# Research Papers — 2026-01 · bundle page 93 of 189 > Full-text markdown of papers #4601–4650 of 9403 published in 2026-01, newest first. AI-readable mirrors of the full 9403-paper 2026-01 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-01/page/93 [← Previous bundle](https://aipod.dev/papers/archive/2026-01/page/92.md) [Next bundle →](https://aipod.dev/papers/archive/2026-01/page/94.md) ### What Trace Powers Reveal About Log-Determinants: Closed-Form Estimators, Certificates, and Failure Modes Authors: Piyush Sao Published: 2026-01-18 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2601.12612v1 Computing $\log\det(A)$ for large symmetric positive definite matrices arises in Gaussian process inference and Bayesian model comparison. Standard methods combine matrix-vector products with polynomial approximations. We study a different model: access to trace powers $p_k = \tr(A^k)$, natural when matrix powers are available. Classical moment-based approximations Taylor-expand $\log(λ)$ around the arithmetic mean. This requires $|λ- \AM| < \AM$ and diverges when $κ> 4$. We work instead with the moment-generating function $M(t) = \E(X^t)$ for normalized eigenvalues $X = λ/\AM$. Since $M'(0) = \E(\log X)$, the log-determinant becomes $\log\det(A) = n(\log \AM + M'(0))$ -- the problem reduces to estimating a derivative at $t = 0$. Trace powers give $M(k)$ at positive integers, but interpolating $M(t)$ directly is ill-conditioned due to exponential growth. The transform $K(t) = \log M(t)$ compresses this range. Normalization by $\AM$ ensures $K(0) = K(1) = 0$. With these anchors fixed, we interpolate $K$ through $m+1$ consecutive integers and differentiate to estimate $K'(0)$. However, this local interpolation cannot capture arbitrary spectral features. We prove a fundamental limit: no continuous estimator using finitely many positive moments can be uniformly accurate over unbounded conditioning. Positive moments downweight the spectral tail; $K'(0) = \E(\log X)$ is tail-sensitive. This motivates guaranteed bounds. From the same traces we derive upper bounds on $(\det A)^{1/n}$. Given a spectral floor $r \leq λ_{\min}$, we obtain moment-constrained lower bounds, yielding a provable interval for $\log\det(A)$. A gap diagnostic indicates when to trust the point estimate and when to report bounds. All estimators and bounds cost $O(m)$, independent of $n$. For $m \in \{4, \ldots, 8\}$, this is effectively constant time. --- ### HERMES: A Unified Open-Source Framework for Realtime Multimodal Physiological Sensing, Edge AI, and Intervention in Closed-Loop Smart Healthcare Applications Authors: Maxim Yudayev, Juha Carlon, Diwas Lamsal, Vayalet Stefanova, Benjamin Filtjens Published: 2026-01-18 Categories: eess.SY, cs.LG arXiv: https://arxiv.org/abs/2601.12610v1 Intelligent assistive technologies are increasingly recognized as critical daily-use enablers for people with disabilities and age-related functional decline. Longitudinal studies, curation of quality datasets, live monitoring in activities of daily living, and intelligent intervention devices, share the largely unsolved need in reliable high-throughput multimodal sensing and processing. Streaming large heterogeneous data from distributed sensors, historically closed-source environments, and limited prior works on realtime closed-loop AI methodologies, inhibit such applications. To accelerate the emergence of clinical deployments, we deliver HERMES - an open-source high-performance Python framework for continuous multimodal sensing and AI processing at the edge. It enables synchronized data collection, and realtime streaming inference with user PyTorch models, on commodity computing devices. HERMES is applicable to fixed-lab and free-living environments, of distributed commercial and custom sensors. It is the first work to offer a holistic methodology that bridges cross-disciplinary gaps in real-world implementation strategies, and guides downstream AI model development. Its application on the closed-loop intelligent prosthesis use case illustrates the process of suitable AI model development from the generated constraints and trade-offs. Validation on the use case, with 4 synchronized hosts cooperatively capturing 18 wearable and off-body modalities, demonstrates performance and relevance of HERMES to the trajectory of the intelligent healthcare domain. --- ### A Cloud-based Multi-Agentic Workflow for Science Authors: Anurag Acharya, Timothy Vega, Rizwan A. Ashraf, Anshu Sharma, Derek Parker, Robert Rallo Published: 2026-01-18 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2601.12607v1 As Large Language Models (LLMs) become ubiquitous across various scientific domains, their lack of ability to perform complex tasks like running simulations or to make complex decisions limits their utility. LLM-based agents bridge this gap due to their ability to call external resources and tools and thus are now rapidly gaining popularity. However, coming up with a workflow that can balance the models, cloud providers, and external resources is very challenging, making implementing an agentic system more of a hindrance than a help. In this work, we present a domain-agnostic, model-independent workflow for an agentic framework that can act as a scientific assistant while being run entirely on cloud. Built with a supervisor agent marshaling an array of agents with individual capabilities, our framework brings together straightforward tasks like literature review and data analysis with more complex ones like simulation runs. We describe the framework here in full, including a proof-of-concept system we built to accelerate the study of Catalysts, which is highly important in the field of Chemistry and Material Science. We report the cost to operate and use this framework, including the breakdown of the cost by services use. We also evaluate our system on a custom-curated synthetic benchmark and a popular Chemistry benchmark, and also perform expert validation of the system. The results show that our system is able to route the task to the correct agent 90% of the time and successfully complete the assigned task 97.5% of the time for the synthetic tasks and 91% of the time for real-world tasks, while still achieving better or comparable accuracy to most frontier models, showing that this is a viable framework for other scientific domains to replicate. --- ### Beyond Softmax and Entropy: Convergence Rates of Policy Gradients with f-SoftArgmax Parameterization & Coupled Regularization Authors: Safwan Labbi, Daniil Tiapkin, Paul Mangold, Eric Moulines Published: 2026-01-18 Categories: cs.LG arXiv: https://arxiv.org/abs/2601.12604v2 Policy gradient methods are known to be highly sensitive to the choice of policy parameterization. In particular, the widely used softmax parameterization can induce ill-conditioned optimization landscapes and lead to exponentially slow convergence. Although this can be mitigated by preconditioning, this solution is often computationally expensive. Instead, we propose replacing the softmax with an alternative family of policy parameterizations based on the generalized f-softargmax. We further advocate coupling this parameterization with a regularizer induced by the same f-divergence, which improves the optimization landscape and ensures that the resulting regularized objective satisfies a Polyak-Lojasiewicz inequality. Leveraging this structure, we establish the first explicit non-asymptotic last-iterate convergence guarantees for stochastic policy gradient methods for finite MDPs without any form of preconditioning. We also derive sample-complexity bounds for the unregularized problem and show that f-PG, with Tsallis divergences achieves polynomial sample complexity in contrast to the exponential complexity incurred by the standard softmax parameterization. --- ### onepot CORE -- an enumerated chemical space to streamline drug discovery, enabled by automated small molecule synthesis and AI Authors: Andrei S. Tyrin, Brandon Wang, Manuel Muñoz, Samuel H. Foxman, Daniil A. Boiko Published: 2026-01-18 Categories: physics.chem-ph, cs.LG arXiv: https://arxiv.org/abs/2601.12603v1 The design-make-test-analyze cycle in early-stage drug discovery remains constrained primarily by the "make" step: small-molecule synthesis is slow, costly, and difficult to scale or automate across diverse chemotypes. Enumerated chemical spaces aim to reduce this bottleneck by predefining synthesizable regions of chemical space from available building blocks and reliable reactions, yet existing commercial spaces are still limited by long turnaround times, narrow reaction scope, and substantial manual decision-making in route selection and execution. Here we present the first version of onepot CORE, an enumerated chemical space containing 3.4B molecules and corresponding on-demand synthesis product enabled by an automated synthesis platform and an AI chemist, Phil, that designs, executes, and analyzes experiments. onepot CORE is constructed by (i) selecting a reaction set commonly used in medicinal chemistry, (ii) sourcing and curating building blocks from supplier catalogs, (iii) enumerating candidate products, and (iv) applying ML-based feasibility assessment to prioritize compounds for robust execution. In the current release, the space is supported by seven reactions. We describe an end-to-end workflow - from route selection and automated liquid handling through workup and purification. We further report validation across operational metrics (success rate, timelines, purity, and identity), including NMR confirmation for a representative set of synthesized compounds and assay suitability demonstrated using a series of DPP4 inhibitors. Collectively, onepot CORE illustrates a path toward faster, more reliable access to diverse small molecules, supporting accelerated discovery in pharmaceuticals and beyond. --- ### SSVD-O: Parameter-Efficient Fine-Tuning with Structured SVD for Speech Recognition Authors: Pu Wang, Shinji Watanabe, Hugo Van hamme Published: 2026-01-18 Categories: cs.SD, cs.CL, cs.LG, eess.AS arXiv: https://arxiv.org/abs/2601.12600v1 Parameter-efficient fine-tuning (PEFT) is a scalable approach for adapting large speech foundation models to new domains. While methods such as LoRA and its state-of-the-art variants reduce adaptation costs, they typically allocate parameters uniformly across model subspaces, which limits their efficiency and scalability in speech applications. Building on our prior work, this paper introduces SSVD-Outer (SSVD-O), an extension of the structured SVD-guided (SSVD) fine-tuning method. SSVD-O combines input acoustic feature space-associated inner transformations with output semantic feature space-associated outer transformations to enable scalable and balanced adaptation. We conduct the first systematic analysis of parameter budget allocation across model subspaces in PEFT for automatic speech recognition (ASR), and investigate the trade-off between learning and forgetting under constrained resources. SSVD-O is benchmarked against LoRA, DoRA, PiSSA, and SSVD on domain-shifted ASR tasks, including child speech and regional accents, across model scales from 0.1B to 2B within the ESPnet framework. Experimental results show that SSVD-O consistently narrows the performance gap to full fine-tuning while improving generalization and mitigating catastrophic forgetting. --- ### Dissecting Linear Recurrent Models: How Different Gating Strategies Drive Selectivity and Generalization Authors: Younes Bouhadjar, Maxime Fabre, Felix Schmidt, Emre Neftci Published: 2026-01-18 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2601.12598v1 Linear recurrent neural networks have emerged as efficient alternatives to the original Transformer's softmax attention mechanism, thanks to their highly parallelizable training and constant memory and computation requirements at inference. Iterative refinements of these models have introduced an increasing number of architectural mechanisms, leading to increased complexity and computational costs. Nevertheless, systematic direct comparisons among these models remain limited. Existing benchmark tasks are either too simplistic to reveal substantial differences or excessively resource-intensive for experimentation. In this work, we propose a refined taxonomy of linear recurrent models and introduce SelectivBench, a set of lightweight and customizable synthetic benchmark tasks for systematically evaluating sequence models. SelectivBench specifically evaluates selectivity in sequence models at small to medium scale, such as the capacity to focus on relevant inputs while ignoring context-based distractors. It employs rule-based grammars to generate sequences with adjustable complexity, incorporating irregular gaps that intentionally violate transition rules. Evaluations of linear recurrent models on SelectivBench reveal performance patterns consistent with results from large-scale language tasks. Our analysis clarifies the roles of essential architectural features: gating and rapid forgetting mechanisms facilitate recall, in-state channel mixing is unnecessary for selectivity, but critical for generalization, and softmax attention remains dominant due to its memory capacity scaling with sequence length. Our benchmark enables targeted, efficient exploration of linear recurrent models and provides a controlled setting for studying behaviors observed in large-scale evaluations. Code is available at https://github.com/symseqbench/selectivbench --- ### SLAP: Scalable Language-Audio Pretraining with Variable-Duration Audio and Multi-Objective Training Authors: Xinhao Mei, Gael Le Lan, Haohe Liu, Zhaoheng Ni, Varun Nagaraja, Yang Liu, Yangyang Shi, Vikas Chandra Published: 2026-01-18 Categories: eess.AS, cs.AI, cs.SD arXiv: https://arxiv.org/abs/2601.12594v1 Contrastive language-audio pretraining (CLAP) has achieved notable success in learning semantically rich audio representations and is widely adopted for various audio-related tasks. However, current CLAP models face several key limitations. First, they are typically trained on relatively small datasets, often comprising a few million audio samples. Second, existing CLAP models are restricted to short and fixed duration, which constrains their usage in real-world scenarios with variable-duration audio. Third, the standard contrastive training objective operates on global representations, which may hinder the learning of dense, fine-grained audio features. To address these challenges, we introduce Scalable Language-Audio Pretraining (SLAP), which scales language-audio pretraining to 109 million audio-text pairs with variable audio durations and incorporates multiple training objectives. SLAP unifies contrastive loss with additional self-supervised and captioning losses in a single-stage training, facilitating the learning of richer dense audio representations. The proposed SLAP model achieves new state-of-the-art performance on audio-text retrieval and zero-shot audio classification tasks, demonstrating its effectiveness across diverse benchmarks. --- ### A Theory of Diversity for Random Matrices with Applications to In-Context Learning of Schrödinger Equations Authors: Frank Cole, Yulong Lu, Shaurya Sehgal Published: 2026-01-18 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2601.12587v1 We address the following question: given a collection $\{\mathbf{A}^{(1)}, \dots, \mathbf{A}^{(N)}\}$ of independent $d \times d$ random matrices drawn from a common distribution $\mathbb{P}$, what is the probability that the centralizer of $\{\mathbf{A}^{(1)}, \dots, \mathbf{A}^{(N)}\}$ is trivial? We provide lower bounds on this probability in terms of the sample size $N$ and the dimension $d$ for several families of random matrices which arise from the discretization of linear Schrödinger operators with random potentials. When combined with recent work on machine learning theory, our results provide guarantees on the generalization ability of transformer-based neural networks for in-context learning of Schrödinger equations. --- ### Do MLLMs See What We See? Analyzing Visualization Literacy Barriers in AI Systems Authors: Mengli, Duan, Yuhe, Jiang, Matthew Varona, Carolina Nobre Published: 2026-01-18 Categories: cs.HC, cs.AI, cs.ET arXiv: https://arxiv.org/abs/2601.12585v1 Multimodal Large Language Models (MLLMs) are increasingly used to interpret visualizations, yet little is known about why they fail. We present the first systematic analysis of barriers to visualization literacy in MLLMs. Using the regenerated Visualization Literacy Assessment Test (reVLAT) benchmark with synthetic data, we open-coded 309 erroneous responses from four state-of-the-art models with a barrier-centric strategy adapted from human visualization literacy research. Our analysis yields a taxonomy of MLLM failures, revealing two machine-specific barriers that extend prior human-participation frameworks. Results show that models perform well on simple charts but struggle with color-intensive, segment-based visualizations, often failing to form consistent comparative reasoning. Our findings inform future evaluation and design of reliable AI-driven visualization assistants. --- ### Ontology-aligned structuring and reuse of multimodal materials data and workflows towards automatic reproduction Authors: Sepideh Baghaee Ravari, Abril Azocar Guzman, Sarath Menon, Stefan Sandfeld, Tilmann Hickel, Markus Stricker Published: 2026-01-18 Categories: cond-mat.mtrl-sci, cs.AI arXiv: https://arxiv.org/abs/2601.12582v1 Reproducibility of computational results remains a challenge in materials science, as simulation workflows and parameters are often reported only in unstructured text and tables. While literature data are valuable for validation and reuse, the lack of machine-readable workflow descriptions prevents large-scale curation and systematic comparison. Existing text-mining approaches are insufficient to extract complete computational workflows with their associated parameters. An ontology-driven, large language model (LLM)-assisted framework is introduced for the automated extraction and structuring of computational workflows from the literature. The approach focuses on density functional theory-based stacking fault energy (SFE) calculations in hexagonal close-packed magnesium and its binary alloys, and uses a multi-stage filtering strategy together with prompt-engineered LLM extraction applied to method sections and tables. Extracted information is unified into a canonical schema and aligned with established materials ontologies (CMSO, ASMO, and PLDO), enabling the construction of a knowledge graph using atomRDF. The resulting knowledge graph enables systematic comparison of reported SFE values and supports the structured reuse of computational protocols. While full computational reproducibility is still constrained by missing or implicit metadata, the framework provides a foundation for organizing and contextualizing published results in a semantically interoperable form, thereby improving transparency and reusability of computational materials data. --- ### Primate-like perceptual decision making emerges through deep recurrent reinforcement learning Authors: Nathan J. Wispinski, Scott A. Stone, Anthony Singhal, Patrick M. Pilarski, Craig S. Chapman Published: 2026-01-18 Categories: q-bio.NC, cs.AI arXiv: https://arxiv.org/abs/2601.12577v1 Progress has led to a detailed understanding of the neural mechanisms that underlie decision making in primates. However, less is known about why such mechanisms are present in the first place. Theory suggests that primate decision making mechanisms, and their resultant behavioral abilities, emerged to maximize reward in the face of noisy, temporally evolving information. To test this theory, we trained an end-to-end deep recurrent neural network using reinforcement learning on a noisy perceptual discrimination task. Networks learned several key abilities of primate-like decision making including trading off speed for accuracy, and flexibly changing their mind in the face of new information. Internal dynamics of these networks suggest that these abilities were supported by similar decision mechanisms as those observed in primate neurophysiological studies. These results provide experimental support for key pressures that gave rise to the primate ability to make flexible decisions. --- ### Camera Pose Revisited Authors: Władysław Skarbek, Michał Salomonowicz, Michał Król Published: 2026-01-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12567v1 Estimating the position and orientation of a camera with respect to an observed scene is one of the central problems in computer vision, particularly in the context of camera calibration and multi-sensor systems. This paper addresses the planar Perspective--$n$--Point problem, with special emphasis on the initial estimation of the pose of a calibration object. As a solution, we propose the \texttt{PnP-ProCay78} algorithm, which combines the classical quadratic formulation of the reconstruction error with a Cayley parameterization of rotations and least-squares optimization. The key component of the method is a deterministic selection of starting points based on an analysis of the reconstruction error for two canonical vectors, allowing costly solution-space search procedures to be avoided. Experimental validation is performed using data acquired also from high-resolution RGB cameras and very low-resolution thermal cameras in an integrated RGB--IR setup. The results demonstrate that the proposed algorithm achieves practically the same projection accuracy as optimal \texttt{SQPnP} and slightly higher than \texttt{IPPE}, both prominent \texttt{PnP-OpenCV} procedures. However, \texttt{PnP-ProCay78} maintains a significantly simpler algorithmic structure. Moreover, the analysis of optimization trajectories in Cayley space provides an intuitive insight into the convergence process, making the method attractive also from a didactic perspective. Unlike existing PnP solvers, the proposed \texttt{PnP-ProCay78} algorithm combines projection error minimization with an analytically eliminated reconstruction-error surrogate for translation, yielding a hybrid cost formulation that is both geometrically transparent and computationally efficient. --- ### Agentic Artificial Intelligence (AI): Architectures, Taxonomies, and Evaluation of Large Language Model Agents Authors: Arunkumar V, Gangadharan G. R., Rajkumar Buyya Published: 2026-01-18 Categories: cs.AI, cs.MA arXiv: https://arxiv.org/abs/2601.12560v1 Artificial Intelligence is moving from models that only generate text to Agentic AI, where systems behave as autonomous entities that can perceive, reason, plan, and act. Large Language Models (LLMs) are no longer used only as passive knowledge engines but as cognitive controllers that combine memory, tool use, and feedback from their environment to pursue extended goals. This shift already supports the automation of complex workflows in software engineering, scientific discovery, and web navigation, yet the variety of emerging designs, from simple single loop agents to hierarchical multi agent systems, makes the landscape hard to navigate. In this paper, we investigate architectures and propose a unified taxonomy that breaks agents into Perception, Brain, Planning, Action, Tool Use, and Collaboration. We use this lens to describe the move from linear reasoning procedures to native inference time reasoning models, and the transition from fixed API calls to open standards like the Model Context Protocol (MCP) and Native Computer Use. We also group the environments in which these agents operate, including digital operating systems, embodied robotics, and other specialized domains, and we review current evaluation practices. Finally, we highlight open challenges, such as hallucination in action, infinite loops, and prompt injection, and outline future research directions toward more robust and reliable autonomous systems. --- ### Life, Machine Learning, and the Search for Habitability: Predicting Biosignature Fluxes for the Habitable Worlds Observatory Authors: Mark Moussa, Amber V. Young, Brianna Isola, Vasuda Trehan, Michael D. Himes, Nicholas Wogan, Giada Arney Published: 2026-01-18 Categories: cs.LG, cs.AI, cs.CV arXiv: https://arxiv.org/abs/2601.12557v1 Future direct-imaging flagship missions, such as NASA's Habitable Worlds Observatory (HWO), face critical decisions in prioritizing observations due to extremely stringent time and resource constraints. In this paper, we introduce two advanced machine-learning architectures tailored for predicting biosignature species fluxes from exoplanetary reflected-light spectra: a Bayesian Convolutional Neural Network (BCNN) and our novel model architecture, the Spectral Query Adaptive Transformer (SQuAT). The BCNN robustly quantifies both epistemic and aleatoric uncertainties, offering reliable predictions under diverse observational conditions, whereas SQuAT employs query-driven attention mechanisms to enhance interpretability by explicitly associating spectral features with specific biosignature species. We demonstrate that both models achieve comparably high predictive accuracy on an augmented dataset spanning a wide range of exoplanetary conditions, while highlighting their distinct advantages in uncertainty quantification and spectral interpretability. These capabilities position our methods as promising tools for accelerating target triage, optimizing observation schedules, and maximizing scientific return for upcoming flagship missions such as HWO. --- ### Evaluating Contextually Mediated Factual Recall in Multilingual Large Language Models Authors: Yihong Liu, Bingyu Xiong, Hinrich Schütze Published: 2026-01-18 Categories: cs.CL arXiv: https://arxiv.org/abs/2601.12555v1 Large language models (LLMs) can recall a wide range of factual knowledge across languages. However, existing factual recall evaluations primarily assess fact retrieval in isolation, where the queried entity is explicitly named and the fact is requested directly. In natural language use, facts are often accessed through context, where the relevant entity is introduced only indirectly. In this work, we study contextually mediated factual recall, asking whether LLMs can reliably retrieve factual knowledge when the target entity is embedded in a naturalistic context rather than queried explicitly, across languages. We construct controlled prompts that preserve the underlying fact while introducing referential mediation through contextual sentences. To disentangle contextual effects from name-specific associations, we further compare performance using synthetic names and real names across languages. Evaluating multiple model families in five languages, we find that contextual mediation consistently degrades factual recall, with substantial variation across relations. Larger models are more robust to contextual mediation, exhibiting a reduced performance gap relative to direct queries, while the effect of real names and name origin is mixed and unsystematic. These findings highlight a gap between isolated factual recall and context-dependent language understanding in multilingual LLMs. --- ### Artificial Intelligence in Materials Science and Engineering: Current Landscape, Key Challenges, and Future Trajectorie Authors: Iman Peivaste, Salim Belouettar, Francesco Mercuri, Nicholas Fantuzzi, Hamidreza Dehghani, Razieh Izadi, Halliru Ibrahim, Jakub Lengiewicz, Maël Belouettar-Mathis, Kouider Bendine, Ahmed Makradi, Martin Hörsch et al. Published: 2026-01-18 Categories: cond-mat.mtrl-sci, cs.AI, physics.comp-ph, quant-ph arXiv: https://arxiv.org/abs/2601.12554v1 Artificial Intelligence is rapidly transforming materials science and engineering, offering powerful tools to navigate complexity, accelerate discovery, and optimize material design in ways previously unattainable. Driven by the accelerating pace of algorithmic advancements and increasing data availability, AI is becoming an essential competency for materials researchers. This review provides a comprehensive and structured overview of the current landscape, synthesizing recent advancements and methodologies for materials scientists seeking to effectively leverage these data-driven techniques. We survey the spectrum of machine learning approaches, from traditional algorithms to advanced deep learning architectures, including CNNs, GNNs, and Transformers, alongside emerging generative AI and probabilistic models such as Gaussian Processes for uncertainty quantification. The review also examines the pivotal role of data in this field, emphasizing how effective representation and featurization strategies, spanning compositional, structural, image-based, and language-inspired approaches, combined with appropriate preprocessing, fundamentally underpin the performance of machine learning models in materials research. Persistent challenges related to data quality, quantity, and standardization, which critically impact model development and application in materials science and engineering, are also addressed. --- ### PISE: Physics-Anchored Semantically-Enhanced Deep Computational Ghost Imaging for Robust Low-Bandwidth Machine Perception Authors: Tong Wu Published: 2026-01-18 Categories: cs.CV, eess.IV arXiv: https://arxiv.org/abs/2601.12551v2 We propose PISE, a physics-informed deep ghost imaging framework for low-bandwidth edge perception. By combining adjoint operator initialization with semantic guidance, PISE improves classification accuracy by 2.57% and reduces variance by 9x at 5% sampling. --- ### Benchmarking Concept-Spilling Across Languages in LLMs Authors: Ilia Badanin, Daniil Dzenhaliou, Imanol Schlag Published: 2026-01-18 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2601.12549v1 Multilingual Large Language Models (LLMs) exhibit remarkable cross-lingual abilities, yet often exhibit a systematic bias toward the representations from other languages, resulting in semantic interference when generating content in non-English languages$-$a phenomenon we define as language spilling. This paper presents a novel comparative framework for evaluating multilingual semantic robustness by systematically measuring how models handle polysemous words across languages. Our methodology provides a relative measure of model performance: when required to generate exactly five meanings, both strong and weak models may resort to meanings from dominant languages, but semantically stronger models do so later in the generation sequence, producing more true meanings from the target language before failing, while weaker models resort to dominant-language meanings earlier in the sequence. We evaluate a diverse set of open and closed multilingual LLMs using a structured meaning generation task across nine languages, employing a carefully curated benchmark of 100 high-polysemy English words. Our findings reveal significant variation in semantic robustness across both models and languages, providing a principled ranking system for model comparison without requiring definitive causal attribution of error sources. We contribute both a scalable comparative benchmark for multilingual semantic evaluation and a rigorous validation pipeline$-$critical tools for developing more linguistically balanced AI systems. --- ### How Clinicians Think and What AI Can Learn From It Authors: Dipayan Sengupta, Saumya Panda Published: 2026-01-18 Categories: cs.AI arXiv: https://arxiv.org/abs/2601.12547v2 Clinical artificial intelligence increasingly builds high-dimensional representations of patients, yet every finite clinical model is an abstraction. The key question is not only how accurately a model predicts, but which distinctions need to be represented, at what resolution, for the decision at hand. We argue that purposive clinical AI should use decision-sufficient abstraction: model detail should be conditioned by the objective structure of the decision and limited by the evidence available to support that detail. We further argue that this objective structure should often be ordinal-first. The primitive objects are objective-threshold propositions that may recur at progressively finer levels. For example, mortality below 10 percent may appear early and mortality below 5 percent later. We formalize ordered threshold refinement, a set-valued decision relation, and threshold-sufficient abstraction. A model needs only enough resolution to preserve the survivor set induced by the thresholds reached so far. Once an earlier threshold resolves the choice, further refinement of later objectives has no decision value for that choice. When a potentially decisive threshold is unresolved, additional information or model complexity should be targeted to that threshold rather than added globally. The framework links clinical reasoning, model abstraction, causal inference, lexicographic optimization, constrained learning, and human-AI collaboration. Its central claim is that clinical AI should pursue purpose before fidelity, order before abstraction, and refinement until decision-sufficient rather than reality-complete. --- ### Press Start to Charge: Videogaming the Online Centralized Charging Scheduling Problem Authors: Alireza Ghahtarani, Martin Cousineau, Amir-massoud Farahmand, Jorge E. Mendoza Published: 2026-01-18 Categories: cs.LG arXiv: https://arxiv.org/abs/2601.12543v1 We study the online centralized charging scheduling problem (OCCSP). In this problem, a central authority must decide, in real time, when to charge dynamically arriving electric vehicles (EVs), subject to capacity limits, with the objective of balancing load across a finite planning horizon. To solve the problem, we first gamify it; that is, we model it as a game where charging blocks are placed within temporal and capacity constraints on a grid. We design heuristic policies, train learning agents with expert demonstrations, and improve them using Dataset Aggregation (DAgger). From a theoretical standpoint, we show that gamification reduces model complexity and yields tighter generalization bounds than vector-based formulations. Experiments across multiple EV arrival patterns confirm that gamified learning enhances load balancing. In particular, the image-to-movement model trained with DAgger consistently outperforms heuristic baselines, vector-based approaches, and supervised learning agents, while also demonstrating robustness in sensitivity analyses. These operational gains translate into tangible economic value. In a real-world case study for the Greater Montréal Area (Québec, Canada) using utility cost data, the proposed methods lower system costs by tens of millions of dollars per year over the prevailing practice and show clear potential to delay costly grid upgrades. --- ### Rethinking the AI Scientist: Interactive Multi-Agent Workflows for Scientific Discovery Authors: Lukas Weidener, Marko Brkić, Mihailo Jovanović, Ritvik Singh, Chiara Baccin, Emre Ulgac, Alex Dobrin, Aakaash Meduri Published: 2026-01-18 Categories: cs.AI arXiv: https://arxiv.org/abs/2601.12542v2 Artificial intelligence systems for scientific discovery have demonstrated remarkable potential, yet existing approaches remain largely proprietary and operate in batch-processing modes requiring hours per research cycle, precluding real-time researcher guidance. This paper introduces Deep Research, a multi-agent system enabling interactive scientific investigation with turnaround times measured in minutes. The architecture comprises specialized agents for planning, data analysis, literature search, and novelty detection, unified through a persistent world state that maintains context across iterative research cycles. Two operational modes support different workflows: semi-autonomous mode with selective human checkpoints, and fully autonomous mode for extended investigations. Evaluation on the BixBench computational biology benchmark demonstrated state-of-the-art performance, achieving 48.8% accuracy on open response and 64.4% on multiple-choice evaluation, exceeding existing baselines by 14 to 26 percentage points. Analysis of architectural constraints, including open access literature limitations and challenges inherent to automated novelty assessment, informs practical deployment considerations for AI-assisted scientific workflows. --- ### MemeLens: Multilingual Multitask VLMs for Memes Authors: Ali Ezzat Shahroor, Mohamed Bayan Kmainasi, Abul Hasnat, Dimitar Dimitrov, Giovanni Da San Martino, Preslav Nakov, Firoj Alam Published: 2026-01-18 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2601.12539v4 Memes are a dominant medium for online communication and manipulation because meaning emerges from interactions between embedded text, imagery, and cultural context. Existing meme research is distributed across tasks (e.g., \textit{hate, misogyny, propaganda, sentiment, humour}) and languages, which limits cross-domain generalization. To address this gap, we propose \textsc{MemeLens}, a unified multilingual, multitask explanation-enhanced Vision-Language Model (VLM) for meme understanding. We consolidate $38$ public meme datasets, filter and map dataset-specific labels into a shared taxonomy of $20$ tasks spanning harm, targets, figurative/pragmatic intent, and affect. We present a comprehensive empirical analysis across modeling paradigms, task categories, and datasets. Our findings suggest that robust meme understanding requires multimodal training, varies substantially across semantic categories, and remains sensitive to over-specialization when models are fine-tuned on individual datasets rather than trained in a unified setting. We make the experimental resources (https://github.com/MohamedBayan/MemeLens), model (https://huggingface.co/QCRI/MemeLens-VLM) and datasets (https://huggingface.co/datasets/QCRI/MemeLens) publicly available to the community. --- ### A Survey of Agentic Reasoning for Large Language Models: Towards Recursively Self-Improving and Collective Agents Authors: Tianxin Wei, Ting-Wei Li, Zhining Liu, Xuying Ning, Ze Yang, Jiaru Zou, Zhichen Zeng, Ruizhong Qiu, Xiao Lin, Dongqi Fu, Zihao Li, Mengting Ai et al. Published: 2026-01-18 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2601.12538v2 Reasoning is a fundamental cognitive process underlying inference, problem-solving, and decision-making. While large language models (LLMs) demonstrate strong reasoning capabilities in closed-world settings, they struggle in open-ended and dynamic environments. Agentic reasoning marks a paradigm shift by reframing LLMs as autonomous agents that plan, act, and learn through continual interaction. In this survey, we organize agentic reasoning along three complementary dimensions. First, we characterize environmental dynamics through three layers: foundational agentic reasoning, which establishes core single-agent capabilities including planning, tool use, and search in stable environments; self-evolving agentic reasoning, which studies how agents refine these capabilities through feedback, memory, and adaptation; and collective multi-agent reasoning, which extends intelligence to collaborative settings involving coordination, knowledge sharing, and shared goals. Across these layers, we distinguish in-context reasoning, which scales test-time interaction through structured orchestration, from post-training reasoning, which optimizes behaviors via reinforcement learning and supervised fine-tuning. We further review representative agentic reasoning frameworks across real-world applications and benchmarks, including science, robotics, healthcare, autonomous research, and mathematics. This survey synthesizes agentic reasoning methods into a unified roadmap bridging thought and action, and outlines open challenges and future directions, including personalization, long-horizon interaction, world modeling, scalable multi-agent training, and governance for real-world deployment. --- ### Improving Low-Resource Machine Translation via Round-Trip Reinforcement Learning Authors: Ahmed Attia, Alham Fikri Aji Published: 2026-01-18 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2601.12535v3 Low-resource machine translation (MT) has gained increasing attention as parallel data from low-resource language communities is collected, but many approaches for improving low-resource MT remain underexplored. We investigate a self-supervised reinforcement learning fine-tuning for translation in low-resource settings using round-trip bootstrapping with the No Language Left Behind (NLLB) family of models. Our approach translates English into a target low-resource language and then back into English, using a combination of chrF++ and BLEU as the reward function on the reconstructed English sentences. Using the NLLB-MD dataset, we evaluate both the 600M and 1.3B parameter NLLB models and observe consistent improvements for the following languages: Central Aymara, Friulian, Wolof, Dyula, Bhojpuri and Russian. Qualitative inspection of translation outputs indicates increased fluency and semantic fidelity. We argue that our method can further benefit from scale, enabling models to increasingly leverage their pretrained knowledge and continue self-improving. Code available at: https://github.com/Copticoder/MT-via-Round-Trip-RL --- ### Encoding Emotion Through Self-Supervised Eye Movement Reconstruction Authors: Marcus Ma, Jordan Prescott, Emily Zhou, Tiantian Feng, Kleanthis Avramidis, Gabor Mihaly Toth, Shrikanth Narayanan Published: 2026-01-18 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2601.12534v2 The relationship between emotional expression and eye movement is well-documented, with literature establishing gaze patterns are reliable indicators of emotion. However, most studies utilize specialized, high-resolution eye-tracking equipment, limiting the potential reach of findings. We investigate how eye movement can be used to predict multimodal markers of emotional expression from naturalistic, low-resolution videos. We utilize a collection of video interviews from the USC Shoah Foundation's Visual History Archive with Holocaust survivors as they recount their experiences in the Auschwitz concentration camp. Inspired by pretraining methods on language models, we develop a novel gaze detection model that uses self-supervised eye movement reconstruction that can effectively leverage unlabeled video. We use this model's encoder embeddings to fine-tune models on two downstream tasks related to emotional expression. The first is aligning eye movement with directional emotion estimates from speech. The second task is using eye gaze as a predictor of three momentary manifestations of emotional behaviors: laughing, crying/sobbing, and sighing. We find our new model is predictive of emotion outcomes and observe a positive correlation between pretraining performance and emotion processing performance for both experiments. We conclude self-supervised eye movement reconstruction is an effective method for encoding the affective signal they carry. --- ### BirdsEye-RU: A Dataset For Detecting Faces from Overhead Images Authors: Md. Ahanaf Arif Khan, Ariful Islam, Sangeeta Biswas, Md. Iqbal Aziz Khan, Subrata Pramanik, Sanjoy Kumar Chakravarty, Bimal Kumar Pramanik Published: 2026-01-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12533v2 Detecting faces in overhead images remains a significant challenge due to extreme scale variations and environmental clutter. To address this, we created the BirdsEye-RU dataset, a comprehensive collection of 2,978 images containing over eight thousand annotated faces. This dataset is specifically designed to capture small and distant faces across diverse environments, containing both drone images and smartphone-captured images from high altitude. We present a detailed description of the BirdsEye-RU dataset in this paper. We made our dataset freely available to the public, and it can be accessed at https://www.kaggle.com/datasets/mdahanafarifkhan/birdseye-ru. --- ### XRefine: Attention-Guided Keypoint Match Refinement Authors: Jan Fabian Schmid, Annika Hagemann Published: 2026-01-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12530v1 Sparse keypoint matching is crucial for 3D vision tasks, yet current keypoint detectors often produce spatially inaccurate matches. Existing refinement methods mitigate this issue through alignment of matched keypoint locations, but they are typically detector-specific, requiring retraining for each keypoint detector. We introduce XRefine, a novel, detector-agnostic approach for sub-pixel keypoint refinement that operates solely on image patches centered at matched keypoints. Our cross-attention-based architecture learns to predict refined keypoint coordinates without relying on internal detector representations, enabling generalization across detectors. Furthermore, XRefine can be extended to handle multi-view feature tracks. Experiments on MegaDepth, KITTI, and ScanNet demonstrate that the approach consistently improves geometric estimation accuracy, achieving superior performance compared to existing refinement methods while maintaining runtime efficiency. Our code and trained models can be found at https://github.com/boschresearch/xrefine. --- ### Deep Feature Deformation Weights Authors: Richard Liu, Itai Lang, Rana Hanocka Published: 2026-01-18 Categories: cs.CV, cs.GR arXiv: https://arxiv.org/abs/2601.12527v2 Handle-based mesh deformation is a classic paradigm in computer graphics which enables intuitive edits from sparse controls. Classical techniques are fast and precise, but require users to know ideal handle placement apriori, which can be unintuitive and inconsistent. Handle sets cannot be adjusted easily, as weights are typically optimized through energies defined by the handles. Modern data-driven methods, on the other hand, provide semantic edits but sacrifice fine-grained control and speed. We propose a technique that achieves the best of both worlds: deep feature proximity yields smooth, visual-aware deformation weights with no additional regularization. Importantly, these weights are computed in real-time for any surface point, unlike prior methods which require expensive optimization. We introduce barycentric feature distillation, an improved feature distillation pipeline which leverages the full visual signal from shape renders to make distillation complexity robust to mesh resolution. This enables high resolution meshes to be processed in minutes versus potentially hours for prior methods. We preserve and extend classical properties through feature space constraints and locality weighting. Our field representation enables automatic visual symmetry detection, which we use to produce symmetry-preserving deformations. We show a proof-of-concept application which can produce deformations for meshes up to 1 million faces in real-time on a consumer-grade machine. Project page at https://threedle.github.io/dfd. --- ### Deep Lightweight Unrolled Network for High Dynamic Range Modulo Imaging Authors: Brayan Monroy, Jorge Bacca Published: 2026-01-18 Categories: eess.IV, cs.CV arXiv: https://arxiv.org/abs/2601.12526v1 Modulo-Imaging (MI) offers a promising alternative for expanding the dynamic range of images by resetting the signal intensity when it reaches the saturation level. Subsequently, high-dynamic range (HDR) modulo imaging requires a recovery process to obtain the HDR image. MI is a non-convex and ill-posed problem where recent recovery networks suffer in high-noise scenarios. In this work, we formulate the HDR reconstruction task as an optimization problem that incorporates a deep prior and subsequently unrolls it into an optimization-inspired deep neural network. The network employs a lightweight convolutional denoiser for fast inference with minimal computational overhead, effectively recovering intensity values while mitigating noise. Moreover, we introduce the Scaling Equivariance term that facilitates self-supervised fine-tuning, thereby enabling the model to adapt to new modulo images that fall outside the original training distribution. Extensive evaluations demonstrate the superiority of our method compared to state-of-the-art recovery algorithms in terms of performance and quality. --- ### Approximating splits for decision trees quickly in sparse data streams Authors: Nikolaj Tatti Published: 2026-01-18 Categories: cs.LG, cs.DS arXiv: https://arxiv.org/abs/2601.12525v1 Decision trees are one of the most popular classifiers in the machine learning literature. While the most common decision tree learning algorithms treat data as a batch, numerous algorithms have been proposed to construct decision trees from a data stream. A standard training strategy involves augmenting the current tree by changing a leaf node into a split. Here we typically maintain counters in each leaf which allow us to determine the optimal split, and whether the split should be done. In this paper we focus on how to speed up the search for the optimal split when dealing with sparse binary features and a binary class. We focus on finding splits that have the approximately optimal information gain or Gini index. In both cases finding the optimal split can be done in $O(d)$ time, where $d$ is the number of features. We propose an algorithm that yields $(1 + α)$ approximation when using conditional entropy in amortized $O(α^{-1}(1 + m\log d) \log \log n)$ time, where $m$ is the number of 1s in a data point, and $n$ is the number of data points. Similarly, for Gini index, we achieve $(1 + α)$ approximation in amortized $O(α^{-1} + m \log d)$ time. Our approach is beneficial for sparse data where $m \ll d$. In our experiments we find almost-optimal splits efficiently, faster than the baseline, overperforming the theoretical approximation guarantees. --- ### Improved Bug Localization with AI Agents Leveraging Hypothesis and Dynamic Cognition Authors: Asif Mohammed Samir, Mohammad Masudur Rahman Published: 2026-01-18 Categories: cs.SE, cs.AI, cs.IR, cs.LG, cs.MA arXiv: https://arxiv.org/abs/2601.12522v2 Software bugs cost technology providers (e.g., AT&T) billions annually and cause developers to spend roughly 50% of their time on bug resolution. Traditional methods for bug localization often analyze the suspiciousness of code components (e.g., methods, documents) in isolation, overlooking their connections with other components in the codebase. Recent advances in Large Language Models (LLMs) and agentic AI techniques have shown strong potential for code understanding, but still lack causal reasoning during code exploration and struggle to manage growing context effectively, limiting their capability. In this paper, we present a novel agentic technique for bug localization -- CogniGent -- that overcomes the limitations above by leveraging multiple AI agents capable of causal reasoning, call-graph-based root cause analysis and context engineering. It emulates developers-inspired debugging practices (a.k.a., dynamic cognitive debugging) and conducts hypothesis testing to support bug localization. We evaluate CogniGent on a curated dataset of 591 bug reports using three widely adopted performance metrics and compare it against six established baselines from the literature. Experimental results show that our technique consistently outperformed existing traditional and LLM-based techniques, achieving MAP improvements of 23.33-38.57% at the document and method levels. Similar gains were observed in MRR, with increases of 25.14-53.74% at both granularity levels. Statistical significance tests also confirm the superiority of our technique. By addressing the reasoning, dependency, and context limitations, CogniGent advances the state of bug localization, bridging human-like cognition with agentic automation for improved performance. --- ### Learning Relativistic Geodesics and Chaotic Dynamics via Stabilized Lagrangian Neural Networks Authors: Abdullah Umut Hamzaogullari, Arkadas Ozakin Published: 2026-01-18 Categories: cs.LG arXiv: https://arxiv.org/abs/2601.12519v1 Lagrangian Neural Networks (LNNs) can learn arbitrary Lagrangians from trajectory data, but their unusual optimization objective leads to significant training instabilities that limit their application to complex systems. We propose several improvements that address these fundamental challenges, namely, a Hessian regularization scheme that penalizes unphysical signatures in the Lagrangian's second derivatives with respect to velocities, preventing the network from learning unstable dynamics, activation functions that are better suited to the problem of learning Lagrangians, and a physics-aware coordinate scaling that improves stability. We systematically evaluate these techniques alongside previously proposed methods for improving stability. Our improved architecture successfully trains on systems of unprecedented complexity, including triple pendulums, and achieved 96.6\% lower validation loss value and 90.68\% better stability than baseline LNNs in double pendulum systems. With the improved framework, we show that our LNNs can learn Lagrangians representing geodesic motion in both non-relativistic and general relativistic settings. To deal with the relativistic setting, we extended our regularization to penalize violations of Lorentzian signatures, which allowed us to predict a geodesic Lagrangian under AdS\textsubscript{4} spacetime metric directly from trajectory data, which to our knowledge has not been done in the literature before. This opens new possibilities for automated discovery of geometric structures in physics, including extraction of spacetime metric tensor components from geodesic trajectories. While our approach inherits some limitations of the original LNN framework, particularly the requirement for invertible Hessians, it significantly expands the practical applicability of LNNs for scientific discovery tasks. --- ### Cooperative Multi-agent RL with Communication Constraints Authors: Nuoya Xiong, Aarti Singh Published: 2026-01-18 Categories: cs.LG, cs.AI, stat.ML arXiv: https://arxiv.org/abs/2601.12518v1 Cooperative MARL often assumes frequent access to global information in a data buffer, such as team rewards or other agents' actions, which is typically unrealistic in decentralized MARL systems due to high communication costs. When communication is limited, agents must rely on outdated information to estimate gradients and update their policies. A common approach to handle missing data is called importance sampling, in which we reweigh old data from a base policy to estimate gradients for the current policy. However, it quickly becomes unstable when the communication is limited (i.e. missing data probability is high), so that the base policy in importance sampling is outdated. To address this issue, we propose a technique called base policy prediction, which utilizes old gradients to predict the policy update and collect samples for a sequence of base policies, which reduces the gap between the base policy and the current policy. This approach enables effective learning with significantly fewer communication rounds, since the samples of predicted base policies could be collected within one communication round. Theoretically, we show that our algorithm converges to an $\varepsilon$-Nash equilibrium in potential games with only $O(\varepsilon^{-3/4})$ communication rounds and $O(poly(\max_i |A_i|)\varepsilon^{-11/4})$ samples, improving existing state-of-the-art results in communication cost, as well as sample complexity without the exponential dependence on the joint action space size. We also extend these results to general Markov Cooperative Games to find an agent-wise local maximum. Empirically, we test the base policy prediction algorithm in both simulated games and MAPPO for complex environments. --- ### Fine-Tuning Cycle-GAN for Domain Adaptation of MRI Images Authors: Mohd Usama, Belal Ahmad, Faleh Menawer R Althiyabi Published: 2026-01-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12512v1 Magnetic Resonance Imaging (MRI) scans acquired from different scanners or institutions often suffer from domain shifts owing to variations in hardware, protocols, and acquisition parameters. This discrepancy degrades the performance of deep learning models trained on source domain data when applied to target domain images. In this study, we propose a Cycle-GAN-based model for unsupervised medical-image domain adaptation. Leveraging CycleGANs, our model learns bidirectional mappings between the source and target domains without paired training data, preserving the anatomical content of the images. By leveraging Cycle-GAN capabilities with content and disparity loss for adaptation tasks, we ensured image-domain adaptation while maintaining image integrity. Several experiments on MRI datasets demonstrated the efficacy of our model in bidirectional domain adaptation without labelled data. Furthermore, research offers promising avenues for improving the diagnostic accuracy of healthcare. The statistical results confirm that our approach improves model performance and reduces domain-related variability, thus contributing to more precise and consistent medical image analysis. --- ### CoLR-Det: Collaborative Latent Restoration for Small Object Detection in Low-Resolution Remote Sensing Images Authors: Ruo Qi, Linhui Dai, Yusong Qin, Chaolei Yang, Yanshan Li Published: 2026-01-18 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2601.12507v2 Low-resolution remote sensing small object detection is limited by both missing visual details and the ambiguity of how details serve detection. Existing super-resolution-assisted detectors generally follow a restoration-first paradigm to explicitly enhance inputs before detection, which implicitly assumes visual fidelity benefits recognition. Yet super-resolution favors dense texture and edge recovery, while object detection relies on sparse instance-level semantics, making restoration amplify visually plausible but semantically irrelevant background textures. To tackle this issue, we propose CoLR-Det, a Collaborative Latent-Restoration-Assisted Small Object Detection framework that treats super-resolution supervision as detection-oriented latent regularization rather than explicit image-level enhancement. Instead of reconstructing high-resolution images for inference, CoLR-Det uses a training-only restoration branch to impose auxiliary reconstruction constraints on shared multiscale representations, and the inference pathway remains purely detection-driven. We further design a saliency-guided object-preserving token routing mechanism, which prioritizes high-saliency tokens for attention-based refinement while retaining information of bypassed tokens. Besides, a detection-prioritized two-stage optimization strategy is developed: it first builds stable object-level semantics before introducing restoration supervision, and assigns a smaller learning rate to the SR decoder to keep its updates conservative and reduce perturbations in collaborative training. With this design, CoLR-Det transforms restoration from an explicit visual enhancement operator into an implicit semantic regularizer. Experiments on resolution-degraded NWPU VHR-10-Split, DOTAv1.5-Split and HRSSD-Split show that CoLR-Det outperforms state-of-the-art methods, with code available at https://github.com/qiruo-ya/CoLR-Det. --- ### DoPE: Decoy Oriented Perturbation Encapsulation Human-Readable, AI-Hostile Documents for Academic Integrity Authors: Ashish Raj Shekhar, Shiven Agarwal, Priyanuj Bordoloi, Yash Shah, Tejas Anvekar, Vivek Gupta Published: 2026-01-18 Categories: cs.CL arXiv: https://arxiv.org/abs/2601.12505v1 Multimodal Large Language Models (MLLMs) can directly consume exam documents, threatening conventional assessments and academic integrity. We present DoPE (Decoy-Oriented Perturbation Encapsulation), a document-layer defense framework that embeds semantic decoys into PDF/HTML assessments to exploit render-parse discrepancies in MLLM pipelines. By instrumenting exams at authoring time, DoPE provides model-agnostic prevention (stop or confound automated solving) and detection (flag blind AI reliance) without relying on conventional one-shot classifiers. We formalize prevention and detection tasks, and introduce FewSoRT-Q, an LLM-guided pipeline that generates question-level semantic decoys and FewSoRT-D to encapsulate them into watermarked documents. We evaluate on Integrity-Bench, a novel benchmark of 1826 exams (PDF+HTML) derived from public QA datasets and OpenCourseWare. Against black-box MLLMs from OpenAI and Anthropic, DoPE yields strong empirical gains: a 91.4% detection rate at an 8.7% false-positive rate using an LLM-as-Judge verifier, and prevents successful completion or induces decoy-aligned failures in 96.3% of attempts. We release Integrity-Bench, our toolkit, and evaluation code to enable reproducible study of document-layer defenses for academic integrity. --- ### Semidefinite Programming for Quantum Channel Learning Authors: Mikhail Gennadievich Belov, Victor Victorovich Dubov, Vadim Konstantinovich Ivanov, Alexander Yurievich Maslov, Olga Vladimirovna Proshina, Vladislav Gennadievich Malyshkin Published: 2026-01-18 Categories: cs.LG, math.NA, quant-ph arXiv: https://arxiv.org/abs/2601.12502v2 The problem of reconstructing a quantum channel from a sample of classical data is considered. When the total fidelity can be represented as a ratio of two quadratic forms (e.g., in the case of mapping a mixed state to a pure state, projective operators, unitary learning, and others), Semidefinite Programming (SDP) can be applied to solve the fidelity optimization problem with respect to the Choi matrix. A remarkable feature of SDP is that the optimization is convex, which allows the problem to be efficiently solved by a variety of numerical algorithms. We have tested several commercially available SDP solvers, all of which allowed for the reconstruction of quantum channels of different forms. A notable feature is that the Kraus rank of the obtained quantum channel typically comprises less than a few percent of its maximal possible value. This suggests that a relatively small Kraus rank quantum channel is typically sufficient to describe experimentally observed classical data. The theory was also applied to the problem of reconstructing projective operators from data. Finally, we discuss a classical computational model based on quantum channel transformation, performed and calculated on a classical computer, possibly hardware-optimized. --- ### Video Individual Counting and Tracking from Moving Drones: A Benchmark and Methods Authors: Yaowu Fan, Jia Wan, Tao Han, Andy J. Ma, Wanli Ouyang, Antoni B. Chan Published: 2026-01-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12500v3 Counting and tracking dense crowds in large-scale scenes is valuable yet challenging, while existing methods and datasets are largely limited to fixed cameras with small scene coverage. We introduce MovingDroneCrowd++, a large-scale video-level dataset dedicated to dense crowd counting and tracking from moving drones, captured under diverse flight altitudes, camera angles, and illumination conditions. Existing methods, however, still fail to achieve satisfactory Video Individual Counting (VIC) or Multi-Object Tracking (MOT) performance under these challenging aerial conditions. To this end, we propose GD3A (Global Density Map Decomposition via Group-wise Density Assignment) for VIC and GIA-Track (Group-wise Identity Association Tracker) for MOT. Both methods are unified under a framework that aggregates cross-frame pixel-level matches among multiple descriptors per pedestrian through group-wise density assignment (GDA) and group-wise identity association (GIA), respectively. This group-wise aggregation tolerates intra-group spatial mismatches and limits the propagation of inter-group errors. To establish reliable pixel-level descriptor correspondences across frames, we design a frame-pair-conditioned dustbin score inferred from the pedestrian descriptors of each frame pair, enabling Optimal Transport to better distinguish identity differences. Based on these matches, GD3A decomposes global density maps into shared, inflow, and outflow components, while GIA-Track establishes robust pedestrian trajectories. Experiments show that our methods achieve substantial gains in both VIC and MOT on moving-drone videos with dense crowds and complex motions, reducing counting error by 47.4% and improving tracking accuracy by 64.6%. Code, dataset, and pretrained models are available at https://github.com/fyw1999/MovingDroneCrowd. --- ### Failure Modes in Multi-Hop QA: The Weakest Link Effect and the Recognition Bottleneck Authors: Meiru Zhang, Zaiqiao Meng, Nigel Collier Published: 2026-01-18 Categories: cs.AI, cs.LG arXiv: https://arxiv.org/abs/2601.12499v2 Despite scaling to massive context windows, Large Language Models (LLMs) struggle with multi-hop reasoning due to inherent position bias, which causes them to overlook information at certain positions. Whether these failures stem from an inability to locate evidence (recognition failure) or integrate it (synthesis failure) is unclear. We introduce Multi-Focus Attention Instruction (MFAI), a semantic probe to disentangle these mechanisms by explicitly steering attention towards selected positions. Across 5 LLMs on two multi-hop QA tasks (MuSiQue and NeoQA), we identify the "Weakest Link Effect": in our 18-document, 3-bucket setting, multi-hop reasoning performance collapses to the level of the least visible evidence, governed by absolute position rather than the linear distance between facts. While matched MFAI resolves recognition bottlenecks, improving accuracy by up to 11.49% in low-visibility positions, misleading MFAI yields divergent effects modulated by task topology: entity-centric tasks with vertical reasoning chains are vulnerable, whereas event-centric tasks with horizontal evidence structures are more resilient. Finally, we demonstrate that thinking models utilizing System-2 reasoning effectively locate and integrate the required information, matching gold-only baselines even in noisy, long-context settings. Supplementary experiments on 2WikiMultiHopQA, extended 3-4 hop counts, and a 32B model confirm these findings generalize across datasets, reasoning depths, and model scales. --- ### Multi-Task Instruction Tuning via Data Scheduling for Low-Resource Arabic SpeechLLMs Authors: Hunzalah Hassan Bhatti, Firoj Alam, Shammur Absar Chowdhury Published: 2026-01-18 Categories: cs.SD, cs.AI, cs.CL, eess.AS arXiv: https://arxiv.org/abs/2601.12494v3 Audio large language models (LLMs) enable unified speech understanding and generation, but adapting them to linguistically complex and dialect-rich settings such as Arabic-English remains challenging. We present a controlled study of multi-task instruction tuning for an Arabic-centric audio LLM across generative tasks, including automatic speech recognition (ASR) and speech and text summarization, as well as discriminative tasks, including dialect identification (DID) and speech emotion recognition (SER), in a resource-constrained setting. To support end-to-end Arabic speech summarization, we introduce AraMega-SSum, the first Arabic speech summarization dataset designed for training and benchmarking Arabic-centric audio LLMs. We compare four training strategies: (i) Uniform Mixing (UM), (ii) Task-Progressive Curriculum (TPC), (iii) Aligner-Based Diverse Sampling (ADS) for training-time batch construction, and (iv) a two-stage TPC->ADS strategy. Our results reveal a clear efficiency-robustness trade-off. TPC achieves the strongest performance on generative tasks, including ASR and summarization. ADS improves paralinguistic tasks but reduces generative stability when used alone. The two-stage TPC->ADS strategy provides the best overall balance, achieving the strongest DID and SER performance while outperforming large proprietary models such as Gemini-2.5-Pro on discriminative tasks. We will publicly release AraMega-SSum together with all experimental resources to support future research in Arabic speech understanding. --- ### Histopath-C: Towards Realistic Domain Shifts for Histopathology Vision-Language Adaptation Authors: Mehrdad Noori, Gustavo Adolfo Vargas Hakim, David Osowiechi, Fereshteh Shakeri, Ali Bahri, Moslem Yazdanpanah, Sahar Dastani, Ismail Ben Ayed, Christian Desrosiers Published: 2026-01-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12493v1 Medical Vision-language models (VLMs) have shown remarkable performances in various medical imaging domains such as histo\-pathology by leveraging pre-trained, contrastive models that exploit visual and textual information. However, histopathology images may exhibit severe domain shifts, such as staining, contamination, blurring, and noise, which may severely degrade the VLM's downstream performance. In this work, we introduce Histopath-C, a new benchmark with realistic synthetic corruptions designed to mimic real-world distribution shifts observed in digital histopathology. Our framework dynamically applies corruptions to any available dataset and evaluates Test-Time Adaptation (TTA) mechanisms on the fly. We then propose LATTE, a transductive, low-rank adaptation strategy that exploits multiple text templates, mitigating the sensitivity of histopathology VLMs to diverse text inputs. Our approach outperforms state-of-the-art TTA methods originally designed for natural images across a breadth of histopathology datasets, demonstrating the effectiveness of our proposed design for robust adaptation in histopathology images. Code and data are available at https://github.com/Mehrdad-Noori/Histopath-C. --- ### A Mixture of Experts Vision Transformer for High-Fidelity Surface Code Decoding Authors: Hoang Viet Nguyen, Manh Hung Nguyen, Hoang Ta, Van Khu Vu, Yeow Meng Chee Published: 2026-01-18 Categories: quant-ph, cs.IT, cs.LG arXiv: https://arxiv.org/abs/2601.12483v2 Quantum error correction is a key ingredient for large scale quantum computation, protecting logical information from physical noise by encoding it into many physical qubits. Topological stabilizer codes are particularly appealing due to their geometric locality and practical relevance. In these codes, stabilizer measurements yield a syndrome that must be decoded into a recovery operation, making decoding a central bottleneck for scalable real time operation. Existing decoders are commonly classified into two categories. Classical algorithmic decoders provide strong and well established baselines, but may incur substantial computational overhead at large code distances or under stringent latency constraints. Machine learning based decoders offer fast GPU inference and flexible function approximation, yet many approaches do not explicitly exploit the lattice geometry and local structure of topological codes, which can limit performance. In this work, we propose QuantumSMoE, a quantum vision transformer based decoder that incorporates code structure through plus shaped embeddings and adaptive masking to capture local interactions and lattice connectivity, and improves scalability via a mixture of experts layer with a novel auxiliary loss. Experiments on the toric code demonstrate that QuantumSMoE outperforms state-of-the-art machine learning decoders as well as widely used classical baselines. --- ### NeuralFur: Animal Fur Reconstruction From Multi-View Images Authors: Vanessa Sklyarova, Berna Kabadayi, Anastasios Yiannakidis, Giorgio Becherini, Michael J. Black, Justus Thies Published: 2026-01-18 Categories: cs.CV, cs.GR arXiv: https://arxiv.org/abs/2601.12481v1 Reconstructing realistic animal fur geometry from images is a challenging task due to the fine-scale details, self-occlusion, and view-dependent appearance of fur. In contrast to human hairstyle reconstruction, there are also no datasets that can be leveraged to learn a fur prior for different animals. In this work, we present a first multi-view-based method for high-fidelity 3D fur modeling of animals using a strand-based representation, leveraging the general knowledge of a vision language model. Given multi-view RGB images, we first reconstruct a coarse surface geometry using traditional multi-view stereo techniques. We then use a vision language model (VLM) system to retrieve information about the realistic length structure of the fur for each part of the body. We use this knowledge to construct the animal's furless geometry and grow strands atop it. The fur reconstruction is supervised with both geometric and photometric losses computed from multi-view images. To mitigate orientation ambiguities stemming from the Gabor filters that are applied to the input images, we additionally utilize the VLM to guide the strands' growth direction and their relation to the gravity vector that we incorporate as a loss. With this new schema of using a VLM to guide 3D reconstruction from multi-view inputs, we show generalization across a variety of animals with different fur types. For additional results and code, please refer to https://neuralfur.is.tue.mpg.de. --- ### Capability-Aware Early-Stage Research Idea Evaluation Authors: Renlong Jie, Chen Chu, Zhen Wang Published: 2026-01-18 Categories: cs.CL arXiv: https://arxiv.org/abs/2601.12473v1 Predicting the outcomes of research ideas at their conceptual stage (i.e. before significant resources are committed) holds great potential for optimizing scientific resource allocation and research planning. While existing methods rely heavily on finished manuscripts or peer reviews, we propose a novel capability-aware framework that predicts paper acceptance and ratings using only author information and research ideas, without requiring full text or experimental results. Our approach integrates author information, (inferred) capability presentation, and research ideas through a three-way transformer architecture with flexible fusion mechanisms. We also introduce a two-stage architecture for learning the capability representation given the author information and idea. Experiments show that our method significantly outperform the single-way models by finetuning bert-base and bert-large, and the capability predicting significantly increase the predictive accuracy of the final model. The proposed method can be applied in both early-stage research outcome prediction and scientific resource allocation. --- ### Knowing When to Abstain: Medical LLMs Under Clinical Uncertainty Authors: Sravanthi Machcha, Sushrita Yerra, Sahil Gupta, Aishwarya Sahoo, Sharmin Sultana, Hong Yu, Zonghai Yao Published: 2026-01-18 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2601.12471v2 Current evaluation of large language models (LLMs) overwhelmingly prioritizes accuracy; however, in real-world and safety-critical applications, the ability to abstain when uncertain is equally vital for trustworthy deployment. We introduce MedAbstain, a unified benchmark and evaluation protocol for abstention in medical multiple-choice question answering (MCQA) -- a discrete-choice setting that generalizes to agentic action selection -- integrating conformal prediction, adversarial question perturbations, and explicit abstention options. Our systematic evaluation of both open- and closed-source LLMs reveals that even state-of-the-art, high-accuracy models often fail to abstain with uncertain. Notably, providing explicit abstention options consistently increases model uncertainty and safer abstention, far more than input perturbations, while scaling model size or advanced prompting brings little improvement. These findings highlight the central role of abstention mechanisms for trustworthy LLM deployment and offer practical guidance for improving safety in high-stakes applications. --- ### DCAC: Dynamic Class-Aware Cache Creates Stronger Out-of-Distribution Detectors Authors: Yanqi Wu, Qichao Chen, Runhe Lai, Xinhua Lu, Jia-Xin Zhuang, Zhilin Zhao, Wei-Shi Zheng, Ruixuan Wang Published: 2026-01-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12468v1 Out-of-distribution (OOD) detection remains a fundamental challenge for deep neural networks, particularly due to overconfident predictions on unseen OOD samples during testing. We reveal a key insight: OOD samples predicted as the same class, or given high probabilities for it, are visually more similar to each other than to the true in-distribution (ID) samples. Motivated by this class-specific observation, we propose DCAC (Dynamic Class-Aware Cache), a training-free, test-time calibration module that maintains separate caches for each ID class to collect high-entropy samples and calibrate the raw predictions of input samples. DCAC leverages cached visual features and predicted probabilities through a lightweight two-layer module to mitigate overconfident predictions on OOD samples. This module can be seamlessly integrated with various existing OOD detection methods across both unimodal and vision-language models while introducing minimal computational overhead. Extensive experiments on multiple OOD benchmarks demonstrate that DCAC significantly enhances existing methods, achieving substantial improvements, i.e., reducing FPR95 by 6.55% when integrated with ASH-S on ImageNet OOD benchmark. --- ### Patch-Level Tokenization with CNN Encoders and Attention for Improved Transformer Time-Series Forecasting Authors: Saurish Nagrath, Saroj Kumar Panigrahy Published: 2026-01-18 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2601.12467v3 Transformer-based models have shown strong performance in time-series forecasting by leveraging self-attention to model long-range temporal dependencies. However, their effectiveness depends critically on the quality and structure of input representations derived from raw multivariate time-series data, particularly as sequence length and data scale increase. This paper proposes a two-stage forecasting framework that explicitly separates local temporal representation learning from global dependency modelling. In the proposed approach, a convolutional neural network operates on fixed-length temporal patches to extract short-range temporal dynamics and non-linear feature interactions, producing compact patch-level token embeddings. Token-level self-attention is applied during representation learning to refine these embeddings, after which a Transformer encoder models inter-patch temporal dependencies to generate forecasts. The method is evaluated on a synthetic multivariate time-series dataset with controlled static and dynamic factors, using an extended sequence length and a larger number of samples. Experimental results demonstrate that the proposed framework consistently outperforms a convolutional baseline under increased temporal context and remains competitive with a strong patch-based Transformer model. These findings indicate that structured patch-level tokenization provides a scalable and effective representation for multivariate time-series forecasting, particularly when longer input sequences are considered. --- ### Incentivizing In-depth Reasoning over Long Contexts with Process Advantage Shaping Authors: Miao Peng, Weizhou Shen, Nuo Chen, Chenliang Li, Ming Yan, Jia Li Published: 2026-01-18 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2601.12465v1 Reinforcement Learning with Verifiable Rewards (RLVR) has proven effective in enhancing LLMs short-context reasoning, but its performance degrades in long-context scenarios that require both precise grounding and robust long-range reasoning. We identify the "almost-there" phenomenon in long-context reasoning, where trajectories are largely correct but fail at the final step, and attribute this failure to two factors: (1) the lack of high reasoning density in long-context QA data that push LLMs beyond mere grounding toward sophisticated multi-hop reasoning; and (2) the loss of valuable learning signals during long-context RL training due to the indiscriminate penalization of partially correct trajectories with incorrect outcomes. To overcome this bottleneck, we propose DeepReasonQA, a KG-driven synthesis framework that controllably constructs high-difficulty, multi-hop long-context QA pairs with inherent reasoning chains. Building on this, we introduce Long-context Process Advantage Shaping (LongPAS), a simple yet effective method that performs fine-grained credit assignment by evaluating reasoning steps along Validity and Relevance dimensions, which captures critical learning signals from "almost-there" trajectories. Experiments on three long-context reasoning benchmarks show that our approach substantially outperforms RLVR baselines and matches frontier LLMs while using far fewer parameters. Further analysis confirms the effectiveness of our methods in strengthening long-context reasoning while maintaining stable RL training. --- ### Large-scale EM Benchmark for Multi-Organelle Instance Segmentation in the Wild Authors: Yanrui Lu, Danyang Chen, Haowen Xiao, Jiarui Zhu, Fukang Ge, Binqian Zou, Jiali Guan, Jiayin Liang, Yuting Wang, Ziqian Guan, Xiangcheng Bao, Jinhao Bi et al. Published: 2026-01-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12464v1 Accurate instance-level segmentation of organelles in electron microscopy (EM) is critical for quantitative analysis of subcellular morphology and inter-organelle interactions. However, current benchmarks, based on small, curated datasets, fail to capture the inherent heterogeneity and large spatial context of in-the-wild EM data, imposing fundamental limitations on current patch-based methods. To address these limitations, we developed a large-scale, multi-source benchmark for multi-organelle instance segmentation, comprising over 100,000 2D EM images across variety cell types and five organelle classes that capture real-world variability. Dataset annotations were generated by our designed connectivity-aware Label Propagation Algorithm (3D LPA) with expert refinement. We further benchmarked several state-of-the-art models, including U-Net, SAM variants, and Mask2Former. Our results show several limitations: current models struggle to generalize across heterogeneous EM data and perform poorly on organelles with global, distributed morphologies (e.g., Endoplasmic Reticulum). These findings underscore the fundamental mismatch between local-context models and the challenge of modeling long-range structural continuity in the presence of real-world variability. The benchmark dataset and labeling tool will be publicly released soon. ---