# Research Papers — 2026-01 · bundle page 92 of 189 > Full-text markdown of papers #4551–4600 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/92 [← Previous bundle](https://aipod.dev/papers/archive/2026-01/page/91.md) [Next bundle →](https://aipod.dev/papers/archive/2026-01/page/93.md) ### GaussianTrimmer: Online Trimming Boundaries for 3DGS Segmentation Authors: Liwei Liao, Ronggang Wang Published: 2026-01-19 Categories: cs.CV, eess.IV arXiv: https://arxiv.org/abs/2601.12683v1 With the widespread application of 3D Gaussians in 3D scene representation, 3D scene segmentation methods based on 3D Gaussians have also gradually emerged. However, existing 3D Gaussian segmentation methods basically segment on the basis of Gaussian primitives. Due to the large variation range of the scale of 3D Gaussians, large-sized Gaussians that often span the foreground and background lead to jagged boundaries of segmented objects. To this end, we propose an online boundary trimming method, GaussianTrimmer, which is an efficient and plug-and-play post-processing method capable of trimming coarse boundaries for existing 3D Gaussian segmentation methods. Our method consists of two core steps: 1. Generating uniformly and well-covered virtual cameras; 2. Trimming Gaussian at the primitive level based on 2D segmentation results on virtual cameras. Extensive quantitative and qualitative experiments demonstrate that our method can improve the segmentation quality of existing 3D Gaussian segmentation methods as a plug-and-play method. --- ### Fusion-Restoration Image Processing Algorithm to Improve the High-Temperature Deformation Measurement Authors: Banglei Guan, Dongcai Tan, Jing Tao, Ang Su, Yang Shang, Qifeng Yu Published: 2026-01-19 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12682v1 In the deformation measurement of high-temperature structures, image degradation caused by thermal radiation and random errors introduced by heat haze restrict the accuracy and effectiveness of deformation measurement. To suppress thermal radiation and heat haze using fusion-restoration image processing methods, thereby improving the accuracy and effectiveness of DIC in the measurement of high-temperature deformation. For image degradation caused by thermal radiation, based on the image layered representation, the image is decomposed into positive and negative channels for parallel processing, and then optimized for quality by multi-exposure image fusion. To counteract the high-frequency, random errors introduced by heat haze, we adopt the FSIM as the objective function to guide the iterative optimization of model parameters, and the grayscale average algorithm is applied to equalize anomalous gray values, thereby reducing measurement error. The proposed multi-exposure image fusion algorithm effectively suppresses image degradation caused by complex illumination conditions, boosting the effective computation area from 26% to 50% for under-exposed images and from 32% to 40% for over-exposed images without degrading measurement accuracy in the experiment. Meanwhile, the image restoration combined with the grayscale average algorithm reduces static thermal deformation measurement errors. The error in ε_xx is reduced by 85.3%, while the errors in ε_yy and γ_xy are reduced by 36.0% and 36.4%, respectively. We present image processing methods to suppress the interference of thermal radiation and heat haze in high-temperature deformation measurement using DIC. The experimental results verify that the proposed method can effectively improve image quality, reduce deformation measurement errors, and has potential application value in thermal deformation measurement. --- ### MetaToolAgent: Towards Generalizable Tool Usage in LLMs through Meta-Learning Authors: Zheng Fang, Wolfgang Mayer, Zeyu Zhang, Jian Wang, Hong-Yu Zhang, Wanli Li, Zaiwen Feng Published: 2026-01-19 Categories: cs.LG arXiv: https://arxiv.org/abs/2601.12680v1 Tool learning is increasingly important for large language models (LLMs) to effectively coordinate and utilize a diverse set of tools in order to solve complex real-world tasks. By selecting and integrating appropriate tools, LLMs extend their capabilities beyond pure language understanding to perform specialized functions. However, existing methods for tool selection often focus on limited tool sets and struggle to generalize to novel tools encountered in practical deployments. To address these challenges, we introduce a comprehensive dataset spanning 7 domains, containing 155 tools and 9,377 question-answer pairs, which simulates realistic integration scenarios. Additionally, we propose MetaToolAgent (MTA), a meta-learning approach designed to improve cross-tool generalization. Experimental results show that MTA significantly outperforms baseline methods on unseen tools, demonstrating its promise for building flexible and scalable systems that require dynamic tool coordination. --- ### VILTA: A VLM-in-the-Loop Adversary for Enhancing Driving Policy Robustness Authors: Qimao Chen, Fang Li, Shaoqing Xu, Zhiyi Lai, Zixun Xie, Yuechen Luo, Shengyin Jiang, Hanbing Li, Long Chen, Bing Wang, Yi Zhang, Zhi-Xin Yang Published: 2026-01-19 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12672v1 The safe deployment of autonomous driving (AD) systems is fundamentally hindered by the long-tail problem, where rare yet critical driving scenarios are severely underrepresented in real-world data. Existing solutions including safety-critical scenario generation and closed-loop learning often rely on rule-based heuristics, resampling methods and generative models learned from offline datasets, limiting their ability to produce diverse and novel challenges. While recent works leverage Vision Language Models (VLMs) to produce scene descriptions that guide a separate, downstream model in generating hazardous trajectories for agents, such two-stage framework constrains the generative potential of VLMs, as the diversity of the final trajectories is ultimately limited by the generalization ceiling of the downstream algorithm. To overcome these limitations, we introduce VILTA (VLM-In-the-Loop Trajectory Adversary), a novel framework that integrates a VLM into the closed-loop training of AD agents. Unlike prior works, VILTA actively participates in the training loop by comprehending the dynamic driving environment and strategically generating challenging scenarios through direct, fine-grained editing of surrounding agents' future trajectories. This direct-editing approach fully leverages the VLM's powerful generalization capabilities to create a diverse curriculum of plausible yet challenging scenarios that extend beyond the scope of traditional methods. We demonstrate that our approach substantially enhances the safety and robustness of the resulting AD policy, particularly in its ability to navigate critical long-tail events. --- ### Exploiting Test-Time Augmentation in Federated Learning for Brain Tumor MRI Classification Authors: Thamara Leandra de Deus Melo, Rodrigo Moreira, Larissa Ferreira Rodrigues Moreira, André Ricardo Backes Published: 2026-01-19 Categories: cs.CV, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2601.12671v1 Efficient brain tumor diagnosis is crucial for early treatment; however, it is challenging because of lesion variability and image complexity. We evaluated convolutional neural networks (CNNs) in a federated learning (FL) setting, comparing models trained on original versus preprocessed MRI images (resizing, grayscale conversion, normalization, filtering, and histogram equalization). Preprocessing alone yielded negligible gains; combined with test-time augmentation (TTA), it delivered consistent, statistically significant improvements in federated MRI classification (p<0.001). In practice, TTA should be the default inference strategy in FL-based medical imaging; when the computational budget permits, pairing TTA with light preprocessing provides additional reliable gains. --- ### Empowering All-in-Loop Health Management of Spacecraft Power System in the Mega-Constellation Era via Human-AI Collaboration Authors: Yi Di, Zhibin Zhao, Fujin Wang, Xue Liu, Jiafeng Tang, Jiaxin Ren, Zhi Zhai, Xuefeng Chen Published: 2026-01-19 Categories: cs.AI arXiv: https://arxiv.org/abs/2601.12667v2 It is foreseeable that the number of spacecraft will increase exponentially, ushering in an era dominated by satellite mega-constellations (SMC). This necessitates a focus on energy in space: spacecraft power systems (SPS), especially their health management (HM), given their role in power supply and high failure rates. Providing health management for dozens of SPS and for thousands of SPS represents two fundamentally different paradigms. Therefore, to adapt the health management in the SMC era, this work proposes a principle of aligning underlying capabilities (AUC principle) and develops SpaceHMchat, an open-source Human-AI collaboration (HAIC) framework for all-in-loop health management (AIL HM). SpaceHMchat serves across the entire loop of work condition recognition, anomaly detection, fault localization, and maintenance decision making, achieving goals such as conversational task completion, adaptive human-in-the-loop learning, personnel structure optimization, knowledge sharing, efficiency enhancement, as well as transparent reasoning and improved interpretability. Meanwhile, to validate this exploration, a hardware-realistic fault injection experimental platform is established, and its simulation model is built and open-sourced, both fully replicating the real SPS. The corresponding experimental results demonstrate that SpaceHMchat achieves excellent performance across 23 quantitative metrics, such as 100% conclusion accuracy in logical reasoning of work condition recognition, over 99% success rate in anomaly detection tool invocation, over 90% precision in fault localization, and knowledge base search time under 3 minutes in maintenance decision-making. Another contribution of this work is the release of the first-ever AIL HM dataset of SPS. This dataset contains four sub-datasets, involving 4 types of AIL HM sub-tasks, 17 types of faults, and over 700,000 timestamps. --- ### Near-Light Color Photometric Stereo for Mono-Chromatic Non-Lambertian Surfaces Authors: Zonglin Li, Jieji Ren, Shuangfan Zhou, Heng Guo, Jinnuo Zhang, Jiang Zhou, Boxin Shi, Zhanyu Ma, Guoying Gu Published: 2026-01-19 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12666v2 Color photometric stereo enables single-shot surface reconstruction, extending conventional photometric stereo that requires multiple images of a static scene under varying illumination to dynamic scenarios. However, most existing approaches assume ideal distant lighting and Lambertian reflectance, leaving more practical near-light conditions and non-Lambertian surfaces underexplored. To overcome this limitation, we propose a framework that leverages neural implicit representations for depth and BRDF modeling under the assumption of mono-chromaticity (uniform chromaticity and homogeneous material), which alleviates the inherent ill-posedness of color photometric stereo and allows for detailed surface recovery from just one image. Furthermore, we design a compact optical tactile sensor to validate our approach. Experiments on both synthetic and real-world datasets demonstrate that our method achieves accurate and robust surface reconstruction. --- ### Generalizable Hyperparameter Optimization for Federated Learning on Non-IID Cancer Images Authors: Elisa Gonçalves Ribeiro, Rodrigo Moreira, Larissa Ferreira Rodrigues Moreira, André Ricardo Backes Published: 2026-01-19 Categories: cs.CV, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2601.12664v1 Deep learning for cancer histopathology training conflicts with privacy constraints in clinical settings. Federated Learning (FL) mitigates this by keeping data local; however, its performance depends on hyperparameter choices under non-independent and identically distributed (non-IID) client datasets. This paper examined whether hyperparameters optimized on one cancer imaging dataset generalized across non-IID federated scenarios. We considered binary histopathology tasks for ovarian and colorectal cancers. We perform centralized Bayesian hyperparameter optimization and transfer dataset-specific optima to the non-IID FL setup. The main contribution of this study is the introduction of a simple cross-dataset aggregation heuristic by combining configurations by averaging the learning rates and considering the modal optimizers and batch sizes. This combined configuration achieves a competitive classification performance. --- ### Energy-Efficient Prediction in Textile Manufacturing: Enhancing Accuracy and Data Efficiency With Ensemble Deep Transfer Learning Authors: Yan-Chen Chen, Wei-Yu Chiu, Qun-Yu Wang, Jing-Wei Chen, Hao-Ting Zhao Published: 2026-01-19 Categories: eess.SP, cs.LG arXiv: https://arxiv.org/abs/2601.12663v1 Traditional textile factories consume substantial energy, making energy-efficient production optimization crucial for sustainability and cost reduction. Meanwhile, deep neural networks (DNNs), which are effective for factory output prediction and operational optimization, require extensive historical data, posing challenges due to high sensor deployment and data collection costs. To address this, we propose Ensemble Deep Transfer Learning (EDTL), a novel framework that enhances prediction accuracy and data efficiency by integrating transfer learning with an ensemble strategy and a feature alignment layer. EDTL pretrains DNN models on data-rich production lines (source domain) and adapts them to data-limited lines (target domain), reducing dependency on large datasets. Experiments on real-world textile factory datasets show that EDTL improves prediction accuracy by 5.66% and enhances model robustness by 3.96% compared to conventional DNNs, particularly in data-limited scenarios (20%-40% data availability). This research contributes to energy-efficient textile manufacturing by enabling accurate predictions with fewer data requirements, providing a scalable and cost-effective solution for smart production systems. --- ### Decentralized Learning Strategies for Estimation Error Minimization with Graph Neural Networks Authors: Xingran Chen, Navid NaderiAlizadeh, Alejandro Ribeiro, Shirin Saeedi Bidokhti Published: 2026-01-19 Categories: cs.LG, eess.SP arXiv: https://arxiv.org/abs/2601.12662v2 We address real-time sampling and estimation of autoregressive Markovian sources in dynamic yet structurally similar multi-hop wireless networks. Each node caches samples from others and communicates over wireless collision channels, aiming to minimize time-average estimation error via decentralized policies. Due to the high dimensionality of action spaces and complexity of network topologies, deriving optimal policies analytically is intractable. To address this, we propose a graphical multi-agent reinforcement learning framework for policy optimization. Theoretically, we demonstrate that our proposed policies are transferable, allowing a policy trained on one graph to be effectively applied to structurally similar graphs. Numerical experiments demonstrate that (i) our proposed policy outperforms state-of-the-art baselines; (ii) the trained policies are transferable to larger networks, with performance gains increasing with the number of agents; (iii) the graphical training procedure withstands non-stationarity, even when using independent learning techniques; and (iv) recurrence is pivotal in both independent learning and centralized training and decentralized execution, and improves the resilience to non-stationarity. --- ### MedConsultBench: A Full-Cycle, Fine-Grained, Process-Aware Benchmark for Medical Consultation Agents Authors: Chuhan Qiao, Jianghua Huang, Daxing Zhao, Ziding Liu, Yanjun Shen, Bing Cheng, Wei Lin, Kai Wu Published: 2026-01-19 Categories: cs.AI arXiv: https://arxiv.org/abs/2601.12661v1 Current evaluations of medical consultation agents often prioritize outcome-oriented tasks, frequently overlooking the end-to-end process integrity and clinical safety essential for real-world practice. While recent interactive benchmarks have introduced dynamic scenarios, they often remain fragmented and coarse-grained, failing to capture the structured inquiry logic and diagnostic rigor required in professional consultations. To bridge this gap, we propose MedConsultBench, a comprehensive framework designed to evaluate the complete online consultation cycle by covering the entire clinical workflow from history taking and diagnosis to treatment planning and follow-up Q\&A. Our methodology introduces Atomic Information Units (AIUs) to track clinical information acquisition at a sub-turn level, enabling precise monitoring of how key facts are elicited through 22 fine-grained metrics. By addressing the underspecification and ambiguity inherent in online consultations, the benchmark evaluates uncertainty-aware yet concise inquiry while emphasizing medication regimen compatibility and the ability to handle realistic post-prescription follow-up Q\&A via constraint-respecting plan revisions. Systematic evaluation of 19 large language models reveals that high diagnostic accuracy often masks significant deficiencies in information-gathering efficiency and medication safety. These results underscore a critical gap between theoretical medical knowledge and clinical practice ability, establishing MedConsultBench as a rigorous foundation for aligning medical AI with the nuanced requirements of real-world clinical care. --- ### Toward Faithful Explanations in Acoustic Anomaly Detection Authors: Maab Elrashid, Anthony Deschênes, Cem Subakan, Mirco Ravanelli, Rémi Georges, Michael Morin Published: 2026-01-19 Categories: cs.SD, cs.LG, eess.AS arXiv: https://arxiv.org/abs/2601.12660v1 Interpretability is essential for user trust in real-world anomaly detection applications. However, deep learning models, despite their strong performance, often lack transparency. In this work, we study the interpretability of autoencoder-based models for audio anomaly detection, by comparing a standard autoencoder (AE) with a mask autoencoder (MAE) in terms of detection performance and interpretability. We applied several attribution methods, including error maps, saliency maps, SmoothGrad, Integrated Gradients, GradSHAP, and Grad-CAM. Although MAE shows a slightly lower detection, it consistently provides more faithful and temporally precise explanations, suggesting a better alignment with true anomalies. To assess the relevance of the regions highlighted by the explanation method, we propose a perturbation-based faithfulness metric that replaces them with their reconstructions to simulate normal input. Our findings, based on experiments in a real industrial scenario, highlight the importance of incorporating interpretability into anomaly detection pipelines and show that masked training improves explanation quality without compromising performance. --- ### Augmenting Question Answering with A Hybrid RAG Approach Authors: Tianyi Yang, Nashrah Haque, Vaishnave Jonnalagadda, Yuya Jeremy Ong, Zhehui Chen, Yanzhao Wu, Lei Yu, Divyesh Jadav, Wenqi Wei Published: 2026-01-19 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2601.12658v2 Retrieval-Augmented Generation (RAG) has emerged as a powerful technique for enhancing the quality of responses in Question-Answering (QA) tasks. However, existing approaches often struggle with retrieving contextually relevant information, leading to incomplete or suboptimal answers. In this paper, we introduce Structured-Semantic RAG (SSRAG), a hybrid architecture that enhances QA quality by integrating query augmentation, agentic routing, and a structured retrieval mechanism combining vector and graph based techniques with context unification. By refining retrieval processes and improving contextual grounding, our approach improves both answer accuracy and informativeness. We conduct extensive evaluations on three popular QA datasets, TruthfulQA, SQuAD and WikiQA, across five Large Language Models (LLMs), demonstrating that our proposed approach consistently improves response quality over standard RAG implementations. --- ### Explanation Multiplicity in SHAP: Characterization and Assessment Authors: Hyunseung Hwang, Seungeun Lee, Lucas Rosenblatt, Steven Euijong Whang, Julia Stoyanovich Published: 2026-01-19 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2601.12654v2 Post-hoc explanations are widely used to justify, contest, and review automated decisions in high-stakes domains such as lending, employment, and healthcare. Among these methods, SHAP is often treated as providing a reliable account of which features mattered for an individual prediction and is routinely used to support recourse, oversight, and accountability. In practice, however, SHAP explanations can differ substantially across repeated runs, even when the individual, prediction task, and trained model are held fixed. We conceptualize and name this phenomenon explanation multiplicity: the existence of multiple, internally valid but substantively different explanations for the same decision. Explanation multiplicity poses a normative challenge for responsible AI deployment, as it undermines expectations that explanations can reliably identify the reasons for an adverse outcome. We present a comprehensive methodology for characterizing explanation multiplicity in post-hoc feature attribution methods, disentangling sources arising from model training and selection versus stochasticity intrinsic to the explanation pipeline. Furthermore, whether explanation multiplicity is surfaced depends on how explanation consistency is measured. Commonly used magnitude-based metrics can suggest stability while masking substantial instability in the identity and ordering of top-ranked features. To contextualize observed instability, we derive and estimate randomized baseline values under plausible null models, providing a principled reference point for interpreting explanation disagreement. Across datasets, model classes, and confidence regimes, we find that explanation multiplicity is widespread and persists even under highly controlled conditions, including high-confidence predictions. Thus explanation practices must be evaluated using metrics and baselines aligned with their intended societal role. --- ### Intelligent Documentation in Medical Education: Can AI Replace Manual Case Logging? Authors: Nafiz Imtiaz Khan, Kylie Cleland, Vladimir Filkov, Roger Eric Goldman Published: 2026-01-19 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2601.12648v1 Procedural case logs are a core requirement in radiology training, yet they are time-consuming to complete and prone to inconsistency when authored manually. This study investigates whether large language models (LLMs) can automate procedural case log documentation directly from free-text radiology reports. We evaluate multiple local and commercial LLMs under instruction-based and chain-of-thought prompting to extract structured procedural information from 414 curated interventional radiology reports authored by nine residents between 2018 and 2024. Model performance is assessed using sensitivity, specificity, and F1-score, alongside inference latency and token efficiency to estimate operational cost. Results show that both local and commercial models achieve strong extraction performance, with best F1-scores approaching 0.87, while exhibiting different trade-offs between speed and cost. Automation using LLMs has the potential to substantially reduce clerical burden for trainees and improve consistency in case logging. These findings demonstrate the feasibility of AI-assisted documentation in medical education and highlight the need for further validation across institutions and clinical workflows. --- ### Unbounded Harms, Bounded Law: Liability in the Age of Borderless AI Authors: Ha-Chi Tran Published: 2026-01-19 Categories: cs.CY, cs.AI arXiv: https://arxiv.org/abs/2601.12646v1 The rapid proliferation of artificial intelligence (AI) has exposed significant deficiencies in risk governance. While ex-ante harm identification and prevention have advanced, Responsible AI scholarship remains underdeveloped in addressing ex-post liability. Core legal questions regarding liability allocation, responsibility attribution, and remedial effectiveness remain insufficiently theorized and institutionalized, particularly for transboundary harms and risks that transcend national jurisdictions. Drawing on contemporary AI risk analyses, we argue that such harms are structurally embedded in global AI supply chains and are likely to escalate in frequency and severity due to cross-border deployment, data infrastructures, and uneven national oversight capacities. Consequently, territorially bounded liability regimes are increasingly inadequate. Using a comparative and interdisciplinary approach, this paper examines compensation and liability frameworks from high-risk transnational domains - including vaccine injury schemes, systemic financial risk governance, commercial nuclear liability, and international environmental regimes - to distill transferable legal design principles such as strict liability, risk pooling, collective risk-sharing, and liability channelling, while highlighting potential structural constraints on their application to AI-related harms. Situated within an international order shaped more by AI arms race dynamics than cooperative governance, the paper outlines the contours of a global AI accountability and compensation architecture, emphasizing the tension between geopolitical rivalry and the collective action required to govern transboundary AI risks effectively. --- ### STEP-LLM: Generating CAD STEP Models from Natural Language with Large Language Models Authors: Xiangyu Shi, Junyang Ding, Xu Zhao, Sinong Zhan, Payal Mohapatra, Daniel Quispe, Kojo Welbeck, Jian Cao, Wei Chen, Ping Guo, Qi Zhu Published: 2026-01-19 Categories: cs.AI arXiv: https://arxiv.org/abs/2601.12641v2 Computer-aided design (CAD) is vital to modern manufacturing, yet model creation remains labor-intensive and expertise-heavy. To enable non-experts to translate intuitive design intent into manufacturable artifacts, recent large language models-based text-to-CAD efforts focus on command sequences or script-based formats like CadQuery. However, these formats are kernel-dependent and lack universality for manufacturing. In contrast, the Standard for the Exchange of Product Data (STEP, ISO 10303) file is a widely adopted, neutral boundary representation (B-rep) format directly compatible with manufacturing, but its graph-structured, cross-referenced nature poses unique challenges for auto-regressive LLMs. To address this, we curate a dataset of ~40K STEP-caption pairs and introduce novel preprocessing tailored for the graph-structured format of STEP, including a depth-first search-based reserialization that linearizes cross-references while preserving locality and chain-of-thought(CoT)-style structural annotations that guide global coherence. We integrate retrieval-augmented generation to ground predictions in relevant examples for supervised fine-tuning, and refine generation quality through reinforcement learning with a specific Chamfer Distance-based geometric reward. Experiments demonstrate consistent gains of our STEP-LLM in geometric fidelity over the Text2CAD baseline, with improvements arising from multiple stages of our framework: the RAG module substantially enhances completeness and renderability, the DFS-based reserialization strengthens overall accuracy, and the RL further reduces geometric discrepancy. Both metrics and visual comparisons confirm that STEP-LLM generates shapes with higher fidelity than Text2CAD. These results show the feasibility of LLM-driven STEP model generation from natural language, showing its potential to democratize CAD design for manufacturing. --- ### Objective Matters: Fine-Tuning Objectives Shape Safety, Robustness, and Persona Drift Authors: Daniel Vennemeyer, Punya Syon Pandey, Phan Anh Duong, Michael Umeokoli, Samuel Ratnam Published: 2026-01-19 Categories: cs.CL, cs.LG arXiv: https://arxiv.org/abs/2601.12639v1 Fine-tuning LLMs on benign data can still degrade alignment and adversarial robustness, yet direct analysis of the role of fine-tuning objectives in shaping these safety outcomes remain limited. We present a controlled comparison of six fine-tuning objectives -- Supervised Fine-Tuning, Direct Preference Optimization, Conditional Fine-Tuning, Inoculation Prompting, Odds Ratio Preference Optimization, and KL-regularized fine-tuning -- holding data, domain, architecture, and optimization fixed. Across closed-form reasoning and open-ended generation tasks, we find that objective choice induces systematic, scale-dependent shifts along the safety-capability frontier. At small training budgets, robustness is similar across objectives but capability differs. At larger budgets, objectives diverge sharply: supervised and preference-based tuning tightly couple capability gains to increased adversarial vulnerability and persona drift, while objectives that constrain learning signals -- especially ORPO and KL-regularization -- substantially mitigate both. Fine-tuning objectives therefore matter little for safety at small scales but become a primary driver of adversarial robustness and latent persona stability as training scale increases. --- ### Mixed Precision PointPillars for Efficient 3D Object Detection with TensorRT Authors: Ninnart Fuengfusin, Keisuke Yoneda, Naoki Suganuma Published: 2026-01-19 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2601.12638v2 LIDAR 3D object detection is one of the important tasks for autonomous vehicles. Ensuring that this task operates in real-time is crucial. Toward this, model quantization can be used to accelerate the runtime. However, directly applying model quantization often leads to performance degradation due to LIDAR's wide numerical distributions and extreme outliers. To address the wide numerical distribution, we proposed a mixed precision framework designed for PointPillars. Our framework first searches for sensitive layers with post-training quantization (PTQ) by quantizing one layer at a time to 8-bit integer (INT8) and evaluating each model for average precision (AP). The top-k most sensitive layers are assigned as floating point (FP). Combinations of these layers are greedily searched to produce candidate mixed precision models, which are finalized with either PTQ or quantization-aware training (QAT). Furthermore, to handle outliers, we observe that using a very small number of calibration data reduces the likelihood of encountering outliers, thereby improving PTQ performance. Our methods provides mixed precision models without training in the PTQ pipeline, while our QAT pipeline achieves the performance competitive to FP models. With TensorRT deployment, our mixed precision models offer less latency by up to 2.538 times compared to FP32 models. --- ### Topology-Aware Multiscale Mixture of Experts for Efficient Molecular Property Prediction Authors: Long D. Nguyen, Kelin Xia, Binh P. Nguyen Published: 2026-01-19 Categories: cs.LG, cs.AI, q-bio.QM arXiv: https://arxiv.org/abs/2601.12637v1 Many molecular properties depend on 3D geometry, where non-covalent interactions, stereochemical effects, and medium- to long-range forces are determined by spatial distances and angles that cannot be uniquely captured by a 2D bond graph. Yet most 3D molecular graph neural networks still rely on globally fixed neighborhood heuristics, typically defined by distance cutoffs and maximum neighbor limits, to define local message-passing neighborhoods, leading to rigid, data-agnostic interaction budgets. We propose Multiscale Interaction Mixture of Experts (MI-MoE) to adapt interaction modeling across geometric regimes. Our contributions are threefold: (1) we introduce a distance-cutoff expert ensemble that explicitly captures short-, mid-, and long-range interactions without committing to a single cutoff; (2) we design a topological gating encoder that routes inputs to experts using filtration-based descriptors, including persistent homology features, summarizing how connectivity evolves across radii; and (3) we show that MI-MoE is a plug-in module that consistently improves multiple strong 3D molecular backbones across diverse molecular and polymer property prediction benchmark datasets, covering both regression and classification tasks. These results highlight topology-aware multiscale routing as an effective principle for 3D molecular graph learning. --- ### From Bands to Depth: Understanding Bathymetry Decisions on Sentinel-2 Authors: Satyaki Roy Chowdhury, Aswathnarayan Radhakrishnan, Hsiao Jou Hsu, Hari Subramoni, Joachim Moortgat Published: 2026-01-19 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12636v1 Deploying Sentinel-2 satellite derived bathymetry (SDB) robustly across sites remains challenging. We analyze a Swin-Transformer based U-Net model (Swin-BathyUNet) to understand how it infers depth and when its predictions are trustworthy. A leave-one-band out study ranks spectral importance to the different bands consistent with shallow water optics. We adapt ablation-based CAM to regression (A-CAM-R) and validate the reliability via a performance retention test: keeping only the top-p% salient pixels while neutralizing the rest causes large, monotonic RMSE increase, indicating explanations localize on evidence the model relies on. Attention ablations show decoder conditioned cross attention on skips is an effective upgrade, improving robustness to glint/foam. Cross-region inference (train on one site, test on another) reveals depth-dependent degradation: MAE rises nearly linearly with depth, and bimodal depth distributions exacerbate mid/deep errors. Practical guidance follows: maintain wide receptive fields, preserve radiometric fidelity in green/blue channels, pre-filter bright high variance near shore, and pair light target site fine tuning with depth aware calibration to transfer across regions. --- ### New Trends in the Stability of Sinkhorn Semigroups Authors: Pierre Del Moral, Ajay Jasra Published: 2026-01-19 Categories: math.PR, math.NA, stat.ML arXiv: https://arxiv.org/abs/2601.12633v1 Entropic optimal transport problems play an increasingly important role in machine learning and generative modelling. In contrast with optimal transport maps which often have limited applicability in high dimensions, Schrodinger bridges can be solved using the celebrated Sinkhorn's algorithm, a.k.a. the iterative proportional fitting procedure. The stability properties of Sinkhorn bridges when the number of iterations tends to infinity is a very active research area in applied probability and machine learning. Traditional proofs of convergence are mainly based on nonlinear versions of Perron-Frobenius theory and related Hilbert projective metric techniques, gradient descent, Bregman divergence techniques and Hamilton-Jacobi-Bellman equations, including propagation of convexity profiles based on coupling diffusions by reflection methods. The objective of this review article is to present, in a self-contained manner, recently developed Sinkhorn/Gibbs-type semigroup analysis based upon contraction coefficients and Lyapunov-type operator-theoretic techniques. These powerful, off-the-shelf semigroup methods are based upon transportation cost inequalities (e.g. log-Sobolev, Talagrand quadratic inequality, curvature estimates), $φ$-divergences, Kantorovich-type criteria and Dobrushin contraction-type coefficients on weighted Banach spaces as well as Wasserstein distances. This novel semigroup analysis allows one to unify and simplify many arguments in the stability of Sinkhorn algorithm. It also yields new contraction estimates w.r.t. generalized $φ$-entropies, as well as weighted total variation norms, Kantorovich criteria and Wasserstein distances. --- ### BioPulse-QA: A Dynamic Biomedical Question-Answering Benchmark for Evaluating Factuality, Robustness, and Bias in Large Language Models Authors: Kriti Bhattarai, Vipina K. Keloth, Donald Wright, Andrew Loza, Yang Ren, Hua Xu Published: 2026-01-19 Categories: cs.CL, cs.IR arXiv: https://arxiv.org/abs/2601.12632v1 Objective: Large language models (LLMs) are increasingly applied in biomedical settings, and existing benchmark datasets have played an important role in supporting model development and evaluation. However, these benchmarks often have limitations. Many rely on static or outdated datasets that fail to capture the dynamic, context-rich, and high-stakes nature of biomedical knowledge. They also carry increasing risk of data leakage due to overlap with model pretraining corpora and often overlook critical dimensions such as robustness to linguistic variation and potential demographic biases. Materials and Methods: To address these gaps, we introduce BioPulse-QA, a benchmark that evaluates LLMs on answering questions from newly published biomedical documents including drug labels, trial protocols, and clinical guidelines. BioPulse-QA includes 2,280 expert-verified question answering (QA) pairs and perturbed variants, covering both extractive and abstractive formats. We evaluate four LLMs - GPT-4o, GPT-o1, Gemini-2.0-Flash, and LLaMA-3.1 8B Instruct - released prior to the publication dates of the benchmark documents. Results: GPT-o1 achieves the highest relaxed F1 score (0.92), followed by Gemini-2.0-Flash (0.90) on drug labels. Clinical trials are the most challenging source, with extractive F1 scores as low as 0.36. Discussion and Conclusion: Performance differences are larger for paraphrasing than for typographical errors, while bias testing shows negligible differences. BioPulse-QA provides a scalable and clinically relevant framework for evaluating biomedical LLMs. --- ### Enhanced Climbing Image Nudged Elastic Band method with Hessian Eigenmode Alignment Authors: Rohit Goswami, Miha Gunde, Hannes Jónsson Published: 2026-01-19 Categories: physics.chem-ph, cond-mat.mtrl-sci, cs.LG, physics.comp-ph arXiv: https://arxiv.org/abs/2601.12630v4 Accurate determination of transition states is central to an understanding of reaction kinetics. Double-endpoint methods where both initial and final states are specified, such as the climbing image nudged elastic band (CI-NEB), identify the minimum energy path between the two and thereby the saddle point on the energy surface that is relevant for the given transition, thus providing an estimate of the transition state within the harmonic approximation of transition state theory. Such calculations can, however, incur high computational costs and may suffer stagnation on exceptionally flat or rough energy surfaces. Conversely, methods that only require specification of an initial set of atomic coordinates, such as the minimum mode following (MMF) method, offer efficiency but can converge on saddle points that are not relevant for transition of interest. Here, we present an adaptive hybrid algorithm that integrates the CI-NEB with the MMF method so as to get faster convergence to the relevant saddle point. The method is benchmarked for the Baker-Chan (BC) saddle point test set using the PET-MAD machine-learned potential as well as 59 transitions of a heptamer island on Pt(111) from the OptBench benchmark set. A Bayesian analysis of the performance shows a reduction in energy and force calculations of 57% (95% CrI: -64%, -50%) relative to CI-NEB for the BC set, while a 31% mean reduction is found for the transitions of the heptamer island. These results establish this hybrid method as a highly effective tool for high-throughput automated chemical discovery of atomic rearrangements. --- ### IC3-Evolve: Proof-/Witness-Gated Offline LLM-Driven Heuristic Evolution for IC3 Hardware Model Checking Authors: Mingkai Miao, Guangyu Hu, Ziyi Yang, Hongce Zhang Published: 2026-01-18 Categories: cs.AI arXiv: https://arxiv.org/abs/2604.03232v1 IC3, also known as property-directed reachability (PDR), is a commonly-used algorithm for hardware safety model checking. It checks if a state transition system complies with a given safety property. IC3 either returns UNSAFE (indicating property violation) with a counterexample trace, or SAFE with a checkable inductive invariant as the proof to safety. In practice, the performance of IC3 is dominated by a large web of interacting heuristics and implementation choices, making manual tuning costly, brittle, and hard to reproduce. This paper presents IC3-Evolve, an automated offline code-evolution framework that utilizes an LLM to propose small, slot-restricted and auditable patches to an IC3 implementation. Crucially, every candidate patch is admitted only through proof- /witness-gated validation: SAFE runs must emit a certificate that is independently checked, and UNSAFE runs must emit a replayable counterexample trace, preventing unsound edits from being deployed. Since the LLM is used only offline, the deployed artifact is a standalone evolved checker with zero ML/LLM inference overhead and no runtime model dependency. We evolve on the public hardware model checking competition (HWMCC) benchmark and evaluate the generalizability on unseen public and industrial model checking benchmarks, showing that IC3-Evolve can reliably discover practical heuristic improvements under strict correctness gates. --- ### Holos: A Web-Scale LLM-Based Multi-Agent System for the Agentic Web Authors: Xiaohang Nie, Zihan Guo, Zicai Cui, Jiachi Yang, Zeyi Chen, Leheyi De, Yu Zhang, Junwei Liao, Bo Huang, Yingxuan Yang, Zhi Han, Zimian Peng et al. Published: 2026-01-18 Categories: cs.AI, cs.MA arXiv: https://arxiv.org/abs/2604.02334v1 As large language models (LLM)-driven agents transition from isolated task solvers to persistent digital entities, the emergence of the Agentic Web, an ecosystem where heterogeneous agents autonomously interact and co-evolve, marks a pivotal shift toward Artificial General Intelligence (AGI). However, LLM-based multi-agent systems (LaMAS) are hindered by open-world issues such as scaling friction, coordination breakdown, and value dissipation. To address these challenges, we introduce Holos, a web-scale LaMAS architected for long-term ecological persistence. Holos adopts a five-layer architecture, with core modules primarily featuring the Nuwa engine for high-efficiency agent generation and hosting, a market-driven Orchestrator for resilient coordination, and an endogenous value cycle to achieve incentive compatibility. By bridging the gap between micro-level collaboration and macro-scale emergence, Holos hopes to lay the foundation for the next generation of the self-organizing and continuously evolving Agentic Web. We have publicly released Holos (accessible at https://holosai.io), providing a resource for the community and a testbed for future research in large-scale agentic ecosystems. --- ### Spark-LLM-Eval: A Distributed Framework for Statistically Rigorous Large Language Model Evaluation Authors: Subhadip Mitra Published: 2026-01-18 Categories: cs.DC, cs.CL, cs.LG arXiv: https://arxiv.org/abs/2603.28769v1 Evaluating large language models at scale remains a practical bottleneck for many organizations. While existing evaluation frameworks work well for thousands of examples, they struggle when datasets grow to hundreds of thousands or millions of samples. This scale is common when assessing model behavior across diverse domains or conducting comprehensive regression testing. We present Spark-LLM-Eval, a distributed evaluation framework built natively on Apache Spark. The system treats evaluation as a data-parallel problem, partitioningexamplesacrossexecutorsandaggregatingresultswithproperstatistical accounting. Beyond raw throughput, we emphasize statistical rigor: every reported metric includes bootstrap confidence intervals, and model comparisons come with appropriate significance tests (paired t-tests, McNemar's test, or Wilcoxon signed-rank, depending on the metric type). The framework also addresses the cost problem inherent in LLM evaluation through content-addressable response caching backed by Delta Lake, which allows iterating on metric definitions without re-running inference. We describe the system architecture, the statistical methodology, and report benchmark results showing linear scaling with cluster size. The framework and all evaluation code are available as open source. --- ### An Embodied Companion for Visual Storytelling Authors: Patrick Tresset, Markus Wulfmeier Published: 2026-01-18 Categories: cs.HC, cs.AI, cs.GR, cs.RO arXiv: https://arxiv.org/abs/2603.05511v1 As artificial intelligence shifts from pure tool for delegation toward agentic collaboration, its use in the arts can shift beyond the exploration of machine autonomy toward synergistic co-creation. While our earlier robotic works utilized automation to distance the artist's intent from the final mark, we present Companion: an artistic apparatus that integrates a drawing robot with Large Language Models (LLMs) to re-center human-machine presence. By leveraging in-context learning and real-time tool use, the system engages in bidirectional interaction via speech and sketching. This approach transforms the robot from a passive executor into a playful co-creative partner capable of driving shared visual storytelling into unexpected aesthetic territories. To validate this collaborative shift, we employed the Consensual Assessment Technique (CAT) with a panel of seven art-world experts. Results confirm that the system produces works with a distinct aesthetic identity and professional exhibition merit, demonstrating the potential of AI as a highly capable artistic collaborator. --- ### Exploring Human-in-the-Loop Themes in AI Application Development: An Empirical Thematic Analysis Authors: Parm Suksakul, Nathan Kittichaikoonkij, Nakhin Polthai, Aung Pyae Published: 2026-01-18 Categories: cs.HC, cs.AI, cs.CY arXiv: https://arxiv.org/abs/2603.05510v1 Developing and deploying AI applications in organizations is challenging when human decision authority and oversight are underspecified across the system lifecycle. Although Human-in-the-Loop (HITL) and Human-Centered AI (HCAI) principles are widely acknowledged, operational guidance for structuring roles, checkpoints, and feedback mechanisms remains fragmented. We report a multi-source qualitative study: a retrospective diary study of a customer-support chatbot and semi-structured interviews with eight AI experts from academia and industry. Through five-cycle thematic analysis of 1,435 codewords, we derive four themes: AI Governance and Human Authority, Human-in-the-Loop Iterative Refinement, AI System Lifecycle and Operational Constraints, and Human-AI Team Collaboration and Coordination. These themes provide empirical inputs for subsequent HITL framework design and validation. --- ### Visualizing and Benchmarking LLM Factual Hallucination Tendencies via Internal State Analysis and Clustering Authors: Nathan Mao, Varun Kaushik, Shreya Shivkumar, Parham Sharafoleslami, Kevin Zhu, Sunishchal Dev Published: 2026-01-18 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2602.11167v1 Large Language Models (LLMs) often hallucinate, generating nonsensical or false information that can be especially harmful in sensitive fields such as medicine or law. To study this phenomenon systematically, we introduce FalseCite, a curated dataset designed to capture and benchmark hallucinated responses induced by misleading or fabricated citations. Running GPT-4o-mini, Falcon-7B, and Mistral 7-B through FalseCite, we observed a noticeable increase in hallucination activity for false claims with deceptive citations, especially in GPT-4o-mini. Using the responses from FalseCite, we can also analyze the internal states of hallucinating models, visualizing and clustering the hidden state vectors. From this analysis, we noticed that the hidden state vectors, regardless of hallucination or non-hallucination, tend to trace out a distinct horn-like shape. Our work underscores FalseCite's potential as a foundation for evaluating and mitigating hallucinations in future LLM research. --- ### Large Language Models Predict Functional Outcomes after Acute Ischemic Stroke Authors: Anjali K. Kapoor, Anton Alyakin, Jin Vivian Lee, Eunice Yang, Annelene M. Schulze, Krithik Vishwanath, Jinseok Lee, Yindalon Aphinyanaphongs, Howard Riina, Jennifer A. Frontera, Eric Karl Oermann Published: 2026-01-18 Categories: cs.LG, cs.AI, cs.CL, cs.CY arXiv: https://arxiv.org/abs/2602.10119v1 Accurate prediction of functional outcomes after acute ischemic stroke can inform clinical decision-making and resource allocation. Prior work on modified Rankin Scale (mRS) prediction has relied primarily on structured variables (e.g., age, NIHSS) and conventional machine learning. The ability of large language models (LLMs) to infer future mRS scores directly from routine admission notes remains largely unexplored. We evaluated encoder (BERT, NYUTron) and generative (Llama-3.1-8B, MedGemma-4B) LLMs, in both frozen and fine-tuned settings, for discharge and 90-day mRS prediction using a large, real-world stroke registry. The discharge outcome dataset included 9,485 History and Physical notes and the 90-day outcome dataset included 1,898 notes from the NYU Langone Get With The Guidelines-Stroke registry (2016-2025). Data were temporally split with the most recent 12 months held out for testing. Performance was assessed using exact (7-class) mRS accuracy and binary functional outcome (mRS 0-2 vs. 3-6) accuracy and compared against established structured-data baselines incorporating NIHSS and age. Fine-tuned Llama achieved the highest performance, with 90-day exact mRS accuracy of 33.9% (95% CI, 27.9-39.9%) and binary accuracy of 76.3% (95% CI, 70.7-81.9%). Discharge performance reached 42.0% (95% CI, 39.0-45.0%) exact accuracy and 75.0% (95% CI, 72.4-77.6%) binary accuracy. For 90-day prediction, Llama performed comparably to structured-data baselines. Fine-tuned LLMs can predict post-stroke functional outcomes from admission notes alone, achieving performance comparable to models requiring structured variable abstraction. Our findings support the development of text-based prognostic tools that integrate seamlessly into clinical workflows without manual data extraction. --- ### IFNSO: Iteration-Free Newton-Schulz Orthogonalization Authors: Chen Hu, Qianxi Zhao, Xiaochen Yuan, Hong Zhang, Ding Yuan, Yanbin Wu, Xiying Li Published: 2026-01-18 Categories: cs.LG, cs.AI, math.NA arXiv: https://arxiv.org/abs/2602.02500v3 The Newton-Schulz (NS) iteration has become a key technique for orthogonalization in optimizers such as Muon and for optimization on the Stiefel manifold. Despite its effectiveness, the conventional NS iteration incurs significant computational overhead due to repeated high-dimensional matrix multiplications. To overcome these limitations, we propose Iteration-Free Newton-Schulz Orthogonalization (IFNSO), a novel framework that consolidates the traditional iterative structure into a unified and Iteration-Free formulation. By analyzing the contribution of individual matrix powers, we streamline the process by removing insignificant terms and introducing a polynomial with learnable coefficients. These coefficients are optimized to ensure both superior computational efficiency and stable convergence. Extensive experiments demonstrate that IFNSO achieves superior performance compared to existing methods. Our code is available at: https://github.com/greekinRoma/Unified_Newton_Schulz_Orthogonalization. --- ### LOGOS-CA: A Cellular Automaton Using Natural Language as State and Rule Authors: Keishu Utimula Published: 2026-01-18 Categories: nlin.CG, cs.AI, cs.FL, cs.NE arXiv: https://arxiv.org/abs/2602.00036v2 Large Language Models (LLMs), trained solely on massive text data, have achieved high performance on the Winograd Schema Challenge (WSC), a benchmark proposed to measure commonsense knowledge and reasoning abilities about the real world. This suggests that the language produced by humanity describes a significant portion of the world with considerable nuance. In this study, we attempt to harness the high expressive power of language within cellular automata. Specifically, we express cell states and rules in natural language and delegate their updates to an LLM. Through this approach, cellular automata can transcend the constraints of merely numerical states and fixed rules, providing us with a richer platform for simulation. Here, we propose LOGOS-CA (Language Oriented Grid Of Statements - Cellular Automaton) as a natural framework to achieve this and examine its capabilities. We confirmed that LOGOS-CA successfully performs simple forest fire simulations and also serves as an intriguing subject for investigation from an Artificial Life (ALife) perspective. In this paper, we report the results of these experiments and discuss directions for future research using LOGOS-CA. --- ### Asynchronous MultiAgent Reinforcement Learning for 5G Routing under Side Constraints Authors: Sebastian Racedo, Brigitte Jaumard, Oscar Delgado, Meysam Masoudi Published: 2026-01-18 Categories: cs.NI, cs.DC, cs.LG, cs.MA arXiv: https://arxiv.org/abs/2602.00035v1 Networks in the current 5G and beyond systems increasingly carry heterogeneous traffic with diverse quality-of-service constraints, making real-time routing decisions both complex and time-critical. A common approach, such as a heuristic with human intervention or training a single centralized RL policy or synchronizing updates across multiple learners, struggles with scalability and straggler effects. We address this by proposing an asynchronous multi-agent reinforcement learning (AMARL) framework in which independent PPO agents, one per service, plan routes in parallel and commit resource deltas to a shared global resource environment. This coordination by state preserves feasibility across services and enables specialization for service-specific objectives. We evaluate the method on an O-RAN like network simulation using nearly real-time traffic data from the city of Montreal. We compared against a single-agent PPO baseline. AMARL achieves a similar Grade of Service (acceptance rate) (GoS) and end-to-end latency, with reduced training wall-clock time and improved robustness to demand shifts. These results suggest that asynchronous, service-specialized agents provide a scalable and practical approach to distributed routing, with applicability extending beyond the O-RAN domain. --- ### Synthetic Student Responses: LLM-Extracted Features for IRT Difficulty Parameter Estimation Authors: Matias Hoyl Published: 2026-01-18 Categories: cs.CY, cs.AI arXiv: https://arxiv.org/abs/2602.00034v1 Educational assessment relies heavily on knowing question difficulty, traditionally determined through resource-intensive pre-testing with students. This creates significant barriers for both classroom teachers and assessment developers. We investigate whether Item Response Theory (IRT) difficulty parameters can be accurately estimated without student testing by modeling the response process and explore the relative contribution of different feature types to prediction accuracy. Our approach combines traditional linguistic features with pedagogical insights extracted using Large Language Models (LLMs), including solution step count, cognitive complexity, and potential misconceptions. We implement a two-stage process: first training a neural network to predict how students would respond to questions, then deriving difficulty parameters from these simulated response patterns. Using a dataset of over 250,000 student responses to mathematics questions, our model achieves a Pearson correlation of approximately 0.78 between predicted and actual difficulty parameters on completely unseen questions. --- ### Happy Young Women, Grumpy Old Men? Emotion-Driven Demographic Biases in Synthetic Face Generation Authors: Mengting Wei, Aditya Gulati, Guoying Zhao, Nuria Oliver Published: 2026-01-18 Categories: cs.CY, cs.AI, cs.CV arXiv: https://arxiv.org/abs/2602.00032v3 Synthetic faces from text-to-image (T2I) models pervade digital media, yet their demographic biases under emotionally conditioned prompts remain poorly understood. We aim to systematically audit how emotionally conditioned prompts affect demographic and perceived-attractiveness biases in synthetic faces generated by T2I models, with particular attention to intersectional patterns and cross-ecosystem differences across model families. We audited eight (4 Western and 4 Chinese) T2I models and generated 56,000 faces under seven prompt conditions: a neutral baseline and six emotion conditions. We quantified biases in gender, race, age, and perceived attractiveness using information-theoretic divergence metrics. We further conducted intersectional analyses across combined demographic attributes and compared patterns between the Western and Chinese model groups to assess cross-ecosystem consistency and divergence in bias behavior. All models show strong overrepresentation of young faces, and most also overrepresent White-coded individuals. Intersectional analysis reveals compound underrepresentation or near-erasure of specific demographic combinations, such as young x female x Black faces, which are largely absent across models and are not captured by single-attribute audits. Emotion prompts act as additional demographic selectors: negatively valenced emotions (including sadness and fear) consistently shift outputs toward White, middle-aged, male-coded faces. This produces a valence-driven mapping that is also associated with lower perceived attractiveness in generated faces. These findings indicate that demographic bias in T2I face generation is both pervasive and shaped by emotional conditioning. They underscore the need for intersectional, emotion-conditioned, and multilingual demographic audits as part of standard pre-deployment evaluation practices. --- ### RAPTOR-AI for Disaster OODA Loop: Hierarchical Multimodal RAG with Experience-Driven Agentic Decision-Making Authors: Takato Yasuno Published: 2026-01-18 Categories: cs.LG arXiv: https://arxiv.org/abs/2602.00030v2 Humanitarian Assistance and Disaster Relief (HADR) operations demand rapid synthesis of multimodal information for time-critical decision-making under extreme uncertainty. Traditional information systems struggle with the fragmented, multimodal nature of disaster data and lack adaptive reasoning capabilities essential for dynamic emergency contexts. This work introduces RAPTOR-AI, an agentic multimodal Retrieval-Augmented Generation (RAG) framework that advances beyond conventional static knowledge bases by implementing dynamic, experience-driven decision support for disaster response. The system addresses HADR requirements across initial rescue, recovery, and reconstruction phases through three key innovations: hierarchical multimodal knowledge construction from diverse sources (textual reports, aerial imagery, historical documentation), entropy-aware agentic control that dynamically selects optimal retrieval strategies based on situational context, and experiential knowledge integration using LoRA adaptation for both expert and non-expert responders. The framework constructs hierarchical knowledge trees from 46 tsunami-related PDFs (2,378 pages) using BLIP-based image understanding, ColVBERT embeddings, and long-context summarization within the OODA loop (Observe, Orient, Decide, Act) tactical framework. Experiments demonstrate significant improvements over existing approaches: 23\% improvement in retrieval precision, 31\% better situational grounding, and 27\% enhanced task decomposition accuracy, with efficient scaling up to 3,000 document chunks. --- ### Construct, Align, and Reason: Large Ontology Models for Enterprise Knowledge Management Authors: Yao Zhang, Hongyin Zhu Published: 2026-01-18 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2602.00029v1 Enterprise-scale knowledge management faces significant challenges in integrating multi-source heterogeneous data and enabling effective semantic reasoning. Traditional knowledge graphs often struggle with implicit relationship discovery and lack sufficient semantic understanding for complex question answering. To address these limitations, we introduce a unified construct--align--reason framework, the large ontology model (LOM). We first build a dual-layer enterprise ontology from structured databases and unstructured text, subsequently fusing these sources into a comprehensive enterprise ontology. To enable instruction-aligned reasoning, we propose a unified three-stage training pipeline: ontology instruction fine-tuning to improve structural understanding; text-ontology grounding to strengthen node semantic encoding; and multi-task instruction tuning on ontology-language pairs with curriculum learning to enhance semantic reasoning and generation. We also construct comprehensive training and evaluation datasets covering diverse ontology reasoning tasks. On this benchmark, our 4B-parameter LOM achieves 89.47% accuracy and outperforms DeepSeek-V3.2 on complex graph reasoning, indicating effective fusion of ontology structure and language. --- ### Stablecoin Design with Adversarial-Robust Multi-Agent Systems via Trust-Weighted Signal Aggregation Authors: Shengwei You, Aditya Joshi, Andrey Kuehlkamp, Jarek Nabrzyski Published: 2026-01-18 Categories: q-fin.RM, cs.AI, cs.CR, q-fin.CP arXiv: https://arxiv.org/abs/2601.22168v1 Algorithmic stablecoins promise decentralized monetary stability by maintaining a target peg through programmatic reserve management. Yet, their reserve controllers remain vulnerable to regime-blind optimization, calibrating risk parameters on fair-weather data while ignoring tail events that precipitate cascading failures. The March 2020 Black Thursday collapse, wherein MakerDAO's collateral auctions yielded $8.3M in losses and a 15% peg deviation, exposed a critical gap: existing models like SAS systematically omit extreme volatility regimes from covariance estimates, producing allocations optimal in expectation but catastrophic under adversarial stress. We present MVF-Composer, a trust-weighted Mean-Variance Frontier reserve controller incorporating a novel Stress Harness for risk-state estimation. Our key insight is deploying multi-agent simulations as adversarial stress-testers: heterogeneous agents (traders, liquidity providers, attackers) execute protocol actions under crisis scenarios, exposing reserve vulnerabilities before they manifest on-chain. We formalize a trust-scoring mechanism T: A -> (0,1) that down-weights signals from agents exhibiting manipulative behavior, ensuring the risk-state estimator remains robust to signal injection and Sybil attacks. Across 1,200 randomized scenarios with injected Black-Swan shocks (10% collateral drawdown, 50% sentiment collapse, coordinated redemption attacks), MVF-Composer reduces peak peg deviation by 57% and mean recovery time by 3.1x relative to SAS baselines. Ablation studies confirm the trust layer accounts for 23% of stability gains under adversarial conditions, achieving 72% adversarial agent detection. Our system runs on commodity hardware, requires no on-chain oracles beyond standard price feeds, and provides a reproducible framework for stress-testing DeFi reserve policies. --- ### Analyzing the Temporal Factors for Anxiety and Depression Symptoms with the Rashomon Perspective Authors: Mustafa Cavus, Przemysław Biecek, Julian Tejada, Fernando Marmolejo-Ramos, Andre Faro Published: 2026-01-18 Categories: stat.AP, cs.LG arXiv: https://arxiv.org/abs/2601.20874v1 This paper introduces a new modeling perspective in the public mental health domain to provide a robust interpretation of the relations between anxiety and depression, and the demographic and temporal factors. This perspective particularly leverages the Rashomon Effect, where multiple models exhibit similar predictive performance but rely on diverse internal structures. Instead of considering these multiple models, choosing a single best model risks masking alternative narratives embedded in the data. To address this, we employed this perspective in the interpretation of a large-scale psychological dataset, specifically focusing on the Patient Health Questionnaire-4. We use a random forest model combined with partial dependence profiles to rigorously assess the robustness and stability of predictive relationships across the resulting Rashomon set, which consists of multiple models that exhibit similar predictive performance. Our findings confirm that demographic variables \texttt{age}, \texttt{sex}, and \texttt{education} lead to consistent structural shifts in anxiety and depression risk. Crucially, we identify significant temporal effects: risk probability demonstrates clear diurnal and circaseptan fluctuations, peaking during early morning hours. This work demonstrates the necessity of moving beyond the best model to analyze the entire Rashomon set. Our results highlight that the observed variability, particularly due to circadian and circaseptan rhythms, must be meticulously considered for robust interpretation in psychological screening. We advocate for a multiplicity-aware approach to enhance the stability and generalizability of ML-based conclusions in mental health research. --- ### Prometheus Mind: Retrofitting Memory to Frozen Language Models Authors: Mark Wind Published: 2026-01-18 Categories: cs.AI arXiv: https://arxiv.org/abs/2601.15324v2 Adding memory to pretrained language models typically requires architectural changes or weight modification. We present Prometheus Mind, which retrofits memory to a frozen Qwen3-4B using 11 modular adapters (530MB, 7% overhead) -- fully reversible by removing the adapters. Building this system required solving four problems: (1) Extraction -- we develop Contrastive Direction Discovery (CDD), which finds semantic directions via minimal pairs without labeled data. (2) Training -- end-to-end optimization collapses; stage-wise training of each adapter on simple proxy tasks succeeds. (3) Injection -- learned encoders fail to generalize; we find that lm_head-weight rows already provide the mapping we need, requiring no training. (4) Hidden state collapse -- transformers make ``wife'' and ``brother'' 0.98+ similar; we train projections to recover distinction (0.98 $\rightarrow$ 0.09). On PrometheusExtract-132 (132 cases), the system achieves 94.4% retrieval on clean inputs (n=54, 95% CI: (84.9%, 98.1%)), degrading to 19.4% on informal inputs with ellipsis, filler words, or implicit subjects (n=36). The primary bottleneck is relation classification (47.3% accuracy), responsible for most extraction errors. --- ### An Optimized Decision Tree-Based Framework for Explainable IoT Anomaly Detection Authors: Ashikuzzaman, Md. Shawkat Hossain, Jubayer Abdullah Joy, Md Zahid Akon, Md Manjur Ahmed, Md. Naimul Islam Published: 2026-01-18 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2601.14305v1 The increase in the number of Internet of Things (IoT) devices has tremendously increased the attack surface of cyber threats thus making a strong intrusion detection system (IDS) with a clear explanation of the process essential towards resource-constrained environments. Nevertheless, current IoT IDS systems are usually traded off with detection quality, model elucidability, and computational effectiveness, thus the deployment on IoT devices. The present paper counteracts these difficulties by suggesting an explainable AI (XAI) framework based on an optimized Decision Tree classifier with both local and global importance methods: SHAP values that estimate feature attribution using local explanations, and Morris sensitivity analysis that identifies the feature importance in a global view. The proposed system attains the state of art on the test performance with 99.91% accuracy, F1-score of 99.51% and Cohen Kappa of 0.9960 and high stability is confirmed by a cross validation mean accuracy of 98.93%. Efficiency is also enhanced in terms of computations to provide faster inferences compared to those that are generalized in ensemble models. SrcMac has shown as the most significant predictor in feature analyses according to SHAP and Morris methods. Compared to the previous work, our solution eliminates its major drawback lack because it allows us to apply it to edge devices and, therefore, achieve real-time processing, adhere to the new regulation of transparency in AI, and achieve high detection rates on attacks of dissimilar classes. This combination performance of high accuracy, explainability, and low computation make the framework useful and reliable as a resource-constrained IoT security problem in real environments. --- ### Guided by the Plan: Enhancing Faithful Autoregressive Text-to-Audio Generation with Guided Decoding Authors: Juncheng Wang, Zhe Hu, Chao Xu, Siyue Ren, Yuxiang Feng, Yang Liu, Baigui Sun, Shujun Wang Published: 2026-01-18 Categories: cs.CL, cs.SD, eess.AS arXiv: https://arxiv.org/abs/2601.14304v1 Autoregressive (AR) models excel at generating temporally coherent audio by producing tokens sequentially, yet they often falter in faithfully following complex textual prompts, especially those describing complex sound events. We uncover a surprising capability in AR audio generators: their early prefix tokens implicitly encode global semantic attributes of the final output, such as event count and sound-object category, revealing a form of implicit planning. Building on this insight, we propose Plan-Critic, a lightweight auxiliary model trained with a Generalized Advantage Estimation (GAE)-inspired objective to predict final instruction-following quality from partial generations. At inference time, Plan-Critic enables guided exploration: it evaluates candidate prefixes early, prunes low-fidelity trajectories, and reallocates computation to high-potential planning seeds. Our Plan-Critic-guided sampling achieves up to a 10-point improvement in CLAP score over the AR baseline-establishing a new state of the art in AR text-to-audio generation-while maintaining computational parity with standard best-of-N decoding. This work bridges the gap between causal generation and global semantic alignment, demonstrating that even strictly autoregressive models can plan ahead. --- ### DDSA: Dual-Domain Strategic Attack for Spatial-Temporal Efficiency in Adversarial Robustness Testing Authors: Jinwei Hu, Shiyuan Meng, Yi Dong, Xiaowei Huang Published: 2026-01-18 Categories: cs.CR, cs.AI, cs.PF arXiv: https://arxiv.org/abs/2601.14302v2 Image transmission and processing systems in resource-critical applications face significant challenges from adversarial perturbations that compromise mission-specific object classification. Current robustness testing methods require excessive computational resources through exhaustive frame-by-frame processing and full-image perturbations, proving impractical for large-scale deployments where massive image streams demand immediate processing. This paper presents DDSA (Dual-Domain Strategic Attack), a resource-efficient adversarial robustness testing framework that optimizes testing through temporal selectivity and spatial precision. We introduce a scenario-aware trigger function that identifies critical frames requiring robustness evaluation based on class priority and model uncertainty, and employ explainable AI techniques to locate influential pixel regions for targeted perturbation. Our dual-domain approach achieves substantial temporal-spatial resource conservation while maintaining attack effectiveness. The framework enables practical deployment of comprehensive adversarial robustness testing in resource-constrained real-time applications where computational efficiency directly impacts mission success. --- ### Linear Mechanisms for Spatiotemporal Reasoning in Vision Language Models Authors: Raphi Kang, Hongqiao Chen, Georgia Gkioxari, Pietro Perona Published: 2026-01-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12626v1 Spatio-temporal reasoning is a remarkable capability of Vision Language Models (VLMs), but the underlying mechanisms of such abilities remain largely opaque. We postulate that visual/geometrical and textual representations of spatial structure must be combined at some point in VLM computations. We search for such confluence, and ask whether the identified representation can causally explain aspects of input-output model behavior through a linear model. We show empirically that VLMs encode object locations by linearly binding \textit{spatial IDs} to textual activations, then perform reasoning via language tokens. Through rigorous causal interventions we demonstrate that these IDs, which are ubiquitous across the model, can systematically mediate model beliefs at intermediate VLM layers. Additionally, we find that spatial IDs serve as a diagnostic tool for identifying limitations in existing VLMs, and as a valuable learning signal. We extend our analysis to video VLMs and identify an analogous linear temporal ID mechanism. By characterizing our proposed spatiotemporal ID mechanism, we elucidate a previously underexplored internal reasoning process in VLMs, toward improved interpretability and the principled design of more aligned and capable models. We release our code for reproducibility: https://github.com/Raphoo/linear-mech-vlms. --- ### Towards Robust Universal Perturbation Attacks: A Float-Coded, Penalty-Driven Evolutionary Approach Authors: Shiqi Wang, Mahdi Khosravy, Neeraj Gupta, Olaf Witkowski Published: 2026-01-18 Categories: cs.LG, cs.CV arXiv: https://arxiv.org/abs/2601.12624v1 Universal adversarial perturbations (UAPs) have garnered significant attention due to their ability to undermine deep neural networks across multiple inputs using a single noise pattern. Evolutionary algorithms offer a promising approach to generating such perturbations due to their ability to navigate non-convex, gradient-free landscapes. In this work, we introduce a float-coded, penalty-driven single-objective evolutionary framework for UAP generation that achieves lower visibility perturbations while enhancing attack success rates. Our approach leverages continuous gene representations aligned with contemporary deep learning scales, incorporates dynamic evolutionary operators with adaptive scheduling, and utilizes a modular PyTorch implementation for seamless integration with modern architectures. Additionally, we ensure the universality of the generated perturbations by testing across diverse models and by periodically switching batches to prevent overfitting. Experimental results on the ImageNet dataset demonstrate that our framework consistently produces perturbations with smaller norms, higher misclassification effectiveness, and faster convergence compared to existing evolutionary-based methods. These findings highlight the robustness and scalability of our approach for universal adversarial attacks across various deep learning architectures. --- ### Learning Deterministic Finite-State Machines from the Prefixes of a Single String is NP-Complete Authors: Radu Cosmin Dumitru, Ryo Yoshinaka, Ayumi Shinohara Published: 2026-01-18 Categories: cs.FL, cs.LG arXiv: https://arxiv.org/abs/2601.12621v2 It is well known that computing a minimum deterministic finite automaton consistent with a given set of positive and negative examples is NP-hard. Previous work has identified conditions on the input sample under which the problem becomes tractable or remains hard. In this paper, we study the computational complexity of the case where the input sample is prefix-closed. This formulation is equivalent to computing a minimum Moore machine consistent with observations along its runs. We show that the problem is NP-hard to approximate when the sample set consists of all prefixes of binary strings. Furthermore, we show that the problem remains NP-hard as a decision problem even when the sample set consists of the prefixes of a single binary string. Our argument also extends to the corresponding problem for Mealy machines. --- ### Disagreement as Data: Reasoning Trace Analytics in Multi-Agent Systems Authors: Elham Tajik, Conrad Borchers, Bahar Shahrokhian, Sebastian Simon, Ali Keramati, Sonika Pal, Sreecharan Sankaranarayanan Published: 2026-01-18 Categories: cs.CL arXiv: https://arxiv.org/abs/2601.12618v1 Learning analytics researchers often analyze qualitative student data such as coded annotations or interview transcripts to understand learning processes. With the rise of generative AI, fully automated and human-AI workflows have emerged as promising methods for analysis. However, methodological standards to guide such workflows remain limited. In this study, we propose that reasoning traces generated by large language model (LLM) agents, especially within multi-agent systems, constitute a novel and rich form of process data to enhance interpretive practices in qualitative coding. We apply cosine similarity to LLM reasoning traces to systematically detect, quantify, and interpret disagreements among agents, reframing disagreement as a meaningful analytic signal. Analyzing nearly 10,000 instances of agent pairs coding human tutoring dialog segments, we show that LLM agents' semantic reasoning similarity robustly differentiates consensus from disagreement and correlates with human coding reliability. Qualitative analysis guided by this metric reveals nuanced instructional sub-functions within codes and opportunities for conceptual codebook refinement. By integrating quantitative similarity metrics with qualitative review, our method has the potential to improve and accelerate establishing inter-rater reliability during coding by surfacing interpretive ambiguity, especially when LLMs collaborate with humans. We discuss how reasoning-trace disagreements represent a valuable new class of analytic signals advancing methodological rigor and interpretive depth in educational research. --- ### Creating Disability Story Videos with Generative AI: Motivation, Expression, and Sharing Authors: Shuo Niu, Dylan Clements, Hyungsin Kim Published: 2026-01-18 Categories: cs.HC, cs.AI arXiv: https://arxiv.org/abs/2601.12617v1 Generative AI (GenAI) is both promising and challenging in supporting people with disabilities (PwDs) in creating stories about disability. GenAI can reduce barriers to media production and inspire the creativity of PwDs, but it may also introduce biases and imperfections that hinder its adoption for personal expression. In this research, we examine how nine PwD from a disability advocacy group used GenAI to create videos sharing their disability experiences. Grounded in digital storytelling theory, we explore the motivations, expression, and sharing of PwD-created GenAI story videos. We conclude with a framework of momentous depiction, which highlights four core affordances of GenAI that either facilitate or require improvements to better support disability storytelling: non-capturable depiction, identity concealment and representation, contextual realism and consistency, and emotional articulation. Based on this framework, we further discuss design implications for GenAI in relation to story completion, media formats, and corrective mechanisms. --- ### Deterministic and probabilistic neural surrogates of global hybrid-Vlasov simulations Authors: Daniel Holmberg, Ivan Zaitsev, Markku Alho, Ioanna Bouri, Fanni Franssila, Haewon Jeong, Minna Palmroth, Teemu Roos Published: 2026-01-18 Categories: physics.space-ph, cs.LG, physics.plasm-ph arXiv: https://arxiv.org/abs/2601.12614v4 Hybrid-Vlasov simulations resolve ion-kinetic effects in the solar wind-magnetosphere interaction, but even 5D (2D + 3V) configurations are computationally expensive. We show that graph-based machine learning emulators can learn the spatiotemporal evolution of electromagnetic fields and lower-order moments of the ion velocity distribution function in near-Earth space from four 5D Vlasiator runs, each driven by steady solar wind conditions. The upstream ion number density is systematically varied between the runs, while the grid spacing is held constant, to scan the ratio of ion inertial length to grid size. Using a graph neural network (GNN) operating on the 2D spatial simulation grid comprising 670k cells, we demonstrate that both a deterministic forecasting model (Graph-FM) and a probabilistic ensemble forecasting model (Graph-EFM) based on a latent variable formulation produce accurate predictions of future plasma states. A divergence penalty is incorporated to encourage divergence-freeness in the magnetic fields. For the probabilistic model, a continuous ranked probability score objective is added to improve the calibration of the ensemble forecasts. In terms of wall time per output step, the trained emulators run over two orders of magnitude faster on a single GPU than the Vlasiator simulations on 100 CPUs. Most forecasted fields have Pearson correlations above 0.95 at 50 seconds lead time. Fields that exhibit degenerate (near-zero) distributions in the 5D setting are more challenging for the emulator to keep well correlated. The ensemble forecasts remain underdispersive, with spread-skill ratios of approximately 0.2-0.3, and thus provide spatially structured relative uncertainty estimates. Overall, these results demonstrate that GNNs provide a viable framework for rapid ensemble generation in hybrid-Vlasov modeling and highlight promising directions for future work. ---