Bar-Ilan Faculty of Engineering

Enhancing resilience in distributed networks in the face of malicious agents is an important problem for which many key theoretical results and applications require further development and characterization. This work focuses on the problem of distributed optimization in multi-agent cyberphysical systems, where a legitimate agent's dynamic is influenced both by the values it receives from potentially malicious neighboring agents, and by its own self-serving target function. We develop a new algorithmic and analytical framework to achieve resilience for the class of problems where stochastic values of trust between agents exist and can be exploited. In this case we show that convergence to the true global optimal point can be recovered, both in mean and almost surely, even in the presence of malicious agents. Furthermore, we provide expected convergence rate guarantees in the form of upper bounds on the expected squared distance to the optimal value. Finally, we present numerical results that validate the analytical convergence guarantees we present in this paper even when the malicious agents compose the majority of agents in the network.

Temporal alignment of sequences is a fundamental challenge in many applications, such as computer vision and bioinformatics, where local time shifting needs to be accounted for. Misalignment can lead to poor model generalization, especially in high-dimensional sequences. Existing methods often struggle with optimization when dealing with high-dimensional sparse data, falling into poor alignments. Feature selection is frequently used to enhance model performance for sparse data. However, a fixed set of selected features would not generally work for dynamically changing sequences and would need to be modified based on the state of the sequence. Therefore, modifying the selected feature based on contextual input would result in better alignment. Our suggested method, Conditional Deep Canonical Temporal Time Warping (CDCTW), is designed for temporal alignment in sparse temporal data to address these challenges. CDCTW enhances alignment accuracy for high dimensional time-dependent views be performing dynamic time warping on data embedded in maximally correlated subspace which handles sparsity with novel feature selection method. We validate the effectiveness of CDCTW through extensive experiments on various datasets, demonstrating superior performance over previous techniques.

In recent years, it has become popular to tackle image restoration tasks with a single pretrained diffusion model (DM) and data-fidelity guidance, instead of training a dedicated deep neural network per task. However, such "zero-shot" restoration schemes currently require many Neural Function Evaluations (NFEs) for performing well, which may be attributed to the many NFEs needed in the original generative functionality of the DMs. Recently, faster variants of DMs have been explored for image generation. These include Consistency Models (CMs), which can generate samples via a couple of NFEs. However, existing works that use guided CMs for restoration still require tens of NFEs or fine-tuning of the model per task that leads to performance drop if the assumptions during the fine-tuning are not accurate. In this paper, we propose a zero-shot restoration scheme that uses CMs and operates well with as little as 4 NFEs. It is based on a wise combination of several ingredients: better initialization, back-projection guidance, and above all a novel noise injection mechanism. We demonstrate the advantages of our approach for image super-resolution, deblurring and inpainting. Interestingly, we show that the usefulness of our noise injection technique goes beyond CMs: it can also mitigate the performance degradation of existing guided DM methods when reducing their NFE count.

Training and fine-tuning large language models (LLMs) come with challenges related to memory and computational requirements due to the increasing size of the model weights and the optimizer states. Various techniques have been developed to tackle these challenges, such as low-rank adaptation (LoRA), which involves introducing a parallel trainable low-rank matrix to the fixed pre-trained weights at each layer. However, these methods often fall short compared to the full-rank weight training approach, as they restrict the parameter search to a low-rank subspace. This limitation can disrupt training dynamics and require a full-rank warm start to mitigate the impact. In this paper, we introduce a new method inspired by a phenomenon we formally prove: as training progresses, the rank of the estimated layer gradients gradually decreases, and asymptotically approaches rank one. Leveraging this, our approach involves adaptively reducing the rank of the gradients during Adam optimization steps, using an efficient online-updating low-rank projections rule. We further present a randomized SVD scheme for efficiently finding the projection matrix. Our technique enables full-parameter fine-tuning with adaptive low-rank gradient updates, significantly reducing overall memory requirements during training compared to state-of-the-art methods while improving model performance in both pretraining and fine-tuning. Finally, we provide a convergence analysis of our method and demonstrate its merits for training and fine-tuning language and biological foundation models.

Accurate and reliable identification of the RTF between microphones with respect to a desired source is an essential component in the design of microphone array beamformers, specifically the MVDR criterion. Since an accurate estimation of the RTF in a noisy and reverberant environment is a cumbersome task, we aim at leveraging prior knowledge of the acoustic enclosure to robustify the RTF estimation by learning the RTF manifold. In this paper, we present a novel robust RTF identification method, tested and trained with real recordings, which relies on learning the RTF manifold using a GCN to infer a robust representation of the RTF in a confined area, and consequently enhance the beamformer's performance.

We study the question of whether a sequence of positive integers is the degree sequence of some outerplanar (a.k.a. 1-page book embeddable) graph G. If so, G is an outerplanar realization of d and d is an outerplanaric sequence. The case where is easy, as d has a realization by a forest (which is trivially an outerplanar graph). In this paper, we consider the family of all sequences d of even sum , where is the number of x’s in d. (The second inequality is a necessary condition for a sequence d with to be outerplanaric.) We partition into two disjoint subfamilies, , such that every sequence in is provably non-outerplanaric, and every sequence in is given a realizing graph G enjoying a 2-page book embedding (and moreover, one of the pages is also bipartite).

Most emotion recognition systems fail in real-life situations (in the wild scenarios) where the audio is contaminated by reverberation. Our study explores new methods to alleviate the performance degradation of Speech Emotion Recognition (SER) algorithms and develop a more robust system for adverse conditions. We propose processing multi-microphone signals to address these challenges and improve emotion classification accuracy. We adopt a state-of-the-art transformer model, the Hierarchical Token-semantic Audio Transformer (HTS-AT), to handle multi-channel audio inputs. We evaluate two strategies: averaging mel-spectrograms across channels and summing patch-embedded representations. Our multimicrophone model achieves superior performance compared to single-channel baselines when tested on real-world reverberant environments.

In this paper, we study the problem of fair multi-agent multi-arm bandit learning when agents do not communicate with each other, except collision information, provided to agents accessing the same arm simultaneously. We provide an algorithm with regret O(N3f(log T)log T) (assuming bounded rewards, with unknown bound), where f(t) is any function diverging to infinity with t. In contrast to optimal algorithms which share the rewards with a selected leader, our algorithm does not require a centralized collection of the arm rewards, allowing each agent to keep its rewards private. We also significantly improved previous privacy-preserving algorithms with the same upper bound on the regret of order O(f(log T)log T) but an exponential dependence on the number of agents. Simulation results present the dependence of the regret on log T.

Introducing a domain shift, such as a change in language or environment, to a well-trained speech enhancement system can cause severe performance degradation. Most current research assumes that a domain shift has already been detected and focuses on either supervised or unsupervised domain adaptation techniques. Here, we address the problem of automatically detecting when a domain shift has occurred. We present a domain shift detection method based on monitoring the confidence of a network that predicts the quality of enhanced speech. The experimental results show that our method can effectively detect a domain mismatch between the training and test sets.

Temporal alignment of sequences is a fundamental challenge in many applications, such as computer vision and bioinformatics, where local time shifting needs to be accounted for. Misalignment can lead to poor model generalization, especially in high-dimensional sequences. Existing methods often struggle with optimization when dealing with high-dimensional sparse data, falling into poor alignments. Feature selection is frequently used to enhance model performance for sparse data. However, a fixed set of selected features would not generally work for dynamically changing sequences and would need to be modified based on the state of the sequence. Therefore, modifying the selected feature based on contextual input would result in better alignment. Our suggested method, Conditional Deep Canonical Temporal Time Warping (CDCTW), is designed for temporal alignment in sparse temporal data to address …

Objectives This study aimed to assess the effectiveness of gold nanoparticles conjugated with anti-EGFR monoclonal antibodies (GNPs-EGFR) in distinguishing between benign and malignant salivary gland tumors. Methods A total of 49 oral salivary gland tissue samples were analyzed, including 22 malignant salivary gland tumors (MSGTs), 15 benign salivary gland tumors (BSGTs), and 12 control samples. For each sample, three 5 μm consecutive tissue sections were prepared. The first section was stained with hematoxylin and eosin (H&E) to confirm the diagnosis, the second was immunohistochemically stained for anti-EGFR, and the third was treated with GNPs-EGFR followed by hyperspectral microscopy to analyze the reflectance spectrum. Results Reflectance intensity was significantly higher (p < 0.001) in MSGTs compared to BSGTs and controls, with intensity levels increasing alongside tumor grade. The average hyperspectral reflectance values were strongly correlated with the GNPs-EGFR immunohistochemical score and varied significantly between subgroups (p < 0.001). Conclusions GNPs-EGFR reflection measurements effectively differentiate MSGTs from BSGTs with high sensitivity. This diffusion–reflection technique holds potential as a valuable tool for tumor detection, surgical margin assessment, and intraoperative identification of residual disease in salivary gland tumors. Objectives Methods Results Conclusions

Significance Alcohol consumption monitoring is essential for forensic and healthcare applications. While breath and blood alcohol concentration sensors are currently the most common methods, there is a growing need for faster, non-invasive, and more efficient assessment techniques. The rationale for our binary classification relates to law enforcement applications in countries with strict limits on alcohol consumption such as China, which seeks to prevent driving with even the smallest amount of alcohol in the bloodstream. Aim We propose a remote optical technique for assessing alcohol consumption using speckle pattern analysis, enhanced by machine learning for binary classification. This method offers remote and fast alcohol consumption evaluation without requiring before and after comparisons. Approach Our experimental setup includes a laser directed toward the subject’s radial artery, a camera capturing …

Machine unlearning aims to remove the influence of problematic training data after a model has been trained. The primary challenge in machine unlearning is ensuring that the process effectively removes specified data without compromising the model's overall performance on the remaining dataset. Many existing machine unlearning methods address this challenge by carefully balancing gradient ascent on the unlearn data with the gradient descent on a retain set representing the training data. Here, we propose OrthoGrad, a novel approach that mitigates interference between the unlearn set and the retain set rather than competing ascent and descent processes. Our method projects the gradient of the unlearn set onto the subspace orthogonal to all gradients in the retain batch, effectively avoiding any gradient interference. We demonstrate the effectiveness of OrthoGrad on multiple machine unlearning benchmarks, including automatic speech recognition, outperforming competing methods.

Urban energy efficiency and sustainability are critical challenges, as cities worldwide attempt to balance economic growth, environmental sustainability, and energy consumption. This systematic review examines the dynamics of urban energy management, focusing on how local authorities navigate energy transitions through efficiency measures, renewable energy adoption, and policy interventions. Specifically, it seeks to answer the following research question: how do local authorities implement energy-efficient practices and adopt renewable energy technologies to reduce emissions, optimize cost-effectiveness, and influence urban policy-making? The goal of this study is to assess the effectiveness of these approaches in different urban contexts. By reviewing 47 articles, this study identifies the unique characteristics of urban energy management and highlights the need for tailored, context-specific solutions, such as integrating decentralized renewable energy systems, optimizing building energy performance, and developing policy incentives that consider local socio-economic conditions. The findings reveal varying degrees of success among cities, with particular challenges in lower-income municipalities, where financial and institutional barriers hinder the implementation of sustainable energy projects. This study concludes that localized approaches and long-term strategies are essential for achieving sustainable urban energy transitions, offering a comprehensive perspective on the complexities of urban energy systems and their evolving policy landscape. Future research should focus on assessing the long-term impact of municipal energy …

Graph fragmentation aims to find the smallest vertex subset whose removal breaks a graph into components of bounded size. While this problem has applications in network dismantling and combinatorics, theoretical bounds on optimal solutions remain limited. We derive rigorous bounds for several graph classes, characterize hard instances, and illuminate the relationship between graph structure and optimal fragmentation strategies. Specifically, we show that for random d-regular graphs with n vertices, the minimal size of the fragmenting subset of nodes is asymptotically almost surely |S|≥d−22d−2n−o(n), and that asymptotically almost surely, n−2α(G)−o(n)≤|S|≤n−α(G)+o(n), where α(G) is the independence number of G. For d≫1, we prove that asymptotically almost surely, |S|/n≈1−logd/d. However, we show that the line graphs of random regular graphs are considerably harder to fragment, with |S|/n≥1−c/d for some constant c.

Leakage-resilient cryptography aims to protect cryptographic primitives from so-called “side channel attacks” that exploit their physical implementation to learn their input or secret state. Starting from the works of Ishai, Sahai and Wagner (CRYPTO ‘03) and Micali and Reyzin (TCC ‘04), most works on leakage-resilient cryptography either focus on protecting general computations, such as circuits or multiparty computation protocols, or on specific non-interactive primitives such as storage, encryption, and signatures. This work focuses on leakage resilience for the middle ground, namely for distributed and interactive cryptographic primitives. Our main technical contribution is designing the first secret sharing scheme that is equivocal, resists adaptive probing of a constant fraction of bits from each share, while incurs only a constant blowup in share size. Equivocation is a strong leakage-resilience guarantee, recently …

In content-oblivious computation, n nodes wish to compute a given task over an asynchronous network that suffers from an extremely harsh type of noise, which corrupts the content of all messages across all channels. In a recent work, Censor-Hillel, Cohen, Gelles, and Sela (Distributed Computing, 2023) showed how to perform arbitrary computations in a content-oblivious way in 2-edge connected networks but only if the network has a distinguished node (called root) to initiate the computation. Our goal is to remove this assumption, which was conjectured to be necessary. Achieving this goal essentially reduces to performing a content-oblivious leader election since an elected leader can then serve as the root required to perform arbitrary content-oblivious computations. We focus on ring networks, which are the simplest 2-edge connected graphs. On \emph{oriented} rings, we obtain a leader election algorithm with message complexity O(n * ID_max, where ID_max is the maximal assigned ID. As it turns out, this dependency on ID_max is inherent: we show a lower bound of Omega(n log(ID_max/n)) messages for content-oblivious leader election algorithms. We also extend our results to \emph{non-oriented} rings, where nodes cannot tell which channel leads to which neighbor. In this case, however, the algorithm does not terminate but only reaches quiescence.

We study the question of whether a sequence d=(d 1,…, d n) of positive integers is the degree sequence of some outerplanar graph G. If so, G is an outerplanar realization of d and d is an outerplanaric sequence. The case where∑ d≤ 2 n− 2 is easy, as d has a realization by a forest. In this paper, we consider the family D of all sequences d of even sum 2 n≤∑ d≤ 4 n− 6− 2 ω 1, where ω x is the number of x's in d. We partition D into two disjoint subfamilies, D= D N O P∪ D 2 P B E, such that every sequence in D N O P is provably non-outerplanaric, and every sequence in D 2 P B E is given a realizing graph G enjoying a 2-page book embedding (and moreover, one of the pages is also bipartite).

Objectives Dengue virus (DENV) infection is a significant global health concern, causing severe morbidity and mortality. While many cases present as a mild febrile illness, some progress to life-threatening severe dengue (SD). Early intervention is essential to improve outcomes, but current predictive methods lack specificity, burdening healthcare systems in endemic regions. Circulating long non-coding RNAs (lncRNAs) have emerged as stable and promising biomarkers. This study explored the use of lncRNAs as predictive markers for SD. Methods Differential expression and qPCR arrays were employed to identify lncRNAs associated with SD. Candidate lncRNAs were validated, and their plasma levels were measured in a cohort of Vietnamese dengue patients (n= 377) and healthy controls (n= 128) at admission. Machine learning algorithms were applied to predict the probability of SD progression. Results The …

Standard detection of entanglement relies on local measurements of the individual particles, evaluating their correlations in post-processing. For time-energy entangled photons, either times or energies are measured, but not both due to the mutual quantum uncertainty, providing only partial information of the entanglement. In principle, a global detector could recover the complete information of entanglement in a single shot if it could measure the combined correlated variables and without measuring the individual energies or times. Such a global measurement is possible using the reverse disentangling interaction, like sum-frequency generation (SFG), but nonlinear interactions at the single-photon level are ridiculously inefficient. We overcome this barrier by stimulating the nonlinear SFG interaction with a strong pump, thereby measuring both the energy-sum (SFG spectrum) and the time-difference (response to group delay/dispersion) simultaneously and efficiently. We generate bi-photons with extreme time-energy entanglement (octave-spanning spectrum of 113THz) and measure a relative uncertainty between time-difference and energy-sum of , violating the classical bound by >12 orders of magnitude. The presented coherent SFG dramatically enhances the detection SNR compared to standard methods since it ideally rejects erroneous coincidences in both time and energy, paving the way for sensing applications, such as quantum illumination (radar) and more.

Neurotechnological interfaces have the potential to create new forms of human-machine interactions, by allowing devices to interact directly with neurological signals instead of via intermediates such as keystrokes. Surface electromyography (sEMG) has been used extensively in myoelectric control systems, which use bioelectric activity recorded from muscles during contractions to classify actions. This technology has been used primarily for rehabilitation applications. In order to support the development of myoelectric interfaces for a broader range of human-machine interactions, we present an sEMG dataset obtained during key presses in a typing task. This fine-grained classification dataset consists of 16-channel bilateral sEMG recordings and key logs, collected from 19 individuals in two sessions on different days. We report baseline results on intra-session, inter-session and inter-subject evaluations. Our …