בית הספר
להנדסת חשמל ומחשבים

Presentor: Shy-El Cohen Supervisor: Dr. Eliya Nachmani Degree: MSc in computers and electrical engineering Seminar Subject: Hybrid Mamba–Transformer Decoder for Error-Correcting Codes Abstract: We introduce a novel deep learning method for decoding error correction codes based on the Mamba architecture, enhanced with Transformer layers. Our approach proposes a hybrid decoder that leverages Mamba’s efficient sequential modeling while maintaining the global context capabilities of Transformers. To further improve performance, we design a novel layer-wise masking strategy applied to each Mamba layer, allowing selective attention to relevant code features at different depths. Additionally, we introduce a progressive layer-wise loss, supervising the network at intermediate stages and promoting robust feature extraction throughout the decoding process. Comprehensive experiments across a range of linear codes demonstrate that our method outperforms or matches Transformer-only decoders while improving complexity. Name: Tal Shuster Supervisor: Dr. Eliya Nachmani Degree: MSc in computers and electrical engineering Subject of the seminar: Q2D2: A geometry-aware audio codec leveraging two-dimensional quantization Abstract: Recent neural audio codecs have achieved impressive reconstruction quality, typically relying on quantization methods such as Residual Vector Quantization (RVQ), Vector Quantization (VQ) and Finite Scalar Quantization (FSQ). However, these quantization techniques limit the geometric structure of the latent space, make it harder to capture correlations between features leading to inefficiency in representation learning, codebook utilization and token rate. In this paper we introduce Two-Dimensional Quantization (Q2D2), a quantization scheme in which feature pairs are projected onto structured 2D grids, such as hexagonal, rhombic, or rectangular tiling and quantized to the nearest grid values, yielding an implicit codebook defined by the product of grid levels, with codebook sizes comparable to conventional methods. Despite its simple geometric formulation, Q2D2 improves audio compression efficiency, with low token rates and high codebook utilization while maintaining state of the art reconstruction quality. Specifically, Q2D2 achieves competitive to superior performance in various objective and subjective reconstruction metrics, across extensive experiments in speech domain compared to state of the art models. Comprehensive ablation studies further confirm the effectiveness of our design choices. Name: Alon Helvits Degree: Electrical and Computer Engineering Supervisor: Eliya Nachmani Seminar Subject: Score Based Error Correcting Code Decoder Error-correcting codes are fundamental to reliable communication, but practical soft decoding remains difficult to scale across diverse code families and block lengths. In this seminar, I will present a score-based decoding approach that reframes decoding as a continuous-time denoising problem in symbol space (BPSK over AWGN). Instead of relying on magnitude-only reliability features or requiring explicit SNR estimation, the method learns a denoising field directly from the signed channel observation and uses it and parity constraint (syndrome) to define a probability-flow ODE. Decoding proceeds by numerically integrating this ODE from the noisy received vector toward a parity-consistent codeword, with the solver budget providing a natural accuracy–latenc

Shira Karmi Electrical and Computer Engineering Prof. Tami Riklin Reviv seminar topic: Decoding Functional Networks for Visual Categories via GNN Abstract: Understanding how large-scale brain networks represent visual categories is fundamental to linking perception and cortical organization. Using high-resolution 7T fMRI from the Natural Scenes Dataset, we construct parcel-level functional graphs and train a signed Graph Neural Network that models both positive and negative interactions, with a sparse edge mask and class-specific saliency. The model accurately decodes category-specific functional connectivity states sports, food,vehicles) and reveals reproducible, biologically meaningful subnetworks along the ventral and dorsal visual pathways. This framework bridges machine learning and neuroscience by extending voxel-level category selectivity to a connectivity-based representation of visual processing. Abu Hussein, Tariq1, Abdulhalim Ibrahim1 Optically addressed spatial light modulators (OASLMs) are fundamental components in the development of parallel optical computers, serving critical roles in various applications, including optical correlators, high-speed modulation, compatibility with coherent light sources, image color conversion, image polarization conversion, image amplification, 3D displays, digital holography, and projection displays[1]. The structure of OASLM enables high resolution without the need for complicated fabrication processes because it is a single-pixel device. Our group has recently developed an IR-to-visible image conversion device using an InGaAs photosensor and a nematic LC layer[2]. Full name - Anastasia Ivanovski Supervisors - Prof. Ofer Hadar, school of Electrical and Computer Engineering, Ben Gurion University; Dr. Alaa Jamal, Institute of Agricultural and Biosystems Engineering, Volcani Institute Degree - M.Sc Electrical engineering student Seminar topic - Hybrid Physics-Informed Machine Learning for Real-Time Water Quality Prediction in Aquaculture Ponds Seminar abstract -  Effective aquaculture management relies on accurate forecasting of water temperature and dissolved oxygen, yet traditional models often force a trade-off between physical interpretability and data-driven flexibility. While mechanistic models struggle with environmental variability, purely data-driven approaches lack physical consistency and robustness to sensor noise. In this talk, I will present a hybrid framework that addresses these limitations using Physics-Informed Neural Networks (PINNs) and Symbolic Regression (PySR/SINDy). By embedding thermodynamic laws and diffusion equations directly into the model, we ensure predictive accuracy even in poorly instrumented settings. This approach results in a lightweight, interpretable decision-support tool that provides near real-time early-warning indicators for proactive pond management.

Name: Yuval Meir Studies: Electrical and Computer Engineering with Thesis Under the supervision of : Prof. Shmuel Ben-Yaakov Seminar Title: Second harmonic generation in nonlinear capacitors and application in electric field sensor design Abstract: This seminar presents a study on the design of a novel electric field sensor based on the Second Harmonic Generation (SHG) phenomenon in Non-Linear Capacitors (NLC). The research focuses on developing a comprehensive model to describe the generation of the second harmonic signal under large-signal excitation, explicitly characterizing the correlation between the capacitor's non-linearity profile and the resulting harmonic magnitude. Bridging theoretical analysis with high-fidelity SPICE simulations, the study validates these findings through laboratory experiments and culminates in the design and implementation of a practical proof-of-concept electric field tester, demonstrating the feasibility of this sensing approach. Guy Perets MSc. In Communication Systems Engineering Mentors: Prof. Itshak Lapidot (Afeka), Dr. Yehuda Ben-Shimol (BGU) Subject: Waveform-Based Spoof-Robust Automatic Speaker Verification Using Deep Learning Seminar Summary: Modern speaker verification systems are increasingly targeted by spoofing attacks such as replay and synthetic speech. A waveform-based anti-spoofing system distinguishes genuine speech from manipulated audio by learning authenticity cues directly from the raw signal. It models the characteristic patterns of natural speech production and real recording chains, and flags samples that deviate in ways consistent with spoofing while aiming to stay reliable across diverse conditions and attack styles. Tuvia Hausdorff M.sc in Electrical and Computers Engineering Supervised by Dr. Dan Vilenchik Title: Robust Voice-Based Depression Detection in Real-World Settings using Voice Conversion GenAI Summary: My research explores how speech-based machine learning systems can robustly detect depression in real-world clinical settings, where data is limited, noisy, and demographically biased. I focus on using self-supervised speech representations and generative audio models to reduce dependence on labeled data and improve generalization across speakers and recording conditions. In particular, I study label-free and weakly supervised adaptation techniques, including voice conversion–based data augmentation and test-time adaptation, while keeping the pipeline aligned with clinical constraints such as sensitivity, uncertainty, and failure modes. The goal is to move voice-based mental health models from controlled benchmarks toward reliable, deployable systems that engineers can reason about, evaluate, and integrate into real products.

Full name: Itamar Elmakias Degree program: PhD Supervisor: Dr. Dan Vilenchik (cc to this email) Seminar title: Choosing the Right Feature Selection Algorithm: Dataset Hardness, Algorithm Cost, and Practical Guidelines Seminar abstract: Feature Selection (FS) is a central component in modern machine learning pipelines, particularly for high-dimensional classification tasks. While a large body of research proposes new FS algorithms, practitioners still lack clear guidance on when FS is truly beneficial, which algorithms to use, and how to balance performance gains against computational cost and stability. In the first part of this seminar, I will present findings from our recent work on dataset hardness characterization . Using a large-scale empirical study across many real-world datasets, we show that the effectiveness of FS strongly depends on intrinsic dataset properties, and that the common assumption that FS is universally beneficial does not hold. We introduce a practical taxonomy of datasets based on their response to FS, providing an empirical lens for understanding when FS is likely to help. In the second part, I will present ongoing work that shifts the focus from datasets to algorithms. We analyze FS algorithms through multiple operational dimensions, including runtime, stability across cross-validation folds, average and maximum performance gains, and sensitivity to the number of selected features. Based on these analyses, we propose a cost-aware, bucketed framework for FS algorithm selection, offering actionable guidelines for choosing an FS method under different time budgets and dataset regimes. Overall, the seminar aims to bridge the gap between theoretical FS research and practical decision-making, providing evidence-based heuristics for selecting FS algorithms rather than treating them as black-box components.

Title: Resilient Decision-Making for Multi-Robot Systems in the Presence of Adversaries Speaker: Roee M. Francos, Robotics, Embedded Autonomy, and Communication Theory (REACT) Lab, School of Engineering and Applied Sciences, Harvard University This talk presents a unified perspective on resilient decision-making for multi-robot systems in environments where adversaries may influence routing, search, and detection tasks. Recently, significant progress in coordination, control, and navigation of multi-robot systems has been achieved, driven primarily by the rapid commercialization of unmanned aerial systems and drones. Yet, real-world deployment remains challenging. Most research assumes cooperative or optimally performing agents, overlooking adversarial or suboptimal behavior due to uncertainty, faults, or environmental disturbances. Resilience to malicious or malfunctioning agents remains a key limitation, as such agents can degrade efficiency and destabilize routing policies, where stability is defined as bounded cost over time. Existing stability guarantees for cooperative fleets collapse when agents deviate from plans, underscoring the need for adversarially aware planning and routing theory in safety-critical applications. I will present new theoretical and experimental results on resilient multi-agent policies for coordination, control, and learning under uncertainty and adversarial influence, focusing on routing and traffic management for fleets of aerial and ground agents for which I develop adaptive algorithms that provide provable stability, resilience, and safety guarantees. Finally, I will show how resilient cooperative search strategies address challenges posed by intelligent, coordinated external adversaries, providing provable guarantees for coverage and detection in settings such as search and rescue and pursuit-evasion. I conclude by outlining future research directions at the intersection of safety, learning, and large-scale autonomy. Bio: Roee M. Francos is currently a Computer Science Postdoctoral Fellow at the Robotics, Embedded Autonomy, and Communication Theory (REACT) Lab at Harvard University working with Prof. Stephanie Gil, focusing on development of multi-agent resilient decision-making and coordination algorithms. In 2023, he completed his PhD in Computer Science under supervision of Prof. Freddy Bruckstein, at the Technion-Israel Institute of Technology. He received the B.Sc. in Electrical and Computer Engineering from Ben-Gurion University. His research interests are in multi-agent teamwork, autonomous robotics, intelligent transportation systems, bio-inspired robotics and computer vision, focusing on collaborative algorithms for motion planning of autonomous vehicles and multi-robot learning. Roee is a recipient of the 2023 Robotics Science and Systems (RSS) Pioneers Award and the 2025 IEEE Multi-Robot & Multi-Agent Systems (MRS) Young Pioneer Award.