# Research Papers — 2026-01 · bundle page 97 of 189 > Full-text markdown of papers #4801–4850 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/97 [← Previous bundle](https://aipod.dev/papers/archive/2026-01/page/96.md) [Next bundle →](https://aipod.dev/papers/archive/2026-01/page/98.md) ### Neural Isomorphic Fields: A Transformer-based Algebraic Numerical Embedding Authors: Hamidreza Sadeghi, Saeedeh Momtazi, Reza Safabakhsh Published: 2026-01-17 Categories: cs.LG, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2601.12095v1 Neural network models often face challenges when processing very small or very large numbers due to issues such as overflow, underflow, and unstable output variations. To mitigate these problems, we propose using embedding vectors for numbers instead of directly using their raw values. These embeddings aim to retain essential algebraic properties while preventing numerical instabilities. In this paper, we introduce, for the first time, a fixed-length number embedding vector that preserves algebraic operations, including addition, multiplication, and comparison, within the field of rational numbers. We propose a novel Neural Isomorphic Field, a neural abstraction of algebraic structures such as groups and fields. The elements of this neural field are embedding vectors that maintain algebraic structure during computations. Our experiments demonstrate that addition performs exceptionally well, achieving over 95 percent accuracy on key algebraic tests such as identity, closure, and associativity. In contrast, multiplication exhibits challenges, with accuracy ranging from 53 percent to 73 percent across various algebraic properties. These findings highlight the model's strengths in preserving algebraic properties under addition while identifying avenues for further improvement in handling multiplication. --- ### PTL-PINNs: Perturbation-Guided Transfer Learning with Physics- Informed Neural Networks for Nonlinear Systems Authors: Duarte Alexandrino, Ben Moseley, Pavlos Protopapas Published: 2026-01-17 Categories: cs.LG arXiv: https://arxiv.org/abs/2601.12093v1 Accurately and efficiently solving nonlinear differential equations is crucial for modeling dynamic behavior across science and engineering. Physics-Informed Neural Networks (PINNs) have emerged as a powerful solution that embeds physical laws in training by enforcing equation residuals. However, these struggle to model nonlinear dynamics, suffering from limited generalization across problems and long training times. To address these limitations, we propose a perturbation-guided transfer learning framework for PINNs (PTL-PINN), which integrates perturbation theory with transfer learning to efficiently solve nonlinear equations. Unlike gradient-based transfer learning, PTL-PINNs solve an approximate linear perturbative system using closed-form expressions, enabling rapid generalization with the time complexity of matrix-vector multiplication. We show that PTL-PINNs achieve accuracy comparable to various Runge-Kutta methods, with computational speeds up to one order of magnitude faster. To benchmark performance, we solve a broad set of problems, including nonlinear oscillators across various damping regimes, the equilibrium-centered Lotka-Volterra system, the KPP-Fisher and the Wave equation. Since perturbation theory sets the accuracy bound of PTL-PINNs, we systematically evaluate its practical applicability. This work connects long-standing perturbation methods with PINNs, demonstrating how perturbation theory can guide foundational models to solve nonlinear systems with speeds comparable to those of classical solvers. --- ### Mitigating Cultural Bias in LLMs via Multi-Agent Cultural Debate Authors: Qian Tan, Lei Jiang, Yuting Zeng, Shuoyang Ding, Xiaohua Xu Published: 2026-01-17 Categories: cs.LG arXiv: https://arxiv.org/abs/2601.12091v1 Large language models (LLMs) exhibit systematic Western-centric bias, yet whether prompting in non-Western languages (e.g., Chinese) can mitigate this remains understudied. Answering this question requires rigorous evaluation and effective mitigation, but existing approaches fall short on both fronts: evaluation methods force outputs into predefined cultural categories without a neutral option, while mitigation relies on expensive multi-cultural corpora or agent frameworks that use functional roles (e.g., Planner--Critique) lacking explicit cultural representation. To address these gaps, we introduce CEBiasBench, a Chinese--English bilingual benchmark, and Multi-Agent Vote (MAV), which enables explicit ``no bias'' judgments. Using this framework, we find that Chinese prompting merely shifts bias toward East Asian perspectives rather than eliminating it. To mitigate such persistent bias, we propose Multi-Agent Cultural Debate (MACD), a training-free framework that assigns agents distinct cultural personas and orchestrates deliberation via a "Seeking Common Ground while Reserving Differences" strategy. Experiments demonstrate that MACD achieves 57.6% average No Bias Rate evaluated by LLM-as-judge and 86.0% evaluated by MAV (vs. 47.6% and 69.0% baseline using GPT-4o as backbone) on CEBiasBench and generalizes to the Arabic CAMeL benchmark, confirming that explicit cultural representation in agent frameworks is essential for cross-cultural fairness. --- ### Detecting 3D Line Segments for 6DoF Pose Estimation with Limited Data Authors: Matej Mok, Lukáš Gajdošech, Michal Mesároš, Martin Madaras, Viktor Kocur Published: 2026-01-17 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12090v2 The task of 6DoF object pose estimation is one of the fundamental problems of 3D vision with many practical applications such as industrial automation. Traditional deep learning approaches for this task often require extensive training data or CAD models, limiting their application in real-world industrial settings where data is scarce and object instances vary. We propose a novel method for 6DoF pose estimation focused specifically on bins used in industrial settings. We exploit the cuboid geometry of bins by first detecting intermediate 3D line segments corresponding to their top edges. Our approach extends the 2D line segment detection network LeTR to operate on structured point cloud data. The detected 3D line segments are then processed using a simple geometric procedure to robustly determine the bin's 6DoF pose. To evaluate our method, we extend an existing dataset with a newly collected and annotated dataset, which we make publicly available. We show that incorporating synthetic training data significantly improves pose estimation accuracy on real scans. Moreover, we show that our method significantly outperforms current state-of-the-art 6DoF pose estimation methods in terms of the pose accuracy (3 cm translation error, 8.2$^\circ$ rotation error) while not requiring instance-specific CAD models during inference. --- ### Learning to Factorize and Adapt: A Versatile Approach Toward Universal Spatio-Temporal Foundation Models Authors: Siru Zhong, Junjie Qiu, Yangyu Wu, Yiqiu Liu, Yuanpeng He, Zhongwen Rao, Bin Yang, Chenjuan Guo, Hao Xu, Yuxuan Liang Published: 2026-01-17 Categories: cs.LG arXiv: https://arxiv.org/abs/2601.12083v1 Spatio-Temporal (ST) Foundation Models (STFMs) promise cross-dataset generalization, yet joint ST pretraining is computationally expensive and grapples with the heterogeneity of domain-specific spatial patterns. Substantially extending our preliminary conference version, we present FactoST-v2, an enhanced factorized framework redesigned for full weight transfer and arbitrary-length generalization. FactoST-v2 decouples universal temporal learning from domain-specific spatial adaptation. The first stage pretrains a minimalist encoder-only backbone using randomized sequence masking to capture invariant temporal dynamics, enabling probabilistic quantile prediction across variable horizons. The second stage employs a streamlined adapter to rapidly inject spatial awareness via meta adaptive learning and prompting. Comprehensive evaluations across diverse domains demonstrate that FactoST-v2 achieves state-of-the-art accuracy with linear efficiency - significantly outperforming existing foundation models in zero-shot and few-shot scenarios while rivaling domain-specific expert baselines. This factorized paradigm offers a practical, scalable path toward truly universal STFMs. Code is available at https://github.com/CityMind-Lab/FactoST. --- ### Conditional Random Fields for Interactive Refinement of Histopathological Predictions Authors: Tiffanie Godelaine, Maxime Zanella, Karim El Khoury, Saïd Mahmoudi, Benoît Macq, Christophe De Vleeschouwer Published: 2026-01-17 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2601.12082v1 Assisting pathologists in the analysis of histopathological images has high clinical value, as it supports cancer detection and staging. In this context, histology foundation models have recently emerged. Among them, Vision-Language Models (VLMs) provide strong yet imperfect zero-shot predictions. We propose to refine these predictions by adapting Conditional Random Fields (CRFs) to histopathological applications, requiring no additional model training. We present HistoCRF, a CRF-based framework, with a novel definition of the pairwise potential that promotes label diversity and leverages expert annotations. We consider three experiments: without annotations, with expert annotations, and with iterative human-in-the-loop annotations that progressively correct misclassified patches. Experiments on five patch-level classification datasets covering different organs and diseases demonstrate average accuracy gains of 16.0% without annotations and 27.5% with only 100 annotations, compared to zero-shot predictions. Moreover, integrating a human in the loop reaches a further gain of 32.6% with the same number of annotations. The code will be made available on https://github.com/tgodelaine/HistoCRF. --- ### Toward Real-World High-Precision Image Matting and Segmentation Authors: Haipeng Zhou, Zhaohu Xing, Hongqiu Wang, Jun Ma, Ping Li, Lei Zhu Published: 2026-01-17 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12080v1 High-precision scene parsing tasks, including image matting and dichotomous segmentation, aim to accurately predict masks with extremely fine details (such as hair). Most existing methods focus on salient, single foreground objects. While interactive methods allow for target adjustment, their class-agnostic design restricts generalization across different categories. Furthermore, the scarcity of high-quality annotation has led to a reliance on inharmonious synthetic data, resulting in poor generalization to real-world scenarios. To this end, we propose a Foreground Consistent Learning model, dubbed as FCLM, to address the aforementioned issues. Specifically, we first introduce a Depth-Aware Distillation strategy where we transfer the depth-related knowledge for better foreground representation. Considering the data dilemma, we term the processing of synthetic data as domain adaptation problem where we propose a domain-invariant learning strategy to focus on foreground learning. To support interactive prediction, we contribute an Object-Oriented Decoder that can receive both visual and language prompts to predict the referring target. Experimental results show that our method quantitatively and qualitatively outperforms SOTA methods. --- ### EmoLat: Text-driven Image Sentiment Transfer via Emotion Latent Space Authors: Jing Zhang, Bingjie Fan, Jixiang Zhu, Zhe Wang Published: 2026-01-17 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12079v1 We propose EmoLat, a novel emotion latent space that enables fine-grained, text-driven image sentiment transfer by modeling cross-modal correlations between textual semantics and visual emotion features. Within EmoLat, an emotion semantic graph is constructed to capture the relational structure among emotions, objects, and visual attributes. To enhance the discriminability and transferability of emotion representations, we employ adversarial regularization, aligning the latent emotion distributions across modalities. Building upon EmoLat, a cross-modal sentiment transfer framework is proposed to manipulate image sentiment via joint embedding of text and EmoLat features. The network is optimized using a multi-objective loss incorporating semantic consistency, emotion alignment, and adversarial regularization. To support effective modeling, we construct EmoSpace Set, a large-scale benchmark dataset comprising images with dense annotations on emotions, object semantics, and visual attributes. Extensive experiments on EmoSpace Set demonstrate that our approach significantly outperforms existing state-of-the-art methods in both quantitative metrics and qualitative transfer fidelity, establishing a new paradigm for controllable image sentiment editing guided by textual input. The EmoSpace Set and all the code are available at http://github.com/JingVIPLab/EmoLat. --- ### Optimizing User Profiles via Contextual Bandits for Retrieval-Augmented LLM Personalization Authors: Linfeng Du, Ye Yuan, Zichen Zhao, Fuyuan Lyu, Emiliano Penaloza, Xiuying Chen, Zipeng Sun, Jikun Kang, Laurent Charlin, Xue Liu, Haolun Wu Published: 2026-01-17 Categories: cs.CL, cs.IR arXiv: https://arxiv.org/abs/2601.12078v2 Large language models (LLMs) excel at general-purpose tasks, yet adapting their responses to individual users remains challenging. Retrieval augmentation provides a lightweight alternative to fine-tuning by conditioning LLMs on user history records, and existing approaches typically select these records based on semantic relevance. We argue that relevance serves as an unreliable proxy for utility: a record may be semantically similar to a query yet fail to improve generation quality or even degrade it due to redundancy or conflicting information. To bridge this gap, we propose PURPLE, a contextual bandit framework that oPtimizes UseR Profiles for LLM pErsonalization. In contrast to a greedy selection of the most relevant records, PURPLE treats profile construction as an order-sensitive generation process and utilizes a Plackett-Luce ranking model to capture complex inter-record dependencies. By training with semantically rich feedback provided by the likelihood of the reference response, our method aligns retrieval directly with generation quality. Extensive experiments on nine personalization tasks demonstrate that PURPLE consistently outperforms strong heuristic and retrieval-augmented baselines in both effectiveness and efficiency, establishing a principled and scalable solution for optimizing user profiles. --- ### CroBIM-V: Memory-Quality Controlled Remote Sensing Referring Video Object Segmentation Authors: H. Jiang, Y. Sun, Z. Dong, T. Liu, Y. Gu Published: 2026-01-17 Categories: cs.CV, cs.CL arXiv: https://arxiv.org/abs/2601.12076v1 Remote sensing video referring object segmentation (RS-RVOS) is challenged by weak target saliency and severe visual information truncation in dynamic scenes, making it extremely difficult to maintain discriminative target representations during segmentation. Moreover, progress in this field is hindered by the absence of large-scale dedicated benchmarks, while existing models are often affected by biased initial memory construction that impairs accurate instance localization in complex scenarios, as well as indiscriminate memory accumulation that encodes noise from occlusions or misclassifications, leading to persistent error propagation. This paper advances RS-RVOS research through dual contributions in data and methodology. First, we construct RS-RVOS Bench, the first large-scale benchmark comprising 111 video sequences, about 25,000 frames, and 213,000 temporal referring annotations. Unlike common RVOS benchmarks where many expressions are written with access to the full video context, our dataset adopts a strict causality-aware annotation strategy in which linguistic references are generated solely from the target state in the initial frame. Second, we propose a memory-quality-aware online referring segmentation framework, termed Memory Quality Control with Segment Anything Model (MQC-SAM). MQC-SAM introduces a temporal motion consistency module for initial memory calibration, leveraging short-term motion trajectory priors to correct structural deviations and establish accurate memory anchoring. Furthermore, it incorporates a decoupled attention-based memory integration mechanism with dynamic quality assessment, selectively updating high-confidence semantic features while filtering unreliable information, thereby effectively preventing error accumulation and propagation. Extensive experiments on RS-RVOS Bench demonstrate that MQC-SAM achieves state-of-the-art performance. --- ### To Copy or Not to Copy: Copying Is Easier to Induce Than Recall Authors: Mehrdad Farahani, Franziska Penzkofer, Richard Johansson Published: 2026-01-17 Categories: cs.CL arXiv: https://arxiv.org/abs/2601.12075v2 Language models used in retrieval-augmented settings must arbitrate between parametric knowledge stored in their weights and contextual information in the prompt. This work presents a mechanistic study of that choice by extracting an \emph{arbitration vector} from model activations on a curated dataset designed to disentangle (i) irrelevant contexts that elicit parametric recall and (ii) relevant but false contexts that elicit copying. The vector is computed as the residual-stream centroid difference between these regimes across 27 relations, and is injected as an additive intervention at selected layers and token spans to steer behavior in two directions: Copy$\rightarrow$Recall (suppressing context use) and Recall$\rightarrow$Copy (inducing the model to copy any token from the context). Experiments on three architectures (decoder-only and encoder/decoder) and two open-domain QA benchmarks show consistent behavior shifts under moderate scaling while monitoring accuracy and fluency. Mechanistic analyses of attention routing, MLP contributions, and layer-wise probability trajectories reveal an asymmetry: inducing copying is an easy ``reactivation'' process that can be triggered at different locations in the input, while restoring recall is a ``suppression'' process that is more fragile and strongly tied to object-token interventions. --- ### Bridging the Gap in Bangla Healthcare: Machine Learning Based Disease Prediction Using a Symptoms-Disease Dataset Authors: Rowzatul Zannat, Abdullah Al Shafi, Abdul Muntakim Published: 2026-01-17 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2601.12068v1 Increased access to reliable health information is essential for non-English-speaking populations, yet resources in Bangla for disease prediction remain limited. This study addresses this gap by developing a comprehensive Bangla symptoms-disease dataset containing 758 unique symptom-disease relationships spanning 85 diseases. To ensure transparency and reproducibility, we also make our dataset publicly available. The dataset enables the prediction of diseases based on Bangla symptom inputs, supporting healthcare accessibility for Bengali-speaking populations. Using this dataset, we evaluated multiple machine learning models to predict diseases based on symptoms provided in Bangla and analyzed their performance on our dataset. Both soft and hard voting ensemble approaches combining top-performing models achieved 98\% accuracy, demonstrating superior robustness and generalization. Our work establishes a foundational resource for disease prediction in Bangla, paving the way for future advancements in localized health informatics and diagnostic tools. This contribution aims to enhance equitable access to health information for Bangla-speaking communities, particularly for early disease detection and healthcare interventions. --- ### ARMARecon: An ARMA Convolutional Filter based Graph Neural Network for Neurodegenerative Dementias Classification Authors: VSS Tejaswi Abburi, Ananya Singhal, Saurabh J. Shigwan, Nitin Kumar Published: 2026-01-17 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12067v1 Early detection of neurodegenerative diseases such as Alzheimer's Disease (AD) and Frontotemporal Dementia (FTD) is essential for reducing the risk of progression to severe disease stages. As AD and FTD propagate along white-matter regions in a global, graph-dependent manner, graph-based neural networks are well suited to capture these patterns. Hence, we introduce ARMARecon, a unified graph learning framework that integrates Autoregressive Moving Average (ARMA) graph filtering with a reconstruction-driven objective to enhance feature representation and improve classification accuracy. ARMARecon effectively models both local and global connectivity by leveraging 20-bin Fractional Anisotropy (FA) histogram features extracted from white-matter regions, while mitigating over-smoothing. Overall, ARMARecon achieves superior performance compared to state-of-the-art methods on the multi-site dMRI datasets ADNI and NIFD. --- ### Learning Stochastic Bridges for Video Object Removal via Video-to-Video Translation Authors: Zijie Lou, Xiangwei Feng, Jiaxin Wang, Jiangtao Yao, Fei Che, Tianbao Liu, Chengjing Wu, Xiaochao Qu, Luoqi Liu, Ting Liu Published: 2026-01-17 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12066v4 Existing video object removal methods predominantly rely on diffusion models following a noise-to-data paradigm, where generation starts from uninformative Gaussian noise. This approach discards the rich structural and contextual priors present in the original input video. Consequently, such methods often lack sufficient guidance, leading to incomplete object erasure or the synthesis of implausible content that conflicts with the scene's physical logic. In this paper, we reformulate video object removal as a video-to-video translation task via a stochastic bridge model. Unlike noise-initialized methods, our framework establishes a direct stochastic path from the source video (with objects) to the target video (objects removed). This bridge formulation effectively leverages the input video as a strong structural prior, guiding the model to perform precise removal while ensuring that the filled regions are logically consistent with the surrounding environment. To address the trade-off where strong bridge priors hinder the removal of large objects, we propose a novel adaptive mask modulation strategy. This mechanism dynamically modulates input embeddings based on mask characteristics, balancing background fidelity with generative flexibility. Extensive experiments demonstrate that our approach significantly outperforms existing methods in both visual quality and temporal consistency. The project page is https://bridgeremoval.github.io/. --- ### Learning Language-Driven Sequence-Level Modal-Invariant Representations for Video-Based Visible-Infrared Person Re-Identification Authors: Xiaomei Yang, Antai Liu, Xizhan Gao, Fa Zhu, Sijie Niu, Giancarlo Fortino Published: 2026-01-17 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12062v2 The core of video-based visible-infrared person re-identification (VVI-ReID) lies in learning sequence-level modal-invariant representations across different modalities. Recent research tends to use modality-shared language prompts generated by CLIP to guide the learning of modal-invariant representations. Despite achieving optimal performance, such methods still face limitations in efficient spatial-temporal modeling, sufficient cross-modal interaction, and explicit modality-level loss guidance. To address these issues, we propose the language-driven sequence-level modal-invariant representation learning (LSMRL) method, which includes spatial-temporal feature learning (STFL) module, semantic diffusion (SD) module and cross-modal interaction (CMI) module. To enable parameter- and computation-efficient spatial-temporal modeling, the STFL module is built upon CLIP with minimal modifications. To achieve sufficient cross-modal interaction and enhance the learning of modal-invariant features, the SD module is proposed to diffuse modality-shared language prompts into visible and infrared features to establish preliminary modal consistency. The CMI module is further developed to leverage bidirectional cross-modal self-attention to eliminate residual modality gaps and refine modal-invariant representations. To explicitly enhance the learning of modal-invariant representations, two modality-level losses are introduced to improve the features' discriminative ability and their generalization to unseen categories. Extensive experiments on large-scale VVI-ReID datasets demonstrate the superiority of LSMRL over AOTA methods. --- ### Codebook-Injected Dialogue Segmentation for Multi-Utterance Constructs Annotation: LLM-Assisted and Gold-Label-Free Evaluation Authors: Jinsook Lee, Kirk Vanacore, Zhuqian Zhou, Bakhtawar Ahtisham, Jeanine Grutter, Rene F. Kizilcec Published: 2026-01-17 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2601.12061v2 Dialogue Act (DA) annotation typically treats communicative or pedagogical intent as localized to individual utterances or turns. This leads annotators to agree on the underlying action while disagreeing on segment boundaries, reducing apparent reliability. We propose codebook-injected segmentation, which conditions boundary decisions on downstream annotation criteria, and evaluate LLM-based segmenters against standard and retrieval-augmented baselines. To assess these without gold labels, we introduce evaluation metrics for span consistency, distinctiveness, and human-AI distributional agreement. We found DA-awareness produces segments that are internally more consistent than text-only baselines. While LLMs excel at creating construct-consistent spans, coherence-based baselines remain superior at detecting global shifts in dialogue flow. Across two datasets, no single segmenter dominates. Improvements in within-segment coherence frequently trade off against boundary distinctiveness and human-AI distributional agreement. These results highlight segmentation as a consequential design choice that should be optimized for downstream objectives rather than a single performance score. --- ### Automating Parameter Selection in Deep Image Prior for Fluorescence Microscopy Image Denoising via Similarity-Based Parameter Transfer Authors: Lina Meyer, Felix Wissel, Tobias Knopp, Susanne Pfefferle, Ralf Fliegert, Maximilian Sandmann, Liana Uebler, Franziska Möckl, Björn-Philipp Diercks, David Lohr, René Werner Published: 2026-01-17 Categories: cs.CV, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2601.12055v1 Unsupervised deep image prior (DIP) addresses shortcomings of training data requirements and limited generalization associated with supervised deep learning. The performance of DIP depends on the network architecture and the stopping point of its iterative process. Optimizing these parameters for a new image requires time, restricting DIP application in domains where many images need to be processed. Focusing on fluorescence microscopy data, we hypothesize that similar images share comparable optimal parameter configurations for DIP-based denoising, potentially enabling optimization-free DIP for fluorescence microscopy. We generated a calibration (n=110) and validation set (n=55) of semantically different images from an open-source dataset for a network architecture search targeted towards ideal U-net architectures and stopping points. The calibration set represented our transfer basis. The validation set enabled the assessment of which image similarity criterion yields the best results. We then implemented AUTO-DIP, a pipeline for automatic parameter transfer, and compared it to the originally published DIP configuration (baseline) and a state-of-the-art image-specific variational denoising approach. We show that a parameter transfer from the calibration dataset to a test image based on only image metadata similarity (e.g., microscope type, imaged specimen) leads to similar and better performance than a transfer based on quantitative image similarity measures. AUTO-DIP outperforms the baseline DIP (DIP with original DIP parameters) as well as the variational denoising approaches for several open-source test datasets of varying complexity, particularly for very noisy inputs. Applications to locally acquired fluorescence microscopy images further proved superiority of AUTO-DIP. --- ### A New Strategy for Artificial Intelligence: Training Foundation Models Directly on Human Brain Data Authors: Maël Donoso Published: 2026-01-17 Categories: q-bio.NC, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2601.12053v2 While foundation models have achieved remarkable results across a diversity of domains, they still rely on human-generated data, such as text, as a fundamental source of knowledge. However, this data is ultimately the product of human brains, the filtered projection of a deeper neural complexity. In this paper, we explore a new strategy for artificial intelligence: moving beyond surface-level statistical regularities by training foundation models directly on human brain data. We hypothesize that neuroimaging data could open a window into elements of human cognition that are not accessible through observable actions, and argue that this additional knowledge could be used, alongside classical training data, to overcome some of the current limitations of foundation models. While previous research has demonstrated the possibility to train classical machine learning, deep learning, or reinforcement learning models on neural patterns, this path remains largely unexplored for high-level cognitive functions. Here, we classify the current limitations of foundation models, as well as the promising brain regions and cognitive processes that could be leveraged to address them, along four levels: perception, valuation, execution, and integration. Then, we propose two general methods that could be implemented to prioritize the use of limited neuroimaging data for strategically chosen, high-value steps in foundation model training: reinforcement learning from human brain (RLHB) and chain of thought from human brain (CoTHB). We also discuss the potential implications for agents, artificial general intelligence, and artificial superintelligence, as well as the ethical, social, and technical challenges and opportunities. --- ### Task-Driven Prompt Learning: A Joint Framework for Multi-modal Cloud Removal and Segmentation Authors: Zaiyan Zhang, Jie Li, Shaowei Shi, Qiangqiang Yuan Published: 2026-01-17 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12052v2 Optical remote sensing imagery is indispensable for Earth observation, yet persistent cloud occlusion limits its downstream utility. Most cloud removal (CR) methods are optimized for low-level fidelity and can over-smooth textures and boundaries that are critical for analysis-ready data (ARD), leading to a mismatch between visually plausible restoration and semantic utility. To bridge this gap, we propose TDP-CR, a task-driven multimodal framework that jointly performs cloud removal and land-cover segmentation. Central to our approach is a Prompt-Guided Fusion (PGF) mechanism, which utilizes a learnable degradation prompt to encode cloud thickness and spatial uncertainty. By combining global channel context with local prompt-conditioned spatial bias, PGF adaptively integrates Synthetic Aperture Radar (SAR) information only where optical data is corrupted. We further introduce a parameter-efficient two-phase training strategy that decouples reconstruction and semantic representation learning. Experiments on the LuojiaSET-OSFCR dataset demonstrate the superiority of our framework: TDP-CR surpasses heavy state-of-the-art baselines by 0.18 dB in PSNR while using only 15\% of the parameters, and achieves a 1.4\% improvement in mIoU consistently against multi-task competitors, effectively delivering analysis-ready data. --- ### A Unified Masked Jigsaw Puzzle Framework for Vision and Language Models Authors: Weixin Ye, Wei Wang, Yahui Liu, Yue Song, Bin Ren, Wei Bi, Rita Cucchiara, Nicu Sebe Published: 2026-01-17 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12051v1 In federated learning, Transformer, as a popular architecture, faces critical challenges in defending against gradient attacks and improving model performance in both Computer Vision (CV) and Natural Language Processing (NLP) tasks. It has been revealed that the gradient of Position Embeddings (PEs) in Transformer contains sufficient information, which can be used to reconstruct the input data. To mitigate this issue, we introduce a Masked Jigsaw Puzzle (MJP) framework. MJP starts with random token shuffling to break the token order, and then a learnable \textit{unknown (unk)} position embedding is used to mask out the PEs of the shuffled tokens. In this manner, the local spatial information which is encoded in the position embeddings is disrupted, and the models are forced to learn feature representations that are less reliant on the local spatial information. Notably, with the careful use of MJP, we can not only improve models' robustness against gradient attacks, but also boost their performance in both vision and text application scenarios, such as classification for images (\textit{e.g.,} ImageNet-1K) and sentiment analysis for text (\textit{e.g.,} Yelp and Amazon). Experimental results suggest that MJP is a unified framework for different Transformer-based models in both vision and language tasks. Code is publicly available via https://github.com/ywxsuperstar/transformerattack --- ### \textit{FocaLogic}: Logic-Based Interpretation of Visual Model Decisions Authors: Chenchen Zhao, Muxi Chen, Qiang Xu Published: 2026-01-17 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2601.12049v1 Interpretability of modern visual models is crucial, particularly in high-stakes applications. However, existing interpretability methods typically suffer from either reliance on white-box model access or insufficient quantitative rigor. To address these limitations, we introduce FocaLogic, a novel model-agnostic framework designed to interpret and quantify visual model decision-making through logic-based representations. FocaLogic identifies minimal interpretable subsets of visual regions-termed visual focuses-that decisively influence model predictions. It translates these visual focuses into precise and compact logical expressions, enabling transparent and structured interpretations. Additionally, we propose a suite of quantitative metrics, including focus precision, recall, and divergence, to objectively evaluate model behavior across diverse scenarios. Empirical analyses demonstrate FocaLogic's capability to uncover critical insights such as training-induced concentration, increasing focus accuracy through generalization, and anomalous focuses under biases and adversarial attacks. Overall, FocaLogic provides a systematic, scalable, and quantitative solution for interpreting visual models. --- ### Less Is More -- Until It Breaks: Security Pitfalls of Vision Token Compression in Large Vision-Language Models Authors: Xiaomei Zhang, Zhaoxi Zhang, Leo Yu Zhang, Yanjun Zhang, Guanhong Tao, Shirui Pan Published: 2026-01-17 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2601.12042v1 Visual token compression is widely adopted to improve the inference efficiency of Large Vision-Language Models (LVLMs), enabling their deployment in latency-sensitive and resource-constrained scenarios. However, existing work has mainly focused on efficiency and performance, while the security implications of visual token compression remain largely unexplored. In this work, we first reveal that visual token compression substantially degrades the robustness of LVLMs: models that are robust under uncompressed inference become highly vulnerable once compression is enabled. These vulnerabilities are state-specific; failure modes emerge only in the compressed setting and completely disappear when compression is disabled, making them particularly hidden and difficult to diagnose. By analyzing the key stages of the compression process, we identify instability in token importance ranking as the primary cause of this robustness degradation. Small and imperceptible perturbations can significantly alter token rankings, leading the compression mechanism to mistakenly discard task-critical information and ultimately causing model failure. Motivated by this observation, we propose a Compression-Aware Attack to systematically study and exploit this vulnerability. CAA directly targets the token selection mechanism and induces failures exclusively under compressed inference. We further extend this approach to more realistic black-box settings and introduce Transfer CAA, where neither the target model nor the compression configuration is accessible. We further evaluate potential defenses and find that they provide only limited protection. Extensive experiments across models, datasets, and compression methods show that visual token compression significantly undermines robustness, revealing a previously overlooked efficiency-security trade-off. --- ### Partial Reasoning in Language Models: Search and Refinement Guided by Uncertainty Authors: Murilo da Luz, Bruno Brandão, Luana Martins, Gustavo Oliveira, Bryan de Oliveira, Luckeciano Melo, Telma Soares Published: 2026-01-17 Categories: cs.AI arXiv: https://arxiv.org/abs/2601.12040v1 The use of Large Language Models (LLMs) for reasoning and planning tasks has drawn increasing attention in Artificial Intelligence research. Despite their remarkable progress, these models still exhibit limitations in multi-step inference scenarios, particularly in mathematical and logical reasoning. We introduce PREGU (Partial Reasoning Guided by Uncertainty). PREGU monitors the entropy of the output distribution during autoregressive generation and halts the process whenever entropy exceeds a defined threshold, signaling uncertainty. From that point, a localized search is performed in the latent space to refine the partial reasoning and select the most coherent answer, using the Soft Reasoning method. Experiments conducted with LLaMA-3-8B, Mistral-7B, and Qwen2-7B across four reasoning benchmarks (GSM8K, GSM-Hard, SVAMP, and StrategyQA) showed performance greater than or similar to Soft Reasoning, indicating that entropy can serve as an effective signal to trigger selective refinement during reasoning. --- ### Nonlinear Dynamic Factor Analysis With a Transformer Network Authors: Oliver Snellman Published: 2026-01-17 Categories: econ.EM, cs.LG arXiv: https://arxiv.org/abs/2601.12039v1 The paper develops a Transformer architecture for estimating dynamic factors from multivariate time series data under flexible identification assumptions. Performance on small datasets is improved substantially by using a conventional factor model as prior information via a regularization term in the training objective. The results are interpreted with Attention matrices that quantify the relative importance of variables and their lags for the factor estimate. Time variation in Attention patterns can help detect regime switches and evaluate narratives. Monte Carlo experiments suggest that the Transformer is more accurate than the linear factor model, when the data deviate from linear-Gaussian assumptions. An empirical application uses the Transformer to construct a coincident index of U.S. real economic activity. --- ### Subargument Argumentation Frameworks: Separating Direct Conflict from Structural Dependency Authors: Beishui Liao Published: 2026-01-17 Categories: cs.AI arXiv: https://arxiv.org/abs/2601.12038v3 Dung's abstract argumentation frameworks model acceptability solely in terms of an attack relation, thereby conflating two conceptually distinct aspects of argumentative reasoning: direct conflict between arguments and the structural dependencies that arise from their internal composition. While this abstraction preserves extension-based semantics, it obscures how justification is grounded in subarguments and how defeats propagate through argument structure. We introduce Subargument Argumentation Frameworks (SAFs), an abstract framework in which direct attack and subargumenthood are represented as independent primitive relations. This separation makes structural dependency explicit at the representational level while leaving its semantic impact to be determined by structure-sensitive notions of defence, admissibility, and complete semantics defined within the framework. We show that projecting SAFs onto attack-only frameworks yields extension-equivalent Dung frameworks under all standard semantics, yet the projection irreversibly loses information about justificatory grounding and structural propagation. SAFs therefore provide strictly greater representational expressiveness while remaining semantically compatible with Dung's theory, thereby offering a principled basis for structure-sensitive accounts of defence, justification, and explanation in abstract argumentation. --- ### Multimodal Feedback for Handheld Tool Guidance: Combining Wrist-Based Haptics with Augmented Reality Authors: Yue Yang, Christoph Leuze, Brian Hargreaves, Bruce Daniel, Fred M Baik Published: 2026-01-17 Categories: cs.HC, cs.CV, eess.SY arXiv: https://arxiv.org/abs/2601.12037v1 We investigate how vibrotactile wrist feedback can enhance spatial guidance for handheld tool movement in optical see-through augmented reality (AR). While AR overlays are widely used to support surgical tasks, visual occlusion, lighting conditions, and interface ambiguity can compromise precision and confidence. To address these challenges, we designed a multimodal system combining AR visuals with a custom wrist-worn haptic device delivering directional and state-based cues. A formative study with experienced surgeons and residents identified key tool maneuvers and preferences for reference mappings, guiding our cue design. In a cue identification experiment (N=21), participants accurately recognized five vibration patterns under visual load, with higher recognition for full-actuator states than spatial direction cues. In a guidance task (N=27), participants using both AR and haptics achieved significantly higher spatial precision (5.8 mm) and usability (SUS = 88.1) than those using either modality alone, despite having modest increases in task time. Participants reported that haptic cues provided reassuring confirmation and reduced cognitive effort during alignment. Our results highlight the promise of integrating wrist-based haptics into AR systems for high-precision, visually complex tasks such as surgical guidance. We discuss design implications for multimodal interfaces supporting confident, efficient tool manipulation. --- ### Don't Start Over: A Cost-Effective Framework for Migrating Personalized Prompts Between LLMs Authors: Ziyi Zhao, Chongming Gao, Yang Zhang, Haoyan Liu, Weinan Gan, Huifeng Guo, Yong Liu, Fuli Feng Published: 2026-01-17 Categories: cs.CL, cs.IR arXiv: https://arxiv.org/abs/2601.12034v1 Personalization in Large Language Models (LLMs) often relies on user-specific soft prompts. However, these prompts become obsolete when the foundation model is upgraded, necessitating costly, full-scale retraining. To overcome this limitation, we propose the Prompt-level User Migration Adapter (PUMA), a lightweight framework to efficiently migrate personalized prompts across incompatible models. PUMA utilizes a parameter-efficient adapter to bridge the semantic gap, combined with a group-based user selection strategy to significantly reduce training costs. Experiments on three large-scale datasets show our method matches or even surpasses the performance of retraining from scratch, reducing computational cost by up to 98%. The framework demonstrates strong generalization across diverse model architectures and robustness in advanced scenarios like chained and aggregated migrations, offering a practical path for the sustainable evolution of personalized AI by decoupling user assets from the underlying models. --- ### Preserving Fairness and Safety in Quantized LLMs Through Critical Weight Protection Authors: Muhammad Alif Al Hakim, Alfan Farizki Wicaksono, Fajri Koto Published: 2026-01-17 Categories: cs.CL arXiv: https://arxiv.org/abs/2601.12033v2 Quantization is widely adopted to reduce the computational cost of large language models (LLMs); however, its implications for fairness and safety, particularly in dynamic quantization and multilingual contexts, remain underexplored. In this work, we conduct a systematic study of how static and dynamic quantization methods impact fairness and safety across benchmarks measuring intrinsic and extrinsic bias and safety alignment. For fairness, we evaluate English, French, Dutch, Spanish, and Turkish; for safety, we focus on English, Korean, and Arabic. Our findings reveal that quantization consistently degrades fairness and safety, with dynamic methods demonstrating greater stability than static ones. Moreover, fairness degradation varies across languages, while safety deterioration is especially pronounced in non-English settings. To address these risks, we introduce Critical Weight Protection, a novel technique that identifies and preserves fairness- and safety-critical weights during quantization. This approach effectively mitigates bias and safety deterioration without costly retraining or alignment, maintaining trustworthiness while retaining efficiency. --- ### Speaking to Silicon: Neural Communication with Bitcoin Mining ASICs Authors: Francisco Angulo de Lafuente, Vladimir Veselov, Richard Goodman Published: 2026-01-17 Categories: cs.NE, cs.AR, cs.CR, cs.LG arXiv: https://arxiv.org/abs/2601.12032v1 This definitive research memoria presents a comprehensive, mathematically verified paradigm for neural communication with Bitcoin mining Application-Specific Integrated Circuits (ASICs), integrating five complementary frameworks: thermodynamic reservoir computing, hierarchical number system theory, algorithmic analysis, network latency optimization, and machine-checked mathematical formalization. We establish that obsolete cryptocurrency mining hardware exhibits emergent computational properties enabling bidirectional information exchange between AI systems and silicon substrates. The research program demonstrates: (1) reservoir computing with NARMA-10 Normalized Root Mean Square Error (NRMSE) of 0.8661; (2) the Thermodynamic Probability Filter (TPF) achieving 92.19% theoretical energy reduction; (3) the Virtual Block Manager achieving +25% effective hashrate; and (4) hardware universality across multiple ASIC families including Antminer S9, Lucky Miner LV06, and Goldshell LB-Box. A significant contribution is the machine-checked mathematical formalization using Lean 4 and Mathlib, providing unambiguous definitions, machine-verified theorems, and reviewer-proof claims. Key theorems proven include: independence implies zero leakage, predictor beats baseline implies non-independence (the logical core of TPF), energy savings theoretical maximum, and Physical Unclonable Function (PUF) distinguishability witnesses. Vladimir Veselov's hierarchical number system theory explains why early-round information contains predictive power. This work establishes a new paradigm: treating ASICs not as passive computational substrates but as active conversational partners whose thermodynamic state encodes exploitable computational information. --- ### ARC: Active and Reflection-driven Context Management for Long-Horizon Information Seeking Agents Authors: Yilun Yao, Shan Huang, Elsie Dai, Zhewen Tan, Zhenyu Duan, Shousheng Jia, Yanbing Jiang, Tong Yang Published: 2026-01-17 Categories: cs.AI arXiv: https://arxiv.org/abs/2601.12030v1 Large language models are increasingly deployed as research agents for deep search and long-horizon information seeking, yet their performance often degrades as interaction histories grow. This degradation, known as context rot, reflects a failure to maintain coherent and task-relevant internal states over extended reasoning horizons. Existing approaches primarily manage context through raw accumulation or passive summarization, treating it as a static artifact and allowing early errors or misplaced emphasis to persist. Motivated by this perspective, we propose ARC, which is the first framework to systematically formulate context management as an active, reflection-driven process that treats context as a dynamic internal reasoning state during execution. ARC operationalizes this view through reflection-driven monitoring and revision, allowing agents to actively reorganize their working context when misalignment or degradation is detected. Experiments on challenging long-horizon information-seeking benchmarks show that ARC consistently outperforms passive context compression methods, achieving up to an 11% absolute improvement in accuracy on BrowseComp-ZH with Qwen2.5-32B-Instruct. --- ### Beyond Sentiment: A Multi-Agent Pipeline for Actionable Business Advice from Reviews Authors: Kartikey Singh Bhandari, Tanish Jain, Archit Agrawal, Dhruv Kumar, Praveen Kumar, Pratik Narang Published: 2026-01-17 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2601.12024v2 Customer reviews contain valuable signals about service quality, but converting large-scale review corpora into actionable business recommendations remains difficult. Standard sentiment/aspect analysis is largely descriptive, while direct prompting of large language models (LLMs) often yields generic and repetitive advice that is weakly grounded in user feedback. We propose a hierarchical decision-support pipeline that explicitly separates signal compression, problem abstraction, candidate generation, objective-based evaluation, and cost-aware routing into different agents. This architectural decomposition produces auditable intermediate artifacts and enables controllable trade-offs between advice quality and token budget. Experiments on Yelp reviews from three service domains show consistent improvements over single-pass LLM baselines across multiple advice quality dimensions, including actionability, relevance, and non-redundancy. A human evaluation further indicates that users generally prefer our system's recommendations. These results highlight the value of structured agentic decomposition for scalable, cost-aware business decision support. --- ### A Kernel Approach for Semi-implicit Variational Inference Authors: Longlin Yu, Ziheng Cheng, Shiyue Zhang, Cheng Zhang Published: 2026-01-17 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2601.12023v1 Semi-implicit variational inference (SIVI) enhances the expressiveness of variational families through hierarchical semi-implicit distributions, but the intractability of their densities makes standard ELBO-based optimization biased. Recent score-matching approaches to SIVI (SIVI-SM) address this issue via a minimax formulation, at the expense of an additional lower-level optimization problem. In this paper, we propose kernel semi-implicit variational inference (KSIVI), a principled and tractable alternative that eliminates the lower-level optimization by leveraging kernel methods. We show that when optimizing over a reproducing kernel Hilbert space, the lower-level problem admits an explicit solution, reducing the objective to the kernel Stein discrepancy (KSD). Exploiting the hierarchical structure of semi-implicit distributions, the resulting KSD objective can be efficiently optimized using stochastic gradient methods. We establish optimization guarantees via variance bounds on Monte Carlo gradient estimators and derive statistical generalization bounds of order $\tilde{\mathcal{O}}(1/\sqrt{n})$. We further introduce a multi-layer hierarchical extension that improves expressiveness while preserving tractability. Empirical results on synthetic and real-world Bayesian inference tasks demonstrate the effectiveness of KSIVI. --- ### DIAMOND-SSS: Diffusion-Augmented Multi-View Optimization for Data-efficient SubSurface Scattering Authors: Guillermo Figueroa-Araneda, Iris Diana Jimenez, Florian Hofherr, Manny Ko, Hector Andrade-Loarca, Daniel Cremers Published: 2026-01-17 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12020v1 Subsurface scattering (SSS) gives translucent materials -- such as wax, jade, marble, and skin -- their characteristic soft shadows, color bleeding, and diffuse glow. Modeling these effects in neural rendering remains challenging due to complex light transport and the need for densely captured multi-view, multi-light datasets (often more than 100 views and 112 OLATs). We present DIAMOND-SSS, a data-efficient framework for high-fidelity translucent reconstruction from extremely sparse supervision -- even as few as ten images. We fine-tune diffusion models for novel-view synthesis and relighting, conditioned on estimated geometry and trained on less than 7 percent of the dataset, producing photorealistic augmentations that can replace up to 95 percent of missing captures. To stabilize reconstruction under sparse or synthetic supervision, we introduce illumination-independent geometric priors: a multi-view silhouette consistency loss and a multi-view depth consistency loss. Across all sparsity regimes, DIAMOND-SSS achieves state-of-the-art quality in relightable Gaussian rendering, reducing real capture requirements by up to 90 percent compared to SSS-3DGS. --- ### Acting Flatterers via LLMs Sycophancy: Combating Clickbait with LLMs Opposing-Stance Reasoning Authors: Chaowei Zhang, Xiansheng Luo, Zewei Zhang, Yi Zhu, Jipeng Qiang, Longwei Wang Published: 2026-01-17 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2601.12019v1 The widespread proliferation of online content has intensified concerns about clickbait, deceptive or exaggerated headlines designed to attract attention. While Large Language Models (LLMs) offer a promising avenue for addressing this issue, their effectiveness is often hindered by Sycophancy, a tendency to produce reasoning that matches users' beliefs over truthful ones, which deviates from instruction-following principles. Rather than treating sycophancy as a flaw to be eliminated, this work proposes a novel approach that initially harnesses this behavior to generate contrastive reasoning from opposing perspectives. Specifically, we design a Self-renewal Opposing-stance Reasoning Generation (SORG) framework that prompts LLMs to produce high-quality agree and disagree reasoning pairs for a given news title without requiring ground-truth labels. To utilize the generated reasoning, we develop a local Opposing Reasoning-based Clickbait Detection (ORCD) model that integrates three BERT encoders to represent the title and its associated reasoning. The model leverages contrastive learning, guided by soft labels derived from LLM-generated credibility scores, to enhance detection robustness. Experimental evaluations on three benchmark datasets demonstrate that our method consistently outperforms LLM prompting, fine-tuned smaller language models, and state-of-the-art clickbait detection baselines. --- ### SAR-Based Marine Oil Spill Detection Using the DeepSegFusion Architecture Authors: Pavan Kumar Yata, Pediredla Pradeep, Goli Himanish, Swathi M Published: 2026-01-17 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12015v1 Detection of oil spills from satellite images is essential for both environmental surveillance and maritime safety. Traditional threshold-based methods frequently encounter performance degradation due to very high false alarm rates caused by look-alike phenomena such as wind slicks and ship wakes. Here, a hybrid deep learning model, DeepSegFusion, is presented for oil spill segmentation in Synthetic Aperture Radar (SAR) images. The model uses SegNet and DeepLabV3+ integrated with an attention-based feature fusion mechanism to achieve better boundary precision as well as improved contextual understanding. Results obtained on SAR oil spill datasets, including ALOS PALSAR imagery, confirm that the proposed DeepSegFusion model achieves an accuracy of 94.85%, an Intersection over Union (IoU) of 0.5685, and a ROC-AUC score of 0.9330. The proposed method delivers more than three times fewer false detections compared to individual baseline models and traditional non-segmentation methods, achieving a reduction of 64.4%. These results indicate that DeepSegFusion is a stable model under various marine conditions and can therefore be used in near real-time oil spill monitoring scenarios. --- ### Are LLMs Ready for TOON? Benchmarking Structural Correctness-Sustainability Trade-offs in Novel Structured Output Formats Authors: Elio Masciari, Vincenzo Moscato, Enea Vincenzo Napolitano, Gian Marco Orlando, Marco Perillo, Diego Russo Published: 2026-01-17 Categories: cs.AI, cs.SE arXiv: https://arxiv.org/abs/2601.12014v1 Large Language Models (LLMs) are increasingly required to generate structured, machine-readable outputs for downstream systems. While recent benchmarks have focused on evaluating the structural correctness of such outputs, the environmental impact of inference for different output formats has largely been overlooked. In this paper, we argue that structured output formats should be assessed not only in terms of correctness, but also with respect to their environmental efficiency. To this end, we introduce a sustainability-aware evaluation framework for structured generation that measures token usage, generation time, and estimated carbon emissions. Within this framework, we propose the Environment-Aware Generation Correctness Score (GCS_env), a unified metric that integrates structural correctness with carbon-aware efficiency. Using this framework, we systematically benchmark the novel TOON format against established representations (JSON, XML, YAML) across multiple LLMs spanning different architectures and parameter scales. Our results reveal a consistent trade-off: TOON yields markedly more compact outputs and lower emissions, but lower structural correctness when models lack native support. We show that increased model capacity reduces this gap and that environment-aware scoring can shift format rankings depending on deployment priorities. highlighting the need for sustainability-inclusive benchmarking and provides empirical evidence that compact representations such as TOON can offer practical advantages in large-scale, carbon-conscious LLM deployments. --- ### Why Loss Re-weighting Works If You Stop Early: Training Dynamics of Unconstrained Features Authors: Yize Zhao, Christos Thrampoulidis Published: 2026-01-17 Categories: cs.LG arXiv: https://arxiv.org/abs/2601.12011v1 The application of loss reweighting in modern deep learning presents a nuanced picture. While it fails to alter the terminal learning phase in overparameterized deep neural networks (DNNs) trained on high-dimensional datasets, empirical evidence consistently shows it offers significant benefits early in training. To transparently demonstrate and analyze this phenomenon, we introduce a small-scale model (SSM). This model is specifically designed to abstract the inherent complexities of both the DNN architecture and the input data, while maintaining key information about the structure of imbalance within its spectral components. On the one hand, the SSM reveals how vanilla empirical risk minimization preferentially learns to distinguish majority classes over minorities early in training, consequently delaying minority learning. In stark contrast, reweighting restores balanced learning dynamics, enabling the simultaneous learning of features associated with both majorities and minorities. --- ### SMc2f: Robust Scenario Mining for Robotic Autonomy from Coarse to Fine Authors: Yifei Chen, Ross Greer Published: 2026-01-17 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.12010v1 The safety validation of autonomous robotic vehicles hinges on systematically testing their planning and control stacks against rare, safety-critical scenarios. Mining these long-tail events from massive real-world driving logs is therefore a critical step in the robotic development lifecycle. The goal of the Scenario Mining task is to retrieve useful information to enable targeted re-simulation, regression testing, and failure analysis of the robot's decision-making algorithms. RefAV, introduced by the Argoverse team, is an end-to-end framework that uses large language models (LLMs) to spatially and temporally localize scenarios described in natural language. However, this process performs retrieval on trajectory labels, ignoring the direct connection between natural language and raw RGB images, which runs counter to the intuition of video retrieval; it also depends on the quality of upstream 3D object detection and tracking. Further, inaccuracies in trajectory data lead to inaccuracies in downstream spatial and temporal localization. To address these issues, we propose Robust Scenario Mining for Robotic Autonomy from Coarse to Fine (SMc2f), a coarse-to-fine pipeline that employs vision-language models (VLMs) for coarse image-text filtering, builds a database of successful mining cases on top of RefAV and automatically retrieves exemplars to few-shot condition the LLM for more robust retrieval, and introduces text-trajectory contrastive learning to pull matched pairs together and push mismatched pairs apart in a shared embedding space, yielding a fine-grained matcher that refines the LLM's candidate trajectories. Experiments on public datasets demonstrate substantial gains in both retrieval quality and efficiency. --- ### Extreme Value Policy Optimization for Safe Reinforcement Learning Authors: Shiqing Gao, Yihang Zhou, Shuai Shao, Haoyu Luo, Yiheng Bing, Jiaxin Ding, Luoyi Fu, Xinbing Wang Published: 2026-01-17 Categories: cs.LG arXiv: https://arxiv.org/abs/2601.12008v1 Ensuring safety is a critical challenge in applying Reinforcement Learning (RL) to real-world scenarios. Constrained Reinforcement Learning (CRL) addresses this by maximizing returns under predefined constraints, typically formulated as the expected cumulative cost. However, expectation-based constraints overlook rare but high-impact extreme value events in the tail distribution, such as black swan incidents, which can lead to severe constraint violations. To address this issue, we propose the Extreme Value policy Optimization (EVO) algorithm, leveraging Extreme Value Theory (EVT) to model and exploit extreme reward and cost samples, reducing constraint violations. EVO introduces an extreme quantile optimization objective to explicitly capture extreme samples in the cost tail distribution. Additionally, we propose an extreme prioritization mechanism during replay, amplifying the learning signal from rare but high-impact extreme samples. Theoretically, we establish upper bounds on expected constraint violations during policy updates, guaranteeing strict constraint satisfaction at a zero-violation quantile level. Further, we demonstrate that EVO achieves a lower probability of constraint violations than expectation-based methods and exhibits lower variance than quantile regression methods. Extensive experiments show that EVO significantly reduces constraint violations during training while maintaining competitive policy performance compared to baselines. --- ### Robust Verification of Concurrent Stochastic Games Authors: Angel Y. He, David Parker Published: 2026-01-17 Categories: cs.LO, cs.AI, cs.GT, cs.MA, eess.SY arXiv: https://arxiv.org/abs/2601.12003v2 Autonomous systems often operate in multi-agent settings and need to make concurrent, strategic decisions, typically in uncertain environments. Verification and control problems for these systems can be tackled with concurrent stochastic games (CSGs), but this model requires transition probabilities to be precisely specified - an unrealistic requirement in many real-world settings. We introduce *robust CSGs* and their subclass *interval CSGs* (ICSGs), which capture epistemic uncertainty about transition probabilities in CSGs. We propose a novel framework for *robust* verification of these models under worst-case assumptions about transition uncertainty. Specifically, we develop the underlying theoretical foundations and efficient algorithms, for finite- and infinite-horizon objectives in both zero-sum and nonzero-sum settings, the latter based on (social-welfare optimal) Nash equilibria. We build an implementation in the PRISM-games model checker and demonstrate the feasibility of robust verification of ICSGs across a selection of large benchmarks. --- ### Kernel-Based Learning of Safety Barriers Authors: Oliver Schön, Zhengang Zhong, Sadegh Soudjani Published: 2026-01-17 Categories: cs.AI, cs.LG, eess.SY arXiv: https://arxiv.org/abs/2601.12002v1 The rapid integration of AI algorithms in safety-critical applications such as autonomous driving and healthcare is raising significant concerns about the ability to meet stringent safety standards. Traditional tools for formal safety verification struggle with the black-box nature of AI-driven systems and lack the flexibility needed to scale to the complexity of real-world applications. In this paper, we present a data-driven approach for safety verification and synthesis of black-box systems with discrete-time stochastic dynamics. We employ the concept of control barrier certificates, which can guarantee safety of the system, and learn the certificate directly from a set of system trajectories. We use conditional mean embeddings to embed data from the system into a reproducing kernel Hilbert space (RKHS) and construct an RKHS ambiguity set that can be inflated to robustify the result to out-of-distribution behavior. We provide the theoretical results on how to apply the approach to general classes of temporal logic specifications beyond safety. For the data-driven computation of safety barriers, we leverage a finite Fourier expansion to cast a typically intractable semi-infinite optimization problem as a linear program. The resulting spectral barrier allows us to leverage the fast Fourier transform to generate the relaxed problem efficiently, offering a scalable yet distributionally robust framework for verifying safety. Our work moves beyond restrictive assumptions on system dynamics and uncertainty, as demonstrated on two case studies including a black-box system with a neural network controller. --- ### Hybrid IDS Using Signature-Based and Anomaly-Based Detection Authors: Messaouda Boutassetta, Amina Makhlouf, Newfel Messaoudi, Abdelmadjid Benmachiche, Ines Boutabia Published: 2026-01-17 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2601.11998v1 Intrusion detection systems (IDS) are essential for protecting computer systems and networks against a wide range of cyber threats that continue to evolve over time. IDS are commonly categorized into two main types, each with its own strengths and limitations, such as difficulty in detecting previously unseen attacks and the tendency to generate high false positive rates. This paper presents a comprehensive survey and a conceptual overview of Hybrid IDS, which integrate signature-based and anomaly-based detection techniques to enhance attack detection capabilities. The survey examines recent research on Hybrid IDS, classifies existing models into functional categories, and discusses their advantages, limitations, and application domains, including financial systems, air traffic control, and social networks. In addition, recent trends in Hybrid IDS research, such as machine learning-based approaches and cloud-based deployments, are reviewed. Finally, this work outlines potential future research directions aimed at developing more cost-effective Hybrid IDS solutions with improved ability to detect emerging and sophisticated cyberattacks. --- ### MongoDB Injection Query Classification Model using MongoDB Log files as Training Data Authors: Shaunak Perni, Minal Shirodkar, Ramdas Karmalli Published: 2026-01-17 Categories: cs.CR, cs.DB, cs.LG arXiv: https://arxiv.org/abs/2601.11996v1 NoSQL Injection attacks are a class of cybersecurity attacks where an attacker sends a specifically engineered query to a NoSQL database which then performs an unauthorized operation. To defend against such attacks, rule based systems were initially developed but then were found to be ineffective to innovative injection attacks hence a model based approach was developed. Most model based detection systems, during testing gave exponentially positive results but were trained only on the query statement sent to the server. However due to the scarcity of data and class imbalances these model based systems were found to be not effective against all attacks in the real world. This paper explores classifying NoSQL injection attacks sent to a MongoDB server based on Log Data, and other extracted features excluding raw query statements. The log data was collected from a simulated attack on an empty MongoDB server which was then processed and explored. A discriminant analysis was carried out to determine statistically significant features to discriminate between injection and benign queries resulting in a dataset of significant features. Several Machine learning based classification models using an AutoML library, "FLAML", as well as 6 manually programmed models were trained on this dataset , which were then trained on 50 randomized samples of data, cross validated and evaluated. The study found that the best model was the "FLAML" library's "XGBoost limited depth" model with an accuracy of 71%. --- ### Learning Audio-Visual Embeddings with Inferred Latent Interaction Graphs Authors: Donghuo Zeng, Hao Niu, Yanan Wang, Masato Taya Published: 2026-01-17 Categories: cs.MM, cs.AI, cs.IR, cs.LG, cs.SD arXiv: https://arxiv.org/abs/2601.11995v1 Learning robust audio-visual embeddings requires bringing genuinely related audio and visual signals together while filtering out incidental co-occurrences - background noise, unrelated elements, or unannotated events. Most contrastive and triplet-loss methods use sparse annotated labels per clip and treat any co-occurrence as semantic similarity. For example, a video labeled "train" might also contain motorcycle audio and visual, because "motorcycle" is not the chosen annotation; standard methods treat these co-occurrences as negatives to true motorcycle anchors elsewhere, creating false negatives and missing true cross-modal dependencies. We propose a framework that leverages soft-label predictions and inferred latent interactions to address these issues: (1) Audio-Visual Semantic Alignment Loss (AV-SAL) trains a teacher network to produce aligned soft-label distributions across modalities, assigning nonzero probability to co-occurring but unannotated events and enriching the supervision signal. (2) Inferred Latent Interaction Graph (ILI) applies the GRaSP algorithm to teacher soft labels to infer a sparse, directed dependency graph among classes. This graph highlights directional dependencies (e.g., "Train (visual)" -> "Motorcycle (audio)") that expose likely semantic or conditional relationships between classes; these are interpreted as estimated dependency patterns. (3) Latent Interaction Regularizer (LIR): A student network is trained with both metric loss and a regularizer guided by the ILI graph, pulling together embeddings of dependency-linked but unlabeled pairs in proportion to their soft-label probabilities. Experiments on AVE and VEGAS benchmarks show consistent improvements in mean average precision (mAP), demonstrating that integrating inferred latent interactions into embedding learning enhances robustness and semantic coherence. --- ### DAOS: A Multimodal In-cabin Behavior Monitoring with Driver Action-Object Synergy Dataset Authors: Yiming Li, Chen Cai, Tianyi Liu, Dan Lin, Wenqian Wang, Wenfei Liang, Bingbing Li, Kim-Hui Yap Published: 2026-01-17 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.11990v1 In driver activity monitoring, movements are mostly limited to the upper body, which makes many actions look similar. To tell these actions apart, human often rely on the objects the driver is using, such as holding a phone compared with gripping the steering wheel. However, most existing driver-monitoring datasets lack accurate object-location annotations or do not link objects to their associated actions, leaving a critical gap for reliable action recognition. To address this, we introduce the Driver Action with Object Synergy (DAOS) dataset, comprising 9,787 video clips annotated with 36 fine-grained driver actions and 15 object classes, totaling more than 2.5 million corresponding object instances. DAOS offers multi-modal, multi-view data (RGB, IR, and depth) from front, face, left, and right perspectives. Although DAOS captures a wide range of cabin objects, only a few are directly relevant to each action for prediction, so focusing on task-specific human-object relations is essential. To tackle this challenge, we propose the Action-Object-Relation Network (AOR-Net). AOR-Net comprehends complex driver actions through multi-level reasoning and a chain-of-action prompting mechanism that models the logical relationships among actions, objects, and their relations. Additionally, the Mixture of Thoughts module is introduced to dynamically select essential knowledge at each stage, enhancing robustness in object-rich and object-scarce conditions. Extensive experiments demonstrate that our model outperforms other state-of-the-art methods on various datasets. --- ### Structural Graph Neural Networks with Anatomical Priors for Explainable Chest X-ray Diagnosis Authors: Khaled Berkani Published: 2026-01-17 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.11987v1 We present a structural graph reasoning framework that incorporates explicit anatomical priors for explainable vision-based diagnosis. Convolutional feature maps are reinterpreted as patch-level graphs, where nodes encode both appearance and spatial coordinates, and edges reflect local structural adjacency. Unlike conventional graph neural networks that rely on generic message passing, we introduce a custom structural propagation mechanism that explicitly models relative spatial relations as part of the reasoning process. This design enables the graph to act as an inductive bias for structured inference rather than a passive relational representation. The proposed model jointly supports node-level lesion-aware predictions and graph-level diagnostic reasoning, yielding intrinsic explainability through learned node importance scores without relying on post-hoc visualization techniques. We demonstrate the approach through a chest X-ray case study, illustrating how structural priors guide relational reasoning and improve interpretability. While evaluated in a medical imaging context, the framework is domain-agnostic and aligns with the broader vision of graph-based reasoning across artificial intelligence systems. This work contributes to the growing body of research exploring graphs as computational substrates for structure-aware and explainable learning. --- ### An AI-IoT Based Smart Wheelchair with Gesture-Controlled Mobility, Deep Learning-Based Obstacle Detection, Multi-Sensor Health Monitoring, and Emergency Alert System Authors: Md. Asiful Islam, Abdul Hasib, Tousif Mahmud Emon, Khandaker Tabin Hasan, A. S. M. Ahsanul Sarkar Akib Published: 2026-01-17 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.11983v1 The growing number of differently-abled and elderly individuals demands affordable, intelligent wheelchairs that combine safe navigation with health monitoring. Traditional wheelchairs lack dynamic features, and many smart alternatives remain costly, single-modality, and limited in health integration. Motivated by the pressing demand for advanced, personalized, and affordable assistive technologies, we propose a comprehensive AI-IoT based smart wheelchair system that incorporates glove-based gesture control for hands-free navigation, real-time object detection using YOLOv8 with auditory feedback for obstacle avoidance, and ultrasonic for immediate collision avoidance. Vital signs (heart rate, SpO$_2$, ECG, temperature) are continuously monitored, uploaded to ThingSpeak, and trigger email alerts for critical conditions. Built on a modular and low-cost architecture, the gesture control achieved a 95.5\% success rate, ultrasonic obstacle detection reached 94\% accuracy, and YOLOv8-based object detection delivered 91.5\% Precision, 90.2\% Recall, and a 90.8\% F1-score. This integrated, multi-modal approach offers a practical, scalable, and affordable solution, significantly enhancing user autonomy, safety, and independence by bridging the gap between innovative research and real-world deployment. --- ### Nip Rumors in the Bud: Retrieval-Guided Topic-Level Adaptation for Test-Time Fake News Video Detection Authors: Jian Lang, Rongpei Hong, Ting Zhong, Yong Wang, Fan Zhou Published: 2026-01-17 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.11981v1 Fake News Video Detection (FNVD) is critical for social stability. Existing methods typically assume consistent news topic distribution between training and test phases, failing to detect fake news videos tied to emerging events and unseen topics. To bridge this gap, we introduce RADAR, the first framework that enables test-time adaptation to unseen news videos. RADAR pioneers a new retrieval-guided adaptation paradigm that leverages stable (source-close) videos from the target domain to guide robust adaptation of semantically related but unstable instances. Specifically, we propose an Entropy Selection-Based Retrieval mechanism that provides videos with stable (low-entropy), relevant references for adaptation. We also introduce a Stable Anchor-Guided Alignment module that explicitly aligns unstable instances' representations to the source domain via distribution-level matching with their stable references, mitigating severe domain discrepancies. Finally, our novel Target-Domain Aware Self-Training paradigm can generate informative pseudo-labels augmented by stable references, capturing varying and imbalanced category distributions in the target domain and enabling RADAR to adapt to the fast-changing label distributions. Extensive experiments demonstrate that RADAR achieves superior performance for test-time FNVD, enabling strong on-the-fly adaptation to unseen fake news video topics. --- ### Process In-Context Learning: Enhancing Mathematical Reasoning via Dynamic Demonstration Insertion Authors: Ang Gao, Changshuo Zhang, Xiao Zhang, Deyang Li, Minjun Zhao, Fangchao Liu, Xinyu Zhang Published: 2026-01-17 Categories: cs.AI, cs.LG arXiv: https://arxiv.org/abs/2601.11979v1 In-context learning (ICL) has proven highly effective across diverse large language model (LLM) tasks. However, its potential for enhancing tasks that demand step-by-step logical deduction, such as mathematical reasoning, remains underexplored. A core limitation of existing ICL approaches is their static use of demonstrations: examples are pre-selected before inference and remain fixed, failing to adapt to the dynamic confusion points that often arise during multi-step reasoning such as ambiguous calculations or logical gaps. These unresolved confusion points can lead to cascading errors that degrade final accuracy. To tackle this issue, we propose Process In-Context Learning (PICL), a dynamic demonstration integration framework designed to boost mathematical reasoning by responding to real-time inference needs. PICL operates in two stages: 1)~it identifies potential confusion points by analyzing semantics and entropy in the reasoning process and summarizes their core characteristics; 2)~upon encountering these points, it retrieves relevant demonstrations from the demonstration pool that match the confusion context and inserts them directly into the ongoing reasoning process to guide subsequent steps. Experiments show that PICL outperforms baseline methods by mitigating mid-inference confusion, highlighting the value of adaptive demonstration insertion in complex mathematical reasoning. --- ### One-Shot Price Forecasting with Covariate-Guided Experts under Privacy Constraints Authors: Ren He, Yinliang Xu, Jinfeng Wang, Jeremy Watson, Jian Song Published: 2026-01-17 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2601.11977v1 Forecasting in power systems often involves multivariate time series with complex dependencies and strict privacy constraints across regions. Traditional forecasting methods require significant expert knowledge and struggle to generalize across diverse deployment scenarios. Recent advancements in pre-trained time series models offer new opportunities, but their zero-shot performance on domain-specific tasks remains limited. To address these challenges, we propose a novel MoE Encoder module that augments pretrained forecasting models by injecting a sparse mixture-of-experts layer between tokenization and encoding. This design enables two key capabilities: (1) trans forming multivariate forecasting into an expert-guided univariate task, allowing the model to effectively capture inter-variable relations, and (2) supporting localized training and lightweight parameter sharing in federated settings where raw data cannot be exchanged. Extensive experiments on public multivariate datasets demonstrate that MoE-Encoder significantly improves forecasting accuracy compared to strong baselines. We further simulate federated environments and show that transferring only MoE-Encoder parameters allows efficient adaptation to new regions, with minimal performance degradation. Our findings suggest that MoE-Encoder provides a scalable and privacy-aware extension to foundation time series models. ---