Search

Dissipative models for the quasi-static and dynamic response due to slip in an elastic body containing a single slip plane of vanishing thickness are developed. Discrete dislocations with continuously distributed cores can glide on this plane, and the models are developed as special cases of a fully three-dimensional theory of plasticity induced by dislocation motion. The reduced models are compared and contrasted with the augmented Peierls model of dislocation dynamics. A primary distinguishing feature of the reduced models is the a-priori accounting of space-time conservation of Burgers vector during dislocation evolution. A physical shortcoming of the developed models as well as the Peierls model with regard to a dependence on the choice of a distinguished, coherent reference configuration is discussed, and a testable model without such dependence is also proposed.

Vision-Language Models (VLMs) inherit adversarial vulnerabilities of Large Language Models (LLMs), which are further exacerbated by their multimodal nature. Existing defenses, including adversarial training, input transformations, and heuristic detection, are computationally expensive, architecture-dependent, and fragile against adaptive attacks. We introduce EigenShield, an inference-time defense leveraging Random Matrix Theory to quantify adversarial disruptions in high-dimensional VLM representations. Unlike prior methods that rely on empirical heuristics, EigenShield employs the spiked covariance model to detect structured spectral deviations. Using a Robustness-based Nonconformity Score (RbNS) and quantile-based thresholding, it separates causal eigenvectors, which encode semantic information, from correlational eigenvectors that are susceptible to adversarial artifacts. By projecting embeddings onto the causal subspace, EigenShield filters adversarial noise without modifying model parameters or requiring adversarial training. This architecture-independent, attack-agnostic approach significantly reduces the attack success rate, establishing spectral analysis as a principled alternative to conventional defenses. Our results demonstrate that EigenShield consistently outperforms all existing defenses, including adversarial training, UNIGUARD, and CIDER.

Bipartite quantum states with higher Schmidt numbers have been shown to outperform those with lower Schmidt numbers in various information processing tasks. Therefore, to ensure the efficient use of resources in these tasks, it is essential to certify the Schmidt number of the resource states. Ideally, this certification should rely as little as possible on the certifying devices, ensuring robustness against potential imperfections. In this work, we explore the scope of fully and partially device-independent Schmidt number certification methods. We demonstrate the general impossibility of fully device-independent certification for all states. Specifically, in a restricted setting, we present a class of states with Schmidt number 3, for which it is impossible to certify that their Schmidt number is greater than 2. However, we show that the Schmidt number of all states can be certified in a measurement-device-independent manner via semi-quantum nonlocal games, which assume trust in the preparation devices. Finally, we present an explicit construction of a semi-quantum game for the measurement-device-independent certification of a class of states., Comment: Initial draft; Comments are welcome; 9 pages, 1 figure

We present an analysis of millimeter CO observations to search and quantify signatures of molecular gas outflows. We exploit the large sample of $0.5 < z < 2.6$ galaxies observed as part of the PHIBSS1/2 surveys with the IRAM Plateau de Bure interferometer, focusing on the 154 typical massive star-forming galaxies with CO detections (mainly CO(3-2), but including also CO(2-1) and CO(6-5)) at signal-to-noise (SNR) > 1.5 and available properties (stellar mass, star formation rate, size) from ancillary data. None of the individual spectra exhibit a compelling signature of CO outflow emission even at high SNR > 7. To search for fainter outflow signatures, we carry out an analysis of stacked spectra, including the full sample, as well as subsets, split in terms of stellar mass, redshift, inclination, offset in star formation rate (SFR) from the main sequence, and AGN activity. None of the physically motivated subsamples show any outflow signature. We report a tentative detection in a subset statistically designed to maximize outflow signatures. We derive upper limits on molecular gas outflow rate and mass loading factors $\eta$ based on our results and find $\eta \leq$ 2.2-35.4, depending on the subsample. Much deeper CO data and observations of alternative tracers are needed to decisively constrain the importance of cold molecular gas component of outflows relative to other gas phases., Comment: 19 pages, 8 figures. Submitted to ApJ

Clinical Question Answering (CQA) plays a crucial role in medical decision-making, enabling physicians to extract relevant information from Electronic Medical Records (EMRs). While transformer-based models such as BERT, BioBERT, and ClinicalBERT have demonstrated state-of-the-art performance in CQA, existing models lack the ability to categorize extracted answers, which is critical for structured retrieval, content filtering, and medical decision support. To address this limitation, we introduce a Multi-Task Learning (MTL) framework that jointly trains CQA models for both answer extraction and medical categorization. In addition to predicting answer spans, our model classifies responses into five standardized medical categories: Diagnosis, Medication, Symptoms, Procedure, and Lab Reports. This categorization enables more structured and interpretable outputs, making clinical QA models more useful in real-world healthcare settings. We evaluate our approach on emrQA, a large-scale dataset for medical question answering. Results show that MTL improves F1-score by 2.2% compared to standard fine-tuning, while achieving 90.7% accuracy in answer categorization. These findings suggest that MTL not only enhances CQA performance but also introduces an effective mechanism for categorization and structured medical information retrieval.

Multiple official languages within a country along with languages common with other countries demand content consistency in both shared and unshared languages during information sharing. However, inconsistency due to conflict in content shared and content updates not propagated in languages between countries poses a problem. Towards addressing inconsistency, this research qualitatively studied traits for information sharing among countries inside global brands as depicted by content shared in their country-specific websites. First, inconsistency in content shared is illustrated among websites highlighting the problem in information sharing among countries. Second, content propagation among countries that vary in scales and coupling for specific content categories are revealed. Scales suggested that corporate and customer support related information tend to be shared globally and locally respectively while product related information is both locally and regionally suitable for sharing. Higher occurrences of propagation when sharing corporate related information also showed tendency for high coupling between websites suggesting the suitability for rigid consistency policy compared to other categories. This study also proposed a simplistic approach with pattern of sharing to enable consistent information sharing.

Generating multiple counterfactual explanations for different modes within a class presents a significant challenge, as these modes are distinct yet converge under the same classification. Diffusion probabilistic models (DPMs) have demonstrated a strong ability to capture the underlying modes of data distributions. In this paper, we harness the power of a Diffusion Autoencoder to generate multiple distinct counterfactual explanations. By clustering in the latent space, we uncover the directions corresponding to the different modes within a class, enabling the generation of diverse and meaningful counterfactuals. We introduce a novel methodology, DifCluE, which consistently identifies these modes and produces more reliable counterfactual explanations. Our experimental results demonstrate that DifCluE outperforms the current state-of-the-art in generating multiple counterfactual explanations, offering a significant advancement in model interpretability.

This study examines the growing use of Large Language Models (LLMs) in child-centered applications, highlighting safety and ethical concerns such as bias, harmful content, and cultural insensitivity. Despite their potential to enhance learning, there is a lack of standardized frameworks to mitigate these risks. Through a systematic literature review, we identify key parental and empirical concerns, including toxicity and ethical breaches in AI outputs. Moreover, to address these issues, this paper proposes a protection framework for safe Child-LLM interaction, incorporating metrics for content safety, behavioral ethics, and cultural sensitivity. The framework provides practical tools for evaluating LLM safety, offering guidance for developers, policymakers, and educators to ensure responsible AI deployment for children.

Dynamic Text-Attributed Graphs (DyTAGs) have numerous real-world applications, e.g. social, collaboration, citation, communication, and review networks. In these networks, nodes and edges often contain text descriptions, and the graph structure can evolve over time. Future link prediction, edge classification, relation generation, and other downstream tasks on DyTAGs require powerful representations that encode structural, temporal, and textual information. Although graph neural networks (GNNs) excel at handling structured data, encoding temporal information within dynamic graphs remains a significant challenge. In this work, we propose LLM-driven Knowledge Distillation for Dynamic Text Attributed Graph (LKD4DyTAG) with temporal encoding to address these challenges. We use a simple, yet effective approach to encode temporal information in edges so that graph convolution can simultaneously capture both temporal and structural information in the hidden representations. To leverage LLM's text processing capabilities for learning richer representations on DyTAGs, we distill knowledge from LLM-driven edge representations (based on a neighborhood's text attributes) into saptio-temporal representations using a lightweight GNN model that encodes temporal and structural information. The objective of knowledge distillation enables the GNN to learn representations that more effectively encode the available structural, temporal, and textual information in DyTAG. We conducted extensive experimentation on six real-world DyTAG datasets to verify the effectiveness of our approach LKD4DyTAG for future link prediction and edge classification task. The results show that our approach significantly improves the performance of downstream tasks compared to the baseline models., Comment: Accepted at the AI4TS: AI for Time Series Analysis workshop, AAAI 2025

Automated mammography screening plays an important role in early breast cancer detection. However, current machine learning models, developed on some training datasets, may exhibit performance degradation and bias when deployed in real-world settings. In this paper, we analyze the performance of high-performing AI models on two mammography datasets-the Emory Breast Imaging Dataset (EMBED) and the RSNA 2022 challenge dataset. Specifically, we evaluate how these models perform across different subgroups, defined by six attributes, to detect potential biases using a range of classification metrics. Our analysis identifies certain subgroups that demonstrate notable underperformance, highlighting the need for ongoing monitoring of these subgroups' performance. To address this, we adopt a monitoring method designed to detect performance drifts over time. Upon identifying a drift, this method issues an alert, which can enable timely interventions. This approach not only provides a tool for tracking the performance but also helps ensure that AI models continue to perform effectively across diverse populations.

Reasoning LLMs such as OpenAI o1, o3 and DeepSeek R1 have made significant progress in mathematics and coding, yet find challenging advanced tasks such as International Mathematical Olympiad (IMO) combinatorics problems, Abstraction and Reasoning Corpus (ARC) puzzles, and Humanity's Last Exam (HLE) questions. We use a diverse inference approach that combines multiple models and methods at test time. We find that verifying mathematics and code problems, and rejection sampling on other problems is simple and effective. We automatically verify correctness of solutions to IMO problems by Lean, and ARC puzzles by code, and find that best-of-N effectively answers HLE questions. Our approach increases answer accuracy on IMO combinatorics problems from 33.3% to 77.8%, accuracy on HLE questions from 8% to 37%, and solves 80% of ARC puzzles that 948 humans could not and 26.5% of ARC puzzles that o3 high compute does not. Test-time simulations, reinforcement learning, and meta-learning with inference feedback improve generalization by adapting agent graph representations and varying prompts, code, and datasets. Our approach is reliable, robust, and scalable, and in the spirit of reproducible research, we will make it publicly available upon publication., Comment: 165 pages

We introduce Granite Vision, a lightweight large language model with vision capabilities, specifically designed to excel in enterprise use cases, particularly in visual document understanding. Our model is trained on a comprehensive instruction-following dataset, including document-related tasks, such as content extraction from tables, charts, diagrams, sketches, and infographics, as well as general image tasks. The architecture of Granite Vision is centered around visual modality alignment with a decoder-only, 2 billion parameter Granite large language model. Additionally, we introduce a dedicated safety classification approach in test-time that leverages a sparse set of attention vectors to identify potential harmful inputs. Despite its lightweight architecture, Granite Vision achieves strong results in standard benchmarks related to visual document understanding, as well as on the LiveXiv benchmark, which is designed to avoid test set contamination by using a constantly updated corpus of recently published Arxiv papers. We are releasing the model under the Apache-2 license, allowing for both research and commercial use, while offering complete visibility into the training data and other relevant details. See https://huggingface.co/ibm-granite/ for model weights.

Jailbreak attacks aim to exploit large language models (LLMs) and pose a significant threat to their proper conduct; they seek to bypass models' safeguards and often provoke transgressive behaviors. However, existing automatic jailbreak attacks require extensive computational resources and are prone to converge on suboptimal solutions. In this work, we propose \textbf{C}ompliance \textbf{R}efusal \textbf{I}nitialization (CRI), a novel, attack-agnostic framework that efficiently initializes the optimization in the proximity of the compliance subspace of harmful prompts. By narrowing the initial gap to the adversarial objective, CRI substantially improves adversarial success rates (ASR) and drastically reduces computational overhead -- often requiring just a single optimization step. We evaluate CRI on the widely-used AdvBench dataset over the standard jailbreak attacks of GCG and AutoDAN. Results show that CRI boosts ASR and decreases the median steps to success by up to \textbf{\(\times 60\)}. The project page, along with the reference implementation, is publicly available at \texttt{https://amit1221levi.github.io/CRI-Jailbreak-Init-LLMs-evaluation/}.

Diffusion models enable high-quality and diverse visual content synthesis. However, they struggle to generate rare or unseen concepts. To address this challenge, we explore the usage of Retrieval-Augmented Generation (RAG) with image generation models. We propose ImageRAG, a method that dynamically retrieves relevant images based on a given text prompt, and uses them as context to guide the generation process. Prior approaches that used retrieved images to improve generation, trained models specifically for retrieval-based generation. In contrast, ImageRAG leverages the capabilities of existing image conditioning models, and does not require RAG-specific training. Our approach is highly adaptable and can be applied across different model types, showing significant improvement in generating rare and fine-grained concepts using different base models. Our project page is available at: https://rotem-shalev.github.io/ImageRAG

Shortcut learning, where machine learning models exploit spurious correlations in data instead of capturing meaningful features, poses a significant challenge to building robust and generalizable models. This phenomenon is prevalent across various machine learning applications, including vision, natural language processing, and speech recognition, where models may find unintended cues that minimize training loss but fail to capture the underlying structure of the data. Vision classifiers such as Convolutional Neural Networks (CNNs), Multi-Layer Perceptrons (MLPs), and Vision Transformers (ViTs) leverage distinct architectural principles to process spatial and structural information, making them differently susceptible to shortcut learning. In this study, we systematically evaluate these architectures by introducing deliberate shortcuts into the dataset that are positionally correlated with class labels, creating a controlled setup to assess whether models rely on these artificial cues or learn actual distinguishing features. We perform both quantitative evaluation by training on the shortcut-modified dataset and testing them on two different test sets -- one containing the same shortcuts and another without them -- to determine the extent of reliance on shortcuts. Additionally, qualitative evaluation is performed by using network inversion-based reconstruction techniques to analyze what the models internalize in their weights, aiming to reconstruct the training data as perceived by the classifiers. We evaluate shortcut learning behavior across multiple benchmark datasets, including MNIST, Fashion-MNIST, SVHN, and CIFAR-10, to compare the susceptibility of different vision classifier architectures to shortcut reliance and assess their varying degrees of sensitivity to spurious correlations.

Chronic kidney disease (CKD) is a major global health issue, affecting over 10% of the population and causing significant mortality. While kidney biopsy remains the gold standard for CKD diagnosis and treatment, the lack of comprehensive benchmarks for kidney pathology segmentation hinders progress in the field. To address this, we organized the Kidney Pathology Image Segmentation (KPIs) Challenge, introducing a dataset that incorporates preclinical rodent models of CKD with over 10,000 annotated glomeruli from 60+ Periodic Acid Schiff (PAS)-stained whole slide images. The challenge includes two tasks, patch-level segmentation and whole slide image segmentation and detection, evaluated using the Dice Similarity Coefficient (DSC) and F1-score. By encouraging innovative segmentation methods that adapt to diverse CKD models and tissue conditions, the KPIs Challenge aims to advance kidney pathology analysis, establish new benchmarks, and enable precise, large-scale quantification for disease research and diagnosis.

Sorting is one of the most basic primitives in many algorithms and data analysis tasks. Comparison-based sorting algorithms, like quick-sort and merge-sort, are known to be optimal when the outcome of each comparison is error-free. However, many real-world sorting applications operate in scenarios where the outcome of each comparison can be noisy. In this work, we explore settings where a bounded number of comparisons are potentially corrupted by erroneous agents, resulting in arbitrary, adversarial outcomes. We model the sorting problem as a query-limited tournament graph where edges involving erroneous nodes may yield arbitrary results. Our primary contribution is a randomized algorithm inspired by quick-sort that, in expectation, produces an ordering close to the true total order while only querying $\tilde{O}(n)$ edges. We achieve a distance from the target order $\pi$ within $(3 + \epsilon)|B|$, where $B$ is the set of erroneous nodes, balancing the competing objectives of minimizing both query complexity and misalignment with $\pi$. Our algorithm needs to carefully balance two aspects: identify a pivot that partitions the vertex set evenly and ensure that this partition is "truthful" and yet query as few "triangles" in the graph $G$ as possible. Since the nodes in $B$ can potentially hide in an intricate manner, our algorithm requires several technical steps. Additionally, we demonstrate significant implications for the Ulam-$k$-Median problem, a classical clustering problem where the metric is defined on the set of permutations on a set of $d$ elements. Chakraborty, Das, and Krauthgamer gave a $(2-\varepsilon)$ FPT approximation algorithm for this problem, where the running time is super-linear in both $n$ and $d$. We use our robust sorting framework to give the first $(2-\varepsilon)$ FPT linear time approximation algorithm for this problem., Comment: Abstract shortened to meet arXiv requirements

Safety-critical perception systems require both reliable uncertainty quantification and principled abstention mechanisms to maintain safety under diverse operational conditions. We present a novel dual-threshold conformalization framework that provides statistically-guaranteed uncertainty estimates while enabling selective prediction in high-risk scenarios. Our approach uniquely combines a conformal threshold ensuring valid prediction sets with an abstention threshold optimized through ROC analysis, providing distribution-free coverage guarantees (>= 1 - alpha) while identifying unreliable predictions. Through comprehensive evaluation on CIFAR-100, ImageNet1K, and ModelNet40 datasets, we demonstrate superior robustness across camera and LiDAR modalities under varying environmental perturbations. The framework achieves exceptional detection performance (AUC: 0.993 to 0.995) under severe conditions while maintaining high coverage (>90.0%) and enabling adaptive abstention (13.5% to 63.4% +/- 0.5) as environmental severity increases. For LiDAR-based perception, our approach demonstrates particularly strong performance, maintaining robust coverage (>84.5%) while appropriately abstaining from unreliable predictions. Notably, the framework shows remarkable stability under heavy perturbations, with detection performance (AUC: 0.995 +/- 0.001) significantly outperforming existing methods across all modalities. Our unified approach bridges the gap between theoretical guarantees and practical deployment needs, offering a robust solution for safety-critical autonomous systems operating in challenging real-world conditions.

We present here a new hybrid scheme that combines a discontinuous Galerkin (DG) method with finite volume (FV) and finite difference (FD) methods. The computational mesh is divided into smaller elements that touch but do not overlap. Like a pure DG method, our new hybrid scheme requires information exchange only at the surface of neighboring elements. This avoids the need for ghostzones that are usually many points deep in traditional FV implementations. Furthermore, unlike traditional FV implementations, that require information exchange between each element and its 26 surrounding neighbors on non-cuboid meshes, our new hybrid method exchanges information only between each element and its six nearest neighbors. Through this reduction in communication, we aim to retain the high scalability of DG when using large supercomputers. The goal is to use DG in elements with smooth matter fields and to fall back onto the more robust FV/FD method in elements that contain non-smooth shocks or star surfaces. For this we devise trouble criteria to decide whether an element should be evolved with DG or FV/FD. We use the Nmesh program to implement and test the new scheme. We successfully evolve various single neutron star cases. These include the challenging cases of a neutron star initially in an unstable equilibrium migrating to a stable configuration and a boosted neutron star. These cases are simulated for the first time here in full 3D with general relativistic hydrodynamics using DG methods. We also describe additional numerical methods, such as the limiters and the atmosphere treatment we need for our simulations., Comment: 34 pages, 20 figures

Large Language and Vision-Language Models (LLMs/VLMs) are increasingly used in safety-critical applications, yet their opaque decision-making complicates risk assessment and reliability. Uncertainty quantification (UQ) helps assess prediction confidence and enables abstention when uncertainty is high. Conformal prediction (CP), a leading UQ method, provides statistical guarantees but relies on static thresholds, which fail to adapt to task complexity and evolving data distributions, leading to suboptimal trade-offs in accuracy, coverage, and informativeness. To address this, we propose learnable conformal abstention, integrating reinforcement learning (RL) with CP to optimize abstention thresholds dynamically. By treating CP thresholds as adaptive actions, our approach balances multiple objectives, minimizing prediction set size while maintaining reliable coverage. Extensive evaluations across diverse LLM/VLM benchmarks show our method outperforms Least Ambiguous Classifiers (LAC) and Adaptive Prediction Sets (APS), improving accuracy by up to 3.2%, boosting AUROC for hallucination detection by 22.19%, enhancing uncertainty-guided selective generation (AUARC) by 21.17%, and reducing calibration error by 70%-85%. These improvements hold across multiple models and datasets while consistently meeting the 90% coverage target, establishing our approach as a more effective and flexible solution for reliable decision-making in safety-critical applications. The code is available at: {https://github.com/sinatayebati/vlm-uncertainty}.

Realizing the vision of using AI agents to automate critical IT tasks depends on the ability to measure and understand effectiveness of proposed solutions. We introduce ITBench, a framework that offers a systematic methodology for benchmarking AI agents to address real-world IT automation tasks. Our initial release targets three key areas: Site Reliability Engineering (SRE), Compliance and Security Operations (CISO), and Financial Operations (FinOps). The design enables AI researchers to understand the challenges and opportunities of AI agents for IT automation with push-button workflows and interpretable metrics. ITBench includes an initial set of 94 real-world scenarios, which can be easily extended by community contributions. Our results show that agents powered by state-of-the-art models resolve only 13.8% of SRE scenarios, 25.2% of CISO scenarios, and 0% of FinOps scenarios. We expect ITBench to be a key enabler of AI-driven IT automation that is correct, safe, and fast.

Weakly-collisional plasmas, such as the solar wind or the intra-cluster medium (ICM) of galaxy clusters, evolve in the presence of dynamically strong magnetic fields. The turbulent dynamo can amplify magnetic fields to such levels by converting turbulent kinetic energy into magnetic energy. While extensively studied in collisional magnetohydrodynamic (MHD) simulations, the weakly-collisional regime has only been explored recently. Here, we determine the properties of the weakly-collisional turbulent dynamo in the exponential ``kinematic" growth phase in both the subsonic and the previously unexplored supersonic regime of turbulence, using hybrid particle-in-cell (HPIC) and MHD simulations. We conduct a large parameter study, fixing the magnetic Reynolds number, Rm = 500, and the initial ratio of the magnetic to kinetic energy, $(E_{\rm{mag}}/E_{\rm{kin}})_{0} = 10^{-10}$, and then vary the kinetic Reynolds number, Re = 500, 50, and 5, for the MHD simulations. In the HPIC runs, only Rm = 500 is controlled, while Re emerges self-consistently from wave-particle interactions. We find that the velocity and magnetic field structures, probability distribution functions, and power spectra of the HPIC runs are similar to that of the MHD dynamo with Re ~ 50-500 and Re ~ 500 in the subsonic and supersonic regimes, respectively. Using MHD scaling relations, we infer $\text{Re}_{\rm inferred}=480^{+170}_{-250}$ and $690^{+360}_{-360}$ in the subsonic and supersonic weakly-collisional plasma, respectively. Overall, we find that the turbulent dynamo shares similar physical properties in both weakly-collisional and collisional plasmas. Our results of the weakly-collisional turbulent dynamo may have relevant applications to the solar wind, weakly-collisional shocks, and the hot ICM., Comment: 16 pages, 13 figures, Submitted to Monthly Notices of the Royal Astronomical Society

E-scooters have become a more dominant mode of transport in recent years. However, the rise in their usage has been accompanied by an increase in injuries, affecting the trust and perceived safety of both users and non-users. Artificial intelligence (AI), as a cutting-edge and widely applied technology, has demonstrated potential to enhance transportation safety, particularly in driver assistance systems. The integration of AI into e-scooters presents a promising approach to addressing these safety concerns. This study aims to explore the factors influencing individuals willingness to use AI-assisted e-scooters. Data were collected using a structured questionnaire, capturing responses from 405 participants. The questionnaire gathered information on demographic characteristics, micromobility usage frequency, road users' perception of safety around e-scooters, perceptions of safety in AI-enabled technology, trust in AI-enabled e-scooters, and involvement in e-scooter crash incidents. To examine the impact of demographic factors on participants' preferences between AI-assisted and regular e-scooters, decision tree analysis is employed, indicating that ethnicity, income, and age significantly influence preferences. To analyze the impact of other factors on the willingness to use AI-enabled e-scooters, a full-scale Structural Equation Model (SEM) is applied, revealing that the perception of safety in AI enabled technology and the level of trust in AI-enabled e-scooters are the strongest predictors.

We apply language modeling techniques to detect ransomware activity in NVMe command sequences. We design and train two types of transformer-based models: the Command-Level Transformer (CLT) performs in-context token classification to determine whether individual commands are initiated by ransomware, and the Patch-Level Transformer (PLT) predicts the volume of data accessed by ransomware within a patch of commands. We present both model designs and the corresponding tokenization and embedding schemes and show that they improve over state-of-the-art tabular methods by up to 24% in missed-detection rate, 66% in data loss prevention, and 84% in identifying data accessed by ransomware., Comment: 25 pages, 8 figures

We propose SPARC, a lightweight continual learning framework for large language models (LLMs) that enables efficient task adaptation through prompt tuning in a lower-dimensional space. By leveraging principal component analysis (PCA), we identify a compact subspace of the training data. Optimizing prompts in this lower-dimensional space enhances training efficiency, as it focuses updates on the most relevant features while reducing computational overhead. Furthermore, since the model's internal structure remains unaltered, the extensive knowledge gained from pretraining is fully preserved, ensuring that previously learned information is not compromised during adaptation. Our method achieves high knowledge retention in both task-incremental and domain-incremental continual learning setups while fine-tuning only 0.04% of the model's parameters. Additionally, by integrating LoRA, we enhance adaptability to computational constraints, allowing for a tradeoff between accuracy and training cost. Experiments on the SuperGLUE benchmark demonstrate that our PCA-based prompt tuning combined with LoRA maintains full knowledge retention while improving accuracy, utilizing only 1% of the model's parameters. These results establish our approach as a scalable and resource-efficient solution for continual learning in LLMs.

Precise alignment in Text-to-Image (T2I) systems is crucial to ensure that generated visuals not only accurately encapsulate user intents but also conform to stringent ethical and aesthetic benchmarks. Incidents like the Google Gemini fiasco, where misaligned outputs triggered significant public backlash, underscore the critical need for robust alignment mechanisms. In contrast, Large Language Models (LLMs) have achieved notable success in alignment. Building on these advancements, researchers are eager to apply similar alignment techniques, such as Direct Preference Optimization (DPO), to T2I systems to enhance image generation fidelity and reliability. We present YinYangAlign, an advanced benchmarking framework that systematically quantifies the alignment fidelity of T2I systems, addressing six fundamental and inherently contradictory design objectives. Each pair represents fundamental tensions in image generation, such as balancing adherence to user prompts with creative modifications or maintaining diversity alongside visual coherence. YinYangAlign includes detailed axiom datasets featuring human prompts, aligned (chosen) responses, misaligned (rejected) AI-generated outputs, and explanations of the underlying contradictions.

The outstanding opto-electronic properties of lead halide perovskites have been related to the formation of polarons. Nevertheless, the observation of the atomistic deformation brought about by one electron-hole pair in these materials has remained elusive. Here, we measure the diffraction patterns of single CsPbBr$_3$ quantum dots (QDs) with and without resonant excitation in the single exciton limit using serial femtosecond crystallography (SFX). By reconstructing the 3D differential diffraction pattern, we observe small shifts of the Bragg peaks indicative of a crystal-wide deformation field. Building on DFT calculations, we show that these shifts are consistent with the lattice distortion induced by a delocalized electron and a localized hole, forming a mixed large/small exciton polaron. This result creates a clear picture of the polaronic deformation in CsPbBr$_3$ QDs, highlights the exceptional sensitivity of SFX to lattice distortions in few-nanometer crystallites, and establishes an experimental platform for future studies of electron-lattice interactions., Comment: Main: 12 pages, 5 figures; Supplemental: 21 pages, 11 figures

Autonomous edge computing in robotics, smart cities, and autonomous vehicles relies on the seamless integration of sensing, processing, and actuation for real-time decision-making in dynamic environments. At its core is the sensing-to-action loop, which iteratively aligns sensor inputs with computational models to drive adaptive control strategies. These loops can adapt to hyper-local conditions, enhancing resource efficiency and responsiveness, but also face challenges such as resource constraints, synchronization delays in multi-modal data fusion, and the risk of cascading errors in feedback loops. This article explores how proactive, context-aware sensing-to-action and action-to-sensing adaptations can enhance efficiency by dynamically adjusting sensing and computation based on task demands, such as sensing a very limited part of the environment and predicting the rest. By guiding sensing through control actions, action-to-sensing pathways can improve task relevance and resource use, but they also require robust monitoring to prevent cascading errors and maintain reliability. Multi-agent sensing-action loops further extend these capabilities through coordinated sensing and actions across distributed agents, optimizing resource use via collaboration. Additionally, neuromorphic computing, inspired by biological systems, provides an efficient framework for spike-based, event-driven processing that conserves energy, reduces latency, and supports hierarchical control--making it ideal for multi-agent optimization. This article highlights the importance of end-to-end co-design strategies that align algorithmic models with hardware and environmental dynamics and improve cross-layer interdependencies to improve throughput, precision, and adaptability for energy-efficient edge autonomy in complex environments.

In this work, real photon production scattering rates from jet-medium interactions in the quark-gluon plasma (QGP) is perturbatively calculated using the higher-twist (HT) formalism. Focus is given towards real photon production from a highly virtual (and highly energetic) quark, taking into account heavy-quark mass scales [Phys. Rev. C 94, 054902 (2016)], fermion-boson conversion processes [Nucl. Phys. A 793, 128-170 (2007)], as well as coherence effects [Phys. Rev. C 105, 024908 (2022)]. A generalized factorization procedure, such as that used in e-A deep-inelastic scattering, is employed to derive an improved single-scattering medium-induced photon emission kernel that goes beyond the traditional in-medium gluon exchange approximation. Diagrams are classified based on the final state particles, and include four types of scattering kernels at $O(\alpha_{s}\alpha_{em})$ giving the following final states: (i) real photon and real quark, (ii) real photon and real gluon (iii) virtual photon corrections to quark-antiquark pair-production and (iv) virtual photon correction to quark-quark production. The collision-kernel, thus derived, includes full phase factors from all non-vanishing diagrams and complete second-order derivative terms in the transverse momentum gradient expansion. Moreover, the calculation includes heavy-quark mass effects, thus exploring heavy-quark energy loss. The in-medium parton distribution functions, and the related jet transport coefficients, have a hard transverse momentum dependence (of the emitted gluon or photon) present within the phase factor. It is observed that the jet transport coefficients resemble the transverse-momentum-dependent parton distribution functions., Comment: 53 pages, 26 figures; Added few corrections

The discovery of new materials is essential for enabling technological advancements. Computational approaches for predicting novel materials must effectively learn the manifold of stable crystal structures within an infinite design space. We introduce Open Materials Generation (OMG), a unifying framework for the generative design and discovery of inorganic crystalline materials. OMG employs stochastic interpolants (SI) to bridge an arbitrary base distribution to the target distribution of inorganic crystals via a broad class of tunable stochastic processes, encompassing both diffusion models and flow matching as special cases. In this work, we adapt the SI framework by integrating an equivariant graph representation of crystal structures and extending it to account for periodic boundary conditions in unit cell representations. Additionally, we couple the SI flow over spatial coordinates and lattice vectors with discrete flow matching for atomic species. We benchmark OMG's performance on two tasks: Crystal Structure Prediction (CSP) for specified compositions, and 'de novo' generation (DNG) aimed at discovering stable, novel, and unique structures. In our ground-up implementation of OMG, we refine and extend both CSP and DNG metrics compared to previous works. OMG establishes a new state-of-the-art in generative modeling for materials discovery, outperforming purely flow-based and diffusion-based implementations. These results underscore the importance of designing flexible deep learning frameworks to accelerate progress in materials science.

Directional couplers are a fundamental building block in integrated photonics, particularly in quantum applications and optimization-based design where precision is critical. Accurate functionality is crucial to ensure reliable operation within classical and quantum circuits. However, discrepancies between simulations and measurements are frequently observed. These inaccuracies can compromise the performance and scalability of integrated photonic systems, underscoring the critical need for advanced, precise simulation methods that bridge the gap between design and implementation. In this work, we show that this discrepancy can be mainly attributed to density changes in the oxide cladding. We conduct a systematic study involving experimental optical measurements, numerical simulations, and direct electron microscopy imaging to investigate this discrepancy in directional couplers. We find that the impact of cladding density variations on performance increases as feature gaps shrink. By incorporating these effects into our simulations using a novel and physically motivated Effective Trench Medium Model (ETMM), we achieve highly accurate reproduction of experimental measurements. We quantify the effects of cladding density variations on the SU(2) symmetry parameters that govern light propagation in directional couplers. This insight is crucial for advancing the precision of compact device fabrication, enabling reliable simulation of photonic integrated devices.

Despite tremendous recent progress, generative video models still struggle to capture real-world motion, dynamics, and physics. We show that this limitation arises from the conventional pixel reconstruction objective, which biases models toward appearance fidelity at the expense of motion coherence. To address this, we introduce VideoJAM, a novel framework that instills an effective motion prior to video generators, by encouraging the model to learn a joint appearance-motion representation. VideoJAM is composed of two complementary units. During training, we extend the objective to predict both the generated pixels and their corresponding motion from a single learned representation. During inference, we introduce Inner-Guidance, a mechanism that steers the generation toward coherent motion by leveraging the model's own evolving motion prediction as a dynamic guidance signal. Notably, our framework can be applied to any video model with minimal adaptations, requiring no modifications to the training data or scaling of the model. VideoJAM achieves state-of-the-art performance in motion coherence, surpassing highly competitive proprietary models while also enhancing the perceived visual quality of the generations. These findings emphasize that appearance and motion can be complementary and, when effectively integrated, enhance both the visual quality and the coherence of video generation. Project website: https://hila-chefer.github.io/videojam-paper.github.io/

We develop a measurement-based protocol for simultaneously purifying arbitrary logical states in multiple quantum error correcting codes with unit fidelity and finite probability, starting from arbitrary thermal states of each code. The protocol entails a time evolution caused by an engineered Hamiltonian, which results in transitions between the logical and error subspaces of the quantum error correcting code mediated by the auxiliary qubit, followed by a projective measurement in an optimum basis on the auxiliary qubit and an appropriate post-selection of the measurement outcomes. We illustrate the results with the three-qubit repetition code and the logical qubit used in quantum state transfer protocol. We further demonstrate that when the measurement base is not optimal, it is possible to achieve both classical fidelity, and fidelity as high as $90\%$ through several iterations of the purifying procedure, thereby establishing its robustness against variations in the measurement basis. By repeating the purification rounds, we show that purifying the cardinal states of the logical Bloch sphere corresponding to logical qubits in quantum state transfer is feasible utilizing paradigmatic quantum spin models as the generator of the time evolution., Comment: 12 pages, 6 figures

In an era dominated by datafication, the reduction of human experiences to quantifiable metrics raises profound philosophical and ethical questions. This paper explores these issues through the lens of Meursault, the protagonist of Albert Camus' The Stranger, whose emotionally detached existence epitomizes the existential concept of absurdity. Using natural language processing (NLP) techniques including emotion detection (BERT), sentiment analysis (VADER), and named entity recognition (spaCy)-this study quantifies key events and behaviors in Meursault's life. Our analysis reveals the inherent limitations of applying algorithmic models to complex human experiences, particularly those rooted in existential alienation and moral ambiguity. By examining how modern AI tools misinterpret Meursault's actions and emotions, this research underscores the broader ethical dilemmas of reducing nuanced human narratives to data points, challenging the foundational assumptions of our data-driven society. The findings presented in this paper serve as a critique of the increasing reliance on data-driven narratives and advocate for incorporating humanistic values in artificial intelligence., Comment: 7 pages, 9 figures, 4 tables

In this work, we present INTACT, a novel two-phase framework designed to enhance the robustness of deep neural networks (DNNs) against noisy LiDAR data in safety-critical perception tasks. INTACT combines meta-learning with adversarial curriculum training (ACT) to systematically address challenges posed by data corruption and sparsity in 3D point clouds. The meta-learning phase equips a teacher network with task-agnostic priors, enabling it to generate robust saliency maps that identify critical data regions. The ACT phase leverages these saliency maps to progressively expose a student network to increasingly complex noise patterns, ensuring targeted perturbation and improved noise resilience. INTACT's effectiveness is demonstrated through comprehensive evaluations on object detection, tracking, and classification benchmarks using diverse datasets, including KITTI, Argoverse, and ModelNet40. Results indicate that INTACT improves model robustness by up to 20% across all tasks, outperforming standard adversarial and curriculum training methods. This framework not only addresses the limitations of conventional training strategies but also offers a scalable and efficient solution for real-world deployment in resource-constrained safety-critical systems. INTACT's principled integration of meta-learning and adversarial training establishes a new paradigm for noise-tolerant 3D perception in safety-critical applications. INTACT improved KITTI Multiple Object Tracking Accuracy (MOTA) by 9.6% (64.1% -> 75.1%) and by 12.4% under Gaussian noise (52.5% -> 73.7%). Similarly, KITTI mean Average Precision (mAP) rose from 59.8% to 69.8% (50% point drop) and 49.3% to 70.9% (Gaussian noise), highlighting the framework's ability to enhance deep learning model resilience in safety-critical object tracking scenarios.

Transformer-based video super-resolution (VSR) models have set new benchmarks in recent years, but their substantial computational demands make most of them unsuitable for deployment on resource-constrained devices. Achieving a balance between model complexity and output quality remains a formidable challenge in VSR. Although lightweight models have been introduced to address this issue, they often struggle to deliver state-of-the-art performance. We propose a novel lightweight, parameter-efficient deep residual deformable convolution network for VSR. Unlike prior methods, our model enhances feature utilization through residual connections and employs deformable convolution for precise frame alignment, addressing motion dynamics effectively. Furthermore, we introduce a single memory tensor to capture information accrued from the past frames and improve motion estimation across frames. This design enables an efficient balance between computational cost and reconstruction quality. With just 2.3 million parameters, our model achieves state-of-the-art SSIM of 0.9175 on the REDS4 dataset, surpassing existing lightweight and many heavy models in both accuracy and resource efficiency. Architectural insights from our model pave the way for real-time VSR on streaming data.

Preference-based reinforcement learning (RL) offers a promising approach for aligning policies with human intent but is often constrained by the high cost of human feedback. In this work, we introduce PrefVLM, a framework that integrates Vision-Language Models (VLMs) with selective human feedback to significantly reduce annotation requirements while maintaining performance. Our method leverages VLMs to generate initial preference labels, which are then filtered to identify uncertain cases for targeted human annotation. Additionally, we adapt VLMs using a self-supervised inverse dynamics loss to improve alignment with evolving policies. Experiments on Meta-World manipulation tasks demonstrate that PrefVLM achieves comparable or superior success rates to state-of-the-art methods while using up to 2 x fewer human annotations. Furthermore, we show that adapted VLMs enable efficient knowledge transfer across tasks, further minimizing feedback needs. Our results highlight the potential of combining VLMs with selective human supervision to make preference-based RL more scalable and practical.

The fast and accurate estimation of planetary mass-loss rates is critical for planet population and evolution modelling. We use machine learning (ML) for fast interpolation across an existing large grid of hydrodynamic upper atmosphere models, providing mass-loss rates for any planet inside the grid boundaries with superior accuracy compared to previously published interpolation schemes. We consider an already available grid comprising about 11000 hydrodynamic upper atmosphere models for training and generate an additional grid of about 250 models for testing purposes. We develop the ML interpolation scheme (dubbed "atmospheric Mass Loss INquiry frameworK"; MLink) using a Dense Neural Network, further comparing the results with what was obtained employing classical approaches (e.g. linear interpolation and radial basis function-based regression). Finally, we study the impact of the different interpolation schemes on the evolution of a small sample of carefully selected synthetic planets. MLink provides high-quality interpolation across the entire parameter space by significantly reducing both the number of points with large interpolation errors and the maximum interpolation error compared to previously available schemes. For most cases, evolutionary tracks computed employing MLink and classical schemes lead to comparable planetary parameters at Gyr-timescales. However, particularly for planets close to the top edge of the radius gap, the difference between the predicted planetary radii at a given age of tracks obtained employing MLink and classical interpolation schemes can exceed the typical observational uncertainties. Machine learning can be successfully used to estimate atmospheric mass-loss rates from model grids paving the way to explore future larger and more complex grids of models computed accounting for more physical processes., Comment: Accepted for publication on A&A

Vision classifiers are often trained on proprietary datasets containing sensitive information, yet the models themselves are frequently shared openly under the privacy-preserving assumption. Although these models are assumed to protect sensitive information in their training data, the extent to which this assumption holds for different architectures remains unexplored. This assumption is challenged by inversion attacks which attempt to reconstruct training data from model weights, exposing significant privacy vulnerabilities. In this study, we systematically evaluate the privacy-preserving properties of vision classifiers across diverse architectures, including Multi-Layer Perceptrons (MLPs), Convolutional Neural Networks (CNNs), and Vision Transformers (ViTs). Using network inversion-based reconstruction techniques, we assess the extent to which these architectures memorize and reveal training data, quantifying the relative ease of reconstruction across models. Our analysis highlights how architectural differences, such as input representation, feature extraction mechanisms, and weight structures, influence privacy risks. By comparing these architectures, we identify which are more resilient to inversion attacks and examine the trade-offs between model performance and privacy preservation, contributing to the development of secure and privacy-respecting machine learning models for sensitive applications. Our findings provide actionable insights into the design of secure and privacy-aware machine learning systems, emphasizing the importance of evaluating architectural decisions in sensitive applications involving proprietary or personal data.

In complex systems, information propagation can be defined as diffused or delocalized, weakly localized, and strongly localized. This study investigates the application of graph neural network models to learn the behavior of a linear dynamical system on networks. A graph convolution and attention-based neural network framework has been developed to identify the steady-state behavior of the linear dynamical system. We reveal that our trained model distinguishes the different states with high accuracy. Furthermore, we have evaluated model performance with real-world data. In addition, to understand the explainability of our model, we provide an analytical derivation for the forward and backward propagation of our framework., Comment: 13 pages, 7 figures

In network research, Community Detection has always been a topic of significant interest in network science, with numerous papers and algorithms proposing to uncover the underlying structures within networks. In this paper, we conduct a comparative analysis of several prominent community detection algorithms applied to the SNAP Social Circles Dataset, derived from the Facebook Social Media network. The algorithms implemented include Louvain, Girvan-Newman, Spectral Clustering, K-Means Clustering, etc. We evaluate the performance of these algorithms based on various metrics such as modularity, normalized cut-ratio, silhouette score, compactness, and separability. Our findings reveal insights into the effectiveness of each algorithm in detecting various meaningful communities within the social network, shedding light on their strength and limitations. This research contributes to the understanding of community detection methods and provides valuable guidance for their application in analyzing real-world social networks., Comment: Presented at IDEA2k24: https://ideaconference.in/ Submitted to Springer Lecture Notes in Electrical Engineering series (https://www.springer.com/series/7818)

The cybersecurity threat landscape is constantly actively making it imperative to develop sound frameworks to protect the IT structures. Based on this introduction, this paper aims to discuss the application of cybersecurity frameworks into the IT security with focus placed on the role of such frameworks in addressing the changing nature of cybersecurity threats. It explores widely used models, including the NIST Cybersecurity Framework, Zero Trust Architecture, and the ISO/IEC 27001, and how they apply to industries including finance, healthcare and government. The discussion also singles out such technologies as Artificial Intelligence (AI) and Machine Learning (ML) as the core for real-time threat detection and response mechanisms. As these integration challenges demonstrate, the study provides tangible and proven approaches to tackle framework implementation issues such as legitimate security issues, limited availability of funds and resources, and compliance with legal requirements. By capturing current trends and exposures, the findings promote strong, portfolio-based and risk-appropriate security approaches adjusted for organizational goals and capable to prevent advanced cyber threats., Comment: 23 Pages

The sustainability of libraries is critical for modern software development, yet many libraries face abandonment, posing significant risks to dependent projects. This study explores the prevalence and patterns of library abandonment in the Maven ecosystem. We investigate abandonment trends over the past decade, revealing that approximately one in four libraries fail to survive beyond their creation year. We also analyze the release activities of libraries, focusing on their lifespan and release speed, and analyze the evolution of these metrics within the lifespan of libraries. We find that while slow release speed and relatively long periods of inactivity are often precursors to abandonment, some abandoned libraries exhibit bursts of high frequent release activity late in their life cycle. Our findings contribute to a new understanding of library abandonment dynamics and offer insights for practitioners to identify and mitigate risks in software ecosystems.

We successfully stabilized a Fabry-P\'erot optical cavity incorporating a bulk diamond crystal at millikelvin temperatures in a cryogen-free dilution refrigerator with the pulse-tube cryocooler running. In stark contrast to previous demonstrations where lasers were locked to the cavities, our setup locks the cavity to a laser. Our measurements of cavity length fluctuation suggest that the setup could stabilize a cavity up to a finesse of $1.2\times 10^4$ without the diamond and $5.8 \times10^3$ with the diamond crystal. The finesse with a diamond crystal of approximately 90 is primarily limited by the absorption loss inside the diamond.

We consider the stable matching problem (e.g. between doctors and hospitals) in a one-to-one matching setting, where preferences are drawn uniformly at random. It is known that when doctors propose and the number of doctors equals the number of hospitals, then the expected rank of doctors for their match is $\Theta(\log n)$, while the expected rank of the hospitals for their match is $\Theta(n/\log n)$, where $n$ is the size of each side of the market. However, when adding even a single doctor, [Ashlagi, Kanoria and Leshno, 2017] show that the tables have turned: doctors have expected rank of $\Theta(n/\log n)$ while hospitals have expected rank of $\Theta(\log n)$. That is, (slight) competition has a much more dramatically harmful effect than the benefit of being on the proposing side. Motivated by settings where agents inflate their value for an item if it is already allocated to them (termed endowment effect), we study the case where hospitals exhibit ``loyalty". We model loyalty as a parameter $k$, where a hospital currently matched to their $\ell$th most preferred doctor accepts proposals from their $\ell-k-1$th most preferred doctors. Hospital loyalty should help doctors mitigate the harmful effect of competition, as many more outcomes are now stable. However, we show that the effect of competition is so dramatic that, even in settings with extremely high loyalty, in unbalanced markets, the expected rank of doctors already becomes $\tilde{\Theta}(\sqrt{n})$ for loyalty $k=n-\sqrt{n}\log n=n(1-o(1))$.

This paper examines how social media algorithms and filter bubbles contribute to the fragmentation of online discourse, fostering ideological divides and undermining shared understanding. Drawing on Michael Sandels philosophical emphasis on community and shared values, the study explores how digital platforms amplify discrimination discourse including sexism, racism, xenophobia, ableism, homophobia, and religious intolerance during periods of heightened societal tension. By analyzing the dynamics of digital communities, the research highlights mechanisms driving the emergence and evolution of discourse fragments in response to real world events. The findings reveal how social media structures exacerbate polarization, restrict cross group dialogue, and erode the collective reasoning essential for a just society. This study situates philosophical perspectives within a computational analysis of social media interactions, offering a nuanced understanding of the challenges posed by fragmented discourse in the digital age., Comment: 6 pages, 7 figures, 1 table

Dissociated neuronal cultures provide a simplified yet effective model system for investigating self-organized prediction and information processing in neural networks. This review consolidates current research demonstrating that these in vitro networks display fundamental computational capabilities, including predictive coding, adaptive learning, goal-directed behavior, and deviance detection. We examine how these cultures develop critical dynamics optimized for information processing, detail the mechanisms underlying learning and memory formation, and explore the relevance of the free energy principle within these systems. Building on these insights, we discuss how findings from dissociated neuronal cultures inform the design of neuromorphic and reservoir computing architectures, with the potential to enhance energy efficiency and adaptive functionality in artificial intelligence. The reduced complexity of neuronal cultures allows for precise manipulation and systematic investigation, bridging theoretical frameworks with practical implementations in bio-inspired computing. Finally, we highlight promising future directions, emphasizing advancements in three-dimensional culture techniques, multi-compartment models, and brain organoids that deepen our understanding of hierarchical and predictive processes in both biological and artificial systems. This review aims to provide a comprehensive overview of how dissociated neuronal cultures contribute to neuroscience and artificial intelligence, ultimately paving the way for biologically inspired computing solutions., Comment: 39 pages, 4 figures

Rectification, the conversion of AC fields into DC currents, is crucial for optoelectronic applications such as energy harvesting and wireless communication. However, it is conventionally absent in centrosymmetric systems due to vanishing second-order optical responses. Here, we demonstrate significant rectification and photogalvanic currents in centrosymmetric metals via third-order nonlinear optical responses, driven by finite Fermi surface and disorder-induced contributions. We unveil distinct band geometric mechanisms -- including Berry curvature quadrupole, Fermi surface injection, and shift effects -- and classify all symmetry-allowed rectification responses. Using graphene as an example, we illustrate rectification tunability via light polarization and helicity, enabling rectification engineering in centrosymmetric materials for energy-efficient photodetection and terahertz applications., Comment: 8 pages, 3 figures

Cache side channel attacks are a sophisticated and persistent threat that exploit vulnerabilities in modern processors to extract sensitive information. These attacks leverage weaknesses in shared computational resources, particularly the last level cache, to infer patterns in data access and execution flows, often bypassing traditional security defenses. Such attacks are especially dangerous as they can be executed remotely without requiring physical access to the victim's device. This study focuses on a specific class of these threats: fingerprinting attacks, where an adversary monitors and analyzes the behavior of co-located processes via cache side channels. This can potentially reveal confidential information, such as encryption keys or user activity patterns. A comprehensive threat model illustrates how attackers sharing computational resources with target systems exploit these side channels to compromise sensitive data. To mitigate such risks, a hybrid deep learning model is proposed for detecting cache side channel attacks. Its performance is compared with five widely used deep learning models: Multi-Layer Perceptron, Convolutional Neural Network, Simple Recurrent Neural Network, Long Short-Term Memory, and Gated Recurrent Unit. The experimental results demonstrate that the hybrid model achieves a detection rate of up to 99.96%. These findings highlight the limitations of existing models, the need for enhanced defensive mechanisms, and directions for future research to secure sensitive data against evolving side channel threats., Comment: 8 pages, 4 figures. Accepted in IEEE's 2nd International Conference on Computational Intelligence and Network Systems