11:00-14:00 there will be a gathering at the Computer Science Taub Building, and a meeting at the Taub 1 Auditorium in the entrance floor, which...

[ Full version ]

[ Full version ]

PartTLB: Dynamic TLB Partitioning for SMT Systems

Simultaneous multithreading (SMT) increases the cost of memory address translation due to sharing of the translation lookaside buffer (TLB) among multiple threads. Current x86 processors use a ``competitively-shared’’ TLB, in which low-locality threads might needlessly waste TLB resources and thus degrade the performance of neighboring high-locality threads. To address this problem, we introduce PartTLB, a new mechanism that: (1) samples the TLB requests of the competing threa...

[ Full version ]

[ Full version ]

Model-Based Simulation for SMT Cores

Studies that evaluate new architectural designs of virtual memory typically employ a ``model-based’’ methodology that relies on simulations of the translation lookaside buffer (TLB) coupled with empirical performance models. We observe that this methodology is limited in that each simulated thread of execution has its own dedicated TLB, whereas modern processors share a single TLB among multiple threads through ``simultaneous multithreading’’ (SMT). Existing model-based re...

[ Full version ]

[ Full version ]

Analyzing Individual Neurons in Language Models

Neural language models have significantly developed in recent years, becoming more and more successful on numerous language tasks. Those models rely on encoding words as hidden vector representations, before utilizing these representations for the task at hand. Their success spiked interest in their interpretability: understanding how they work, and what is encoded within these representations. While many studies have shown that linguistic information is encoded in hidden word rep...

[ Full version ]

[ Full version ]

ceClub: Embedded Systems and Security

Farhad Merchant (Aachen University)

Wednesday, 02.03.2022, 11:30

Room 861, EE Meyer Building and zoom Lecture: 96393404383

Protecting intellectual properties from untrusted design houses and foundries has become highly challenging. In this talk, I will focus on the security aspects of embedded systems. First, I will discuss the logic locking tools and attack methods developed at the Institute for Communication Technologies and Embedded Systems, RWTH Aachen University. In the second part of the talk, I will focus on developing low-power, high-performance embedded systems based on emerging non-volatile ...

[ Full version ]

[ Full version ]

Improving Graph Neural Networks Expressivity Via Spectral and Combinatorial Pre-Colorings

Graph isomorphism testing is usually approached via the comparison of graph invariants. Two popular alternatives that offer a good trade-off between expressive power and computational efficiency are combinatorial (i.e., obtained via the Weisfeiler-Leman (WL) test) and spectral invariants. While the exact power of the latter is still an open question, the former is regularly criticized for its limited power, when a standard configuration of uniform pre-coloring is used. This drawba...

[ Full version ]

[ Full version ]

Semantic Symmetry in Transducers

In model checking, we work toward deciding whether a system satisfies a given specification. Often, a system exhibits some type of symmetry in its structure or in its behaviour. Such symmetries can be exploited by a designer to alleviate some of the complexity of model checking, as well as to gain insight into the behaviour of the system. Thus, we want to decide whether a given system exhibits symmetry.
Symmetry is not a well-defined concept and might come in various forms, eac...

[ Full version ]

[ Full version ]

A Meeting with Potential Students for Technion CS Studies

Thursday, 24.02.2022, 10:00

Zoom Event: Registration

A meeting with potential students who are interested in studies at the Technion and the Faculty of Computer Science will be held online on CS Facebook, onThursday, February 24, 2022, at 10:100 and at 14:00.
Details and registration. ...

[ Full version ]

[ Full version ]

TCE-MLIS 2021 Conference

Thursday, 24.02.2022, 08:30

ELMA Arts Complex, Zichron Ya'acov

MLIS, the Technion AI center, in collaboration with TCE, would like to invite you to participate in the annual MILS-TCE conference.
AI is now a major buzz word everywhere and expectations are sky-rocketing, but what is true state-of-the-art and what can be actually implemented in the AI and Machine Learning fields?
In a series of lectures, Techni...

[ Full version ]

[ Full version ]

REDEEMER: Reinforcement Learning Based CEP Pattern Miner for Knowledge Extraction

Complex Event Processing (CEP) are a set of methods that allow efficient knowledge extraction from massive data streams using complex and highly descriptive patterns. As of today, in many fields, patterns are manually defined by human experts. However, desired patterns often contain convoluted relations that are difficult for humans to detect, and human expertise is scarce in many domains.
We present REDEEMER, a novel reinforcement and active learning approach aimed at mining C...

[ Full version ]

[ Full version ]

DELETE: Using deep learning to minimize latency in CEP systems

The ability to detect complex patterns in massive data streams is critical for many real-time applications. These applications must uphold low latency requirements, delivering alerts and notifications with minimal response delays. Complex event processing (CEP), a leading technology for performing this task, is suitable for the efficient and robust detection of complex patterns. However, the CEP complexity grows exponentially with respect to the length of the pattern and the inten...

[ Full version ]

[ Full version ]

CloudWalker: 3D Point Cloud Learning by Random Walks for Shape Analysis

Adi Mesika

Tuesday, 08.02.2022, 11:30

Zoom Lecture: https://technion.zoom.us/my/chaimbaskin

Point clouds are gaining prominence as a method for representing 3D shapes, but their irregular structure poses a challenge for deep learning methods.
In this paper we propose CloudWalker, a novel method for learning 3D shapes using random walks. Previous works attempt to adapt Convolutional Neural Networks (CNNS) or impose a grid or mesh structure to 3D point clouds. This work presents a different approach for representing and learning the shape from a given point set. The ke...

[ Full version ]

[ Full version ]

Pixel Club: Convex Optimization: Adaptive Learned Solvers and CoordinateGradient Descent

Aviad Aberdam (CS/EE, Technion

Tuesday, 08.02.2022, 10:30

Zoom Lecture: https://technion.zoom.us/my/chaimbaskin

Neural networks that are based on unfolding of an iterativesolver, such as LISTA (learned iterative soft threshold algorithm), are widelyused due to their accelerated performance. Nevertheless, as opposed tonon-learned solvers, these networks are trained on a certain dictionary, andtherefore they are inapplicable for varying model scenarios. This talkintroduces an adaptive learned solver, termed Ada-LISTA, which receives pairsof signals and their corresponding dictionaries as inpu...

[ Full version ]

[ Full version ]

ILP Based Load Balancing in Deduplicated Storage Systems

Deduplication reduces the size of the data stored in large-scale storage systems by replacing duplicate data blocks with references to their unique copies. This creates dependencies between files that contain similar content and complicates the management of data in the system. In the work presented in this seminar, we have addressed the problem of data migration, where files are remapped between different volumes because of system expansion or maintenance. The challenge of det...

[ Full version ]

[ Full version ]

Clustering Based Data Migration in Deduplicated Storage

Deduplication is a leading method for reducing physical storage capacity when duplicate data is present. This method can be applied on chunks, files, containers, and more. Instead of storing the same physical data multiple times, a pointer is created from each logical copy to the same physical copy, saving the space of the duplicate data. Due to this, data is shared between objects, such as files or entire directories, which result in garbage collection overhead and migration chal...

[ Full version ]

[ Full version ]

CGGC Seminar: Trading Memory for Computations: Scaling Range Matching on the Critical Path

Range matching (RM) is a crucial component in computer systems, widely used in address translation, hard drives, network switches, and many more applications. RM is performed whenever one wishes to locate a range that contains an input number, given a large set of ranges. Any page-based mechanism uses RM, as pages are basically ranges. Longest prefix matching (LPM) uses ternary rules, which are also ranges. Firewalls are one example of multidimensional RM since ACL rules consist o...

[ Full version ]

[ Full version ]

ceClub: SmartNIC Inline Processing

The inline processing technique enables data transformation as a system transfers data to or from a processing node. It is used to offload computations and accelerate data-intensive communication tasks, reducing latency and power due to data movement and improving throughput by using the best processing core for the job. However, inline processing poses several challenges: it breaks existing operating system and network stack layers and makes it difficult to reuse previous softwar...

[ Full version ]

[ Full version ]

pISTA: preconditioned Iterative Soft Thresholding Algorithm for Graphical Lasso

We propose a novel quasi-Newton method for solving the sparse inverse covariance estimation problem also known as the graphical least absolute shrinkage and selection operator (GLASSO).
This problem is often solved using a second order quadratic approximation. However, in such algorithms the Hessian term is complex and computationally expensive to handle. To this end,our algorithm uses the inverse of the Hessian as a preconditioner to simplify and approximate the quadratic e...

[ Full version ]

[ Full version ]

Pixel Club: TextAdaIN: Paying Attention to Shortcut Learning in TextRecognizers

Oren Nuriel (AWS)

Tuesday, 25.01.2022, 11:30

Zoom Lecture: https://technion.zoom.us/my/chaimbaskin

Leveragingthe characteristics of convolutional layers, neural networks are extremelyeffective for pattern recognition tasks. However in some cases,their decisions are based on unintended information leading to high performanceon standard benchmarks but also to a lack of generalization to challengingtesting conditions and unintuitive failures. Recentworkhas termed this “shortcut learning” and addressed its presence in multipledomains. In text recognition, we reveal another such...

[ Full version ]

[ Full version ]

Solving Constrained Horn Clauses Lazily and Incrementally

Constrained Horn Clauses (CHCs) is a fragment of First Order Logic (FOL), that has gained a lot of attention in recent years. One of the main reasons for the rising interest in CHCs is the ability to reduce many verification problems to satisfiability of CHCs. For example, program verification can naturally be described as the satisfiability of CHCs modulo a background theory such as linear arithmetic and arrays. To this end, CHC-solvers can be used as the back-end for different v...

[ Full version ]

[ Full version ]

Efficient Self-Supervised Data Collection for Offline Robot Learning

a large batch of real or simulated robot interaction data, using some data collection policy, and then learn from this data to perform various tasks, using offline learning algorithms. Previous work focused on manually designing the data collection policy, and on tasks where suitable policies can easily be designed, such as random picking policies for collecting data about object grasping. For more complex tasks, however, it may be difficult to find a data collection policy that e...

[ Full version ]

[ Full version ]

CS LECTURE: Mathematical Foundations of Robust Geometry and Fabrication

Current geometry methods for creating and manipulating shapes on computers can sometimes be unreliable and fail unpredictably. Such failures make geometry tools hard to use, prevent non-experts from creating geometry on their computers, and limit the use of geometry methods in domains where reliability is critical. We will discuss my recent efforts in proving when existing methods work as intended, my work in making methods more robust to imperfect input, my work in the creatio...

[ Full version ]

[ Full version ]

SMEGA2: Distributed Deep Learning Using a Single Momentum Buffer

As the field of deep learning progresses, and models become larger and larger, training deep neural networks has become a demanding task. The task requires a huge amount of compute power, and can still be very time consuming - especially when using just a single GPU. To tackle this problem, distributed deep learning has come into play, with various asynchronous training algorithms. However, most of these algorithms suffer from decreased accuracy as the number of workers increases....

[ Full version ]

[ Full version ]

Mathematical Techniques for Cryptanalysis

Symmetric ciphers are cryptographic algorithms that use the same cryptographic keys for both encryption and decryption. The key represents a shared secret between users, that is used to maintain a private information link. In our research we focus on cryptanalysis of block ciphers, which are the most widely used realization of symmetric ciphers.
Block ciphers are cryptosystems that consist of two algorithms, an encryption algorithm that accepts a symmetric key and a plaintext ...

[ Full version ]

[ Full version ]

How to Avoid Depth Reconstruction in 3D Vision Tasks: Do We Need Depth in State-Of-The-Art Face Authentication?

Face recognition systems are frequently used in a variety of security applications in our daily lives. Some methods are designed to utilize geometric information extracted from depth sensors to overcome single-image-based recognition technologies’ weaknesses, such as vulnerability to illumination variations, large head poses, and spoofing attacks. However, the accurate acquisition of the depth profile or surface is an expensive and challenging process. We introduce a novel metho...

[ Full version ]

[ Full version ]

Pixel Club: Endless Loops: Detecting and Animating Periodic Patterns in Still Images

Tavi Halperin (The Hebrew University of Jerusalem)

Tuesday, 18.01.2022, 11:30

Zoom Lecture: https://technion.zoom.us/my/chaimbaskin

We present an algorithm for producing a seamless animated loop from a single image. The algorithm detects periodic structures, such as the windows of a building or the steps of a staircase, and generates a non-trivial displacement vector field that maps each segment of the structure onto a neighboring segment along a user- or auto-selected main direction of motion. This displacement field is used, together with suitable temporal and spatial smoothing, to warp the image and produce...

[ Full version ]

[ Full version ]

Bribery attack on Nakamoto Consensus Proof of Stake Protocols

Bitcoin was introduced to the world in 2009 with Proof of Work (PoW) Leader Election as one of its novel building blocks. Since then, much criticism has been made of its high energy consumption. Proof of Stake protocols aims at replacing PoW protocols as a much more efficient version while still maintaining its security properties under the Honest Majority model.
In our work, we show a bribery attack under the Rational Majority model, which breaks the persistence security prope...

[ Full version ]

[ Full version ]

Securing ICS Protocols

Industrial Control Systems (ICSs), also known as Operation Technology (OT) systems, are distributed computerized systems designed to manage, monitor and control industrial processes.
They are widely used in critical infrastructures, such as power plants and water supply, whose continuous operation is of major importance to modern life.
Following the well-known Stuxnet attack on OT systems, a large investment in OT security was started.
Though their cyber protection is crucial, ...

[ Full version ]

[ Full version ]

CS LECTURE: Computing Using Time

George Tzimpragos (UC Santa Barbara and Lawrence Berkeley National Laboratory)

Wednesday, 12.01.2022, 17:30

Zoom Lecture: 96743325005

The development of computing systems able to address our ever-increasing needs, especially as we reach the end of CMOS transistor scaling, requires truly novel methods of computing. My research draws inspiration from biology, rethinks the digital/analog boundary, and challenges conventional wisdom, which typically guides how we perform computation, by reimagining the role of time. In this talk, I first introduce a computational temporal logic that sets the foundation for tempo...

[ Full version ]

[ Full version ]

An Automata Theory Method for the Analysis of Unicycle Pursuit Problems

The Pursuit Problem depicts a scenario where a moving target is pursued by an agent, whose movement is prescribed by some defined policy.
Examples of what can be regarded as solutions to the pursuit problem include the shape of the agent's trajectory, whether the agent ultimately captures the target, and the circumstances of the capture, including the time required for capture to be achieved,
The Unicycle Model is a popular simplification used to describe the kinematics of ...

[ Full version ]

[ Full version ]

Distributed Deep Neural Networks

Training deep neural networks in the distributed asynchronous setting is complicated. In the distributed asynchronous setting, the computational devices run in parallel, causing a delay in the propagation of information between the different computational devices. The delay is often referred to as staleness, which harms the training process and the quality of the deep neural network. This staleness is one of the main difficulties in scaling asynchronous settings to a large number ...

[ Full version ]

[ Full version ]

Pixel Club: Layered Neural Atlases for Consistent Video Editing

Dolev Ofri (Weizmann Institute of Science)

Tuesday, 11.01.2022, 11:30

Zoom Lecture: https://technion.zoom.us/my/chaimbaskin

While image editing and manipulation tools have seen steady progress, allowing complex editing effects to be achieved by novice users, video editing remains a difficult task: applying edits in a temporally consistent manner to all frames remains a key challenge.
In this talk, I’ll present a novel method that tackles this challenge by decomposing an input video into a set of layered 2D atlases, each providing a unified representation of an object/background over the entire vid...

[ Full version ]

[ Full version ]

Incorporating Time into Word Representations

Our world is constantly evolving, and so is the content on the web. Consequently, our languages, often said to mirror the world, are dynamic in nature.
However, most current language representations are static and cannot adapt to changes over time.
New words and semantic evolution have been shown to pose a crucial challenge in many Natural Language Processing and Information Retrieval tasks, leading to a significant performance drop for modern language models.
In this the...

[ Full version ]

[ Full version ]

Incentive-Aligned Strategic Classification

Predictive machine learning tools are increasingly being used to inform decisions regarding humans.
When human users stand to gain from certain predictive outcomes, they may be prone to act strategically to improve those outcomes.
We argue that in many realistic scenarios the system and its users are in fact aligned in their goals.
In this work, we give concrete real-world examples for such environments and demonstrate using a series of experiments that they are incent...

[ Full version ]

[ Full version ]

CS LECTURE: Sublinear-time Graph Algorithms: Motif Counting and Uniform Sampling

Talya Eden (MIT and Boston University)

Sunday, 09.01.2022, 10:30

Room 012 Taub Bld (Learning Center Auditorium)

In this talk I will survey recent developments in approximate subgraph-counting and subgraph-sampling in sublinear-time. Counting and sampling small subgraphs (aka motifs) are two of the most basic primitives in graph analysis, and have been studied extensively, both in theory and in practice. In my talk, I will present the sublinear-time computational model, where access to the input graph is given only through queries, and will explain some of the concepts that underlie my resul...

[ Full version ]

[ Full version ]

Practical WEB Development Workshop: To Code of Not to Code

Wednesday, 05.01.2022, 18:30

Zoom Event: Registration

You are invited to a practical web development workshop that will include an overview of the latest technologies of web application development and practical practice with a demonstration, with the help of Uri Shaked, Voiding warranties at Wokwi.com.
The workshop will take place on Wednesday, January 5, 18:30, in a zoom session, and it is optional but not required to be familiar with at least one programming language such as JavaScript, React, Next.js, as well as installation o...

[ Full version ]

[ Full version ]

Estimating NLP Domain Adaptation Performance Using Model Causal Analysis

Domain adaptation setups were not all born equal, and some domains are easier to adapt to and from than others.
This talk will show and attempt to estimate the difficulty (or ease) of adapting between different domains, based on the causal effect of certain features in the data on the adapting model’s predictions. This question is relevant in many real-life scenarios where computational resources exist in relative abundance, while labeling and data-gathering is time-consumin...

[ Full version ]

[ Full version ]

Pixel Club: Computational Imagingfor Sensing High-speed Phenomena

Despite recent advances in sensor technology, capturing high-speed video at high-spatial resolutionsremains a challenge. This is because, in a conventional camera, the available bandwidth limits either the maximum sampling frequency or thecaptured spatial resolution. In this talk, I am going to cover our recent works that use computational imaging to allow high-speed high-resolution imagingunder certain conditions. First I will describe Diffraction Line Imaging, a novel imaging pr...

[ Full version ]

[ Full version ]

CS LECTURE: Informed Data Science

Amir Gilad (Duke University)

Tuesday, 04.01.2022, 10:30

Room 012 Taub Bld (Learning Center Auditorium)

Data science has become prevalent in various fields that affect day-to-day lives, such as healthcare, banking, and the job market. The process of developing data science applications usually consists of several automatic systems that manipulate and prepare the data in different manners. Examples of automatic data manipulations and preparations include generating synthetic data, interactive data exploration, repairing the data, and labeling it for machine learning. These systems ca...

[ Full version ]

[ Full version ]

CS LECTURE: On the Role of Data in Algorithm Design

Tal Wagner (Microsoft Research Redmond)

Sunday, 02.01.2022, 12:30

Room 012 Taub Bld (Learning Center Auditorium)

Recently, there has been a growing interest in harnessing the power of big datasets and modern machine learning for designing new scalable algorithms. This invites us to rethink the role of data in algorithm design: not just as the input to pre-designed algorithms, but also a factor that enters the algorithm design process itself, driving it in a strong and possibly automated manner. This talk will show how to leverage data and learning for better algorithm design in two fundament...

[ Full version ]

[ Full version ]