# Research Papers — 2026-03 · bundle page 198 of 243

> Full-text markdown of papers #9851–9900 of 12123 published in 2026-03, newest first. AI-readable mirrors of the full 12123-paper 2026-03 archive (50 papers per file).

Canonical page: https://aipod.dev/papers/archive/2026-03/page/198
[← Previous bundle](https://aipod.dev/papers/archive/2026-03/page/197.md)
[Next bundle →](https://aipod.dev/papers/archive/2026-03/page/199.md)

### From Entropy to Calibrated Uncertainty: Training Language Models to Reason About Uncertainty

Authors: Azza Jenane, Nassim Walha, Lukas Kuhn, Florian Buettner
Published: 2026-03-06
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2603.06317v1

Large Language Models (LLMs) that can express interpretable and calibrated uncertainty are crucial in high-stakes domains. While methods to compute uncertainty post-hoc exist, they are often sampling-based and therefore computationally expensive or lack calibration. We propose a three-stage pipeline to post-train LLMs to efficiently infer calibrated uncertainty estimates for their responses. First, we compute fine-grained entropy-based uncertainty scores on the training data, capturing the distributional variability of model outputs in embedding space. Second, these scores are calibrated via Platt scaling, producing reliable and human-interpretable uncertainty signals. Finally, the target LLM is post-trained via reinforcement learning to align its policy with these calibrated signals through a verifiable reward function. Unlike post-hoc uncertainty estimation methods, our approach provides interpretable and computationally efficient uncertainty estimates at test time. Experiments show that models trained with our pipeline achieve better calibration than baselines and generalize to unseen tasks without further processing, suggesting that they learn a robust uncertainty reasoning behavior.

---

### WMoE-CLIP: Wavelet-Enhanced Mixture-of-Experts Prompt Learning for Zero-Shot Anomaly Detection

Authors: Peng Chen, Chao Huang
Published: 2026-03-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2603.06313v1

Vision-language models have recently shown strong generalization in zero-shot anomaly detection (ZSAD), enabling the detection of unseen anomalies without task-specific supervision. However, existing approaches typically rely on fixed textual prompts, which struggle to capture complex semantics, and focus solely on spatial-domain features, limiting their ability to detect subtle anomalies. To address these challenges, we propose a wavelet-enhanced mixture-of-experts prompt learning method for ZSAD. Specifically, a variational autoencoder is employed to model global semantic representations and integrate them into prompts to enhance adaptability to diverse anomaly patterns. Wavelet decomposition extracts multi-frequency image features that dynamically refine textual embeddings through cross-modal interactions. Furthermore, a semantic-aware mixture-of-experts module is introduced to aggregate contextual information. Extensive experiments on 14 industrial and medical datasets demonstrate the effectiveness of the proposed method.

---

### Latent Transfer Attack: Adversarial Examples via Generative Latent Spaces

Authors: Eitan Shaar, Ariel Shaulov, Yalcin Tur, Gal Chechik, Ravid Shwartz-Ziv
Published: 2026-03-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2603.06311v1

Adversarial attacks are a central tool for probing the robustness of modern vision models, yet most methods optimize perturbations directly in pixel space under $\ell_\infty$ or $\ell_2$ constraints. While effective in white-box settings, pixel-space optimization often produces high-frequency, texture-like noise that is brittle to common preprocessing (e.g., resizing and cropping) and transfers poorly across architectures. We propose $\textbf{LTA}$ ($\textbf{L}$atent $\textbf{T}$ransfer $\textbf{A}$ttack), a transfer-based attack that instead optimizes perturbations in the latent space of a pretrained Stable Diffusion VAE. Given a clean image, we encode it into a latent code and optimize the latent representation to maximize a surrogate classifier loss, while softly enforcing a pixel-space $\ell_\infty$ budget after decoding. To improve robustness to resolution mismatch and standard input pipelines, we incorporate Expectation Over Transformations (EOT) via randomized resizing, interpolation, and cropping, and apply periodic latent Gaussian smoothing to suppress emerging artifacts and stabilize optimization. Across a suite of CNN and vision-transformer targets, LTA achieves strong transfer attack success while producing spatially coherent, predominantly low-frequency perturbations that differ qualitatively from pixel-space baselines and occupy a distinct point in the transfer-quality trade-off. Our results highlight pretrained generative latent spaces as an effective and structured domain for adversarial optimization, bridging robustness evaluation with modern generative priors.

---

### Continual Adaptation for Pacific Indigenous Speech Recognition

Authors: Yang Xiao, Aso Mahmudi, Nick Thieberger, Eliathamby Ambikairajah, Eun-Jung Holden, Ting Dang
Published: 2026-03-06
Categories: eess.AS, cs.CL, cs.SD
arXiv: https://arxiv.org/abs/2603.06310v2

Speech foundation models struggle with low-resource Pacific Indigenous languages because of severe data scarcity. Furthermore, full fine-tuning risks catastrophic forgetting. To address this gap, we present an empirical study adapting models to real-world Pacific datasets. We investigate the impact of data volume, adaptation strategies, and representational drift on speech foundation models for various Pacific languages. Additionally, we analyze a continual learning framework for sequential language acquisition. Empirical results across three distinct Pacific Indigenous languages demonstrate that adapting to these linguistically distant languages induces severe internal representational drift. Consequently, these models face a strict plasticity and stability dilemma. While LoRA adapts well initially, it suffers from catastrophic forgetting during sequential learning. Ultimately, this study highlights the urgent need for robust adaptation strategies tailored to underrepresented languages.

---

### Polarized Direct Cross-Attention Message Passing in GNNs for Machinery Fault Diagnosis

Authors: Zongyu Shi, Laibin Zhang, Maoyin Chen
Published: 2026-03-06
Categories: cs.LG
arXiv: https://arxiv.org/abs/2603.06303v1

The reliability of safety-critical industrial systems hinges on accurate and robust fault diagnosis in rotating machinery. Conventional graph neural networks (GNNs) for machinery fault diagnosis face limitations in modeling complex dynamic interactions due to their reliance on predefined static graph structures and homogeneous aggregation schemes. To overcome these challenges, this paper introduces polarized direct cross-attention (PolaDCA), a novel relational learning framework that enables adaptive message passing through data-driven graph construction. Our approach builds upon a direct cross-attention (DCA) mechanism that dynamically infers attention weights from three semantically distinct node features (such as individual characteristics, neighborhood consensus, and neighborhood diversity) without requiring fixed adjacency matrices. Theoretical analysis establishes PolaDCA's superior noise robustness over conventional GNNs. Extensive experiments on industrial datasets (i.e., XJTUSuprgear, CWRUBearing and Three-Phase Flow Facility datasets) demonstrate state-of-the-art diagnostic accuracy and enhanced generalization under varying noise conditions, outperforming seven competitive baseline methods. The proposed framework provides an effective solution for safety-critical industrial applications.

---

### DEX-AR: A Dynamic Explainability Method for Autoregressive Vision-Language Models

Authors: Walid Bousselham, Angie Boggust, Hendrik Strobelt, Hilde Kuehne
Published: 2026-03-06
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2603.06302v1

As Vision-Language Models (VLMs) become increasingly sophisticated and widely used, it becomes more and more crucial to understand their decision-making process. Traditional explainability methods, designed for classification tasks, struggle with modern autoregressive VLMs due to their complex token-by-token generation process and intricate interactions between visual and textual modalities. We present DEX-AR (Dynamic Explainability for AutoRegressive models), a novel explainability method designed to address these challenges by generating both per-token and sequence-level 2D heatmaps highlighting image regions crucial for the model's textual responses. The proposed method offers to interpret autoregressive VLMs-including varying importance of layers and generated tokens-by computing layer-wise gradients with respect to attention maps during the token-by-token generation process. DEX-AR introduces two key innovations: a dynamic head filtering mechanism that identifies attention heads focused on visual information, and a sequence-level filtering approach that aggregates per-token explanations while distinguishing between visually-grounded and purely linguistic tokens. Our evaluation on ImageNet, VQAv2, and PascalVOC, shows a consistent improvement in both perturbation-based metrics, using a novel normalized perplexity measure, as well as segmentation-based metrics.

---

### 3D CBCT Artefact Removal Using Perpendicular Score-Based Diffusion Models

Authors: Susanne Schaub, Florentin Bieder, Matheus L. Oliveira, Yulan Wang, Dorothea Dagassan-Berndt, Michael M. Bornstein, Philippe C. Cattin
Published: 2026-03-06
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2603.06300v1

Cone-beam computed tomography (CBCT) is a widely used 3D imaging technique in dentistry, offering high-resolution images while minimising radiation exposure for patients. However, CBCT is highly susceptible to artefacts arising from high-density objects such as dental implants, which can compromise image quality and diagnostic accuracy. To reduce artefacts, implant inpainting in the sequence of projections plays a crucial role in many artefact reduction approaches. Recently, diffusion models have achieved state-of-the-art results in image generation and have widely been applied to image inpainting tasks. However, to our knowledge, existing diffusion-based methods for implant inpainting operate on independent 2D projections. This approach neglects the correlations among individual projections, resulting in inconsistencies in the reconstructed images. To address this, we propose a 3D dental implant inpainting approach based on perpendicular score-based diffusion models, each trained in two different planes and operating in the projection domain. The 3D distribution of the projection series is modelled by combining the two 2D score-based diffusion models in the sampling scheme. Our results demonstrate the method's effectiveness in producing high-quality, artefact-reduced 3D CBCT images, making it a promising solution for improving clinical imaging.

---

### Enhanced Protein Intrinsic Disorder Prediction Through Dual-View Multiscale Features and Multi-objective Evolutionary Algorithm

Authors: Shaokuan Wang, Pengshan Cui, Yining Qian, An-Yang Lu, Xianpeng Wang
Published: 2026-03-06
Categories: cs.NE
arXiv: https://arxiv.org/abs/2603.06292v1

Intrinsically disordered regions of proteins play a crucial role in cell signaling and drug discovery. However, their high structural flexibility makes accurate residue-level prediction challenging. Existing methods often rely on single-view representations or rigid manual fusion strategies, which fail to effectively balance the complex interplay between local amino acid preferences and long-range sequence patterns. To address these limitations, we propose D2MOE, a Dual-View Multiscale Features and Multi-objective Evolutionary Algorithm, which consists of two stages. First, a dual-view multiscale feature extraction method is introduced. This method integrates evolutionary views with deep semantic views and employs multiscale extractors to capture structural information across diverse receptive fields. Second, a multi-objective evolutionary algorithm is designed to adaptively discover optimal fusion architectures. By co-evolving discrete feature selection and continuous fusion weights, the algorithm adaptively explores optimal cross-feature architectures to enhance predictive accuracy while maintaining model compactness. Experimental results across three benchmark datasets demonstrate that D2MOE consistently outperforms state-of-the-art methods. D2MOE combines the feature extraction capabilities of deep learning with the global search advantages of evolutionary algorithms, enabling efficient feature integration without manual design, and providing a robust computational tool for protein disorder prediction.

---

### The EpisTwin: A Knowledge Graph-Grounded Neuro-Symbolic Architecture for Personal AI

Authors: Giovanni Servedio, Potito Aghilar, Alessio Mattiace, Gianni Carmosino, Francesco Musicco, Gabriele Conte, Vito Walter Anelli, Tommaso Di Noia, Francesco Maria Donini
Published: 2026-03-06
Categories: cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2603.06290v1

Personal Artificial Intelligence is currently hindered by the fragmentation of user data across isolated silos. While Retrieval-Augmented Generation offers a partial remedy, its reliance on unstructured vector similarity fails to capture the latent semantic topology and temporal dependencies essential for holistic sensemaking. We introduce EpisTwin, a neuro-symbolic framework that grounds generative reasoning in a verifiable, user-centric Personal Knowledge Graph. EpisTwin leverages Multimodal Language Models to lift heterogeneous, cross-application data into semantic triples. At inference, EpisTwin enables complex reasoning over the personal semantic graph via an agentic coordinator that combines Graph Retrieval-Augmented Generation with Online Deep Visual Refinement, dynamically re-grounding symbolic entities in their raw visual context. We also introduce PersonalQA-71-100, a synthetic benchmark designed to simulate a realistic user's digital footprint and evaluate EpisTwin performance. Our framework demonstrates robust results across a suite of state-of-the-art judge models, offering a promising direction for trustworthy Personal AI.

---

### FlowMotion: Training-Free Flow Guidance for Video Motion Transfer

Authors: Zhen Wang, Youcan Xu, Jun Xiao, Long Chen
Published: 2026-03-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2603.06289v2

Video motion transfer aims to generate a target video that inherits motion patterns from a source video while rendering new scenes. Existing training-free approaches focus on constructing motion guidance based on the intermediate outputs of pre-trained T2V models, which results in heavy computational overhead and limited flexibility. In this paper, we present FlowMotion, a novel training-free framework that enables efficient and flexible motion transfer by directly leveraging the predicted outputs of flow-based T2V models. Our key insight is that early latent predictions inherently encode rich temporal information. Motivated by this, we propose flow guidance, which extracts motion representations based on latent predictions to align motion patterns between source and generated videos. We further introduce a velocity regularization strategy to stabilize optimization and ensure smooth motion evolution. By operating purely on model predictions, FlowMotion achieves superior time and resource efficiency as well as competitive performance compared with state-of-the-art methods.

---

### Learning Where the Physics Is: Probabilistic Adaptive Sampling for Stiff PDEs

Authors: Akshay Govind Srinivasan, Balaji Srinivasan
Published: 2026-03-06
Categories: cs.CE, cs.AI, cs.LG, math.AP
arXiv: https://arxiv.org/abs/2603.06287v1

Modeling stiff partial differential equations (PDEs) with sharp gradients remains a significant challenge for scientific machine learning. While Physics-Informed Neural Networks (PINNs) struggle with spectral bias and slow training times, Physics-Informed Extreme Learning Machines (PIELMs) offer a rapid, closed-form linear solution but are fundamentally limited by physics-agnostic, random initialization. We introduce the Gaussian Mixture Model Adaptive PIELM (GMM-PIELM), a probabilistic framework that learns a probability density function representing the ``location of physics'' for adaptively sampling kernels of PIELMs. By employing a weighted Expectation-Maximization (EM) algorithm, GMM-PIELM autonomously concentrates radial basis function centers in regions of high numerical error, such as shock fronts and boundary layers. This approach dynamically improves the conditioning of the hidden layer without the expensive gradient-based optimization(of PINNs) or Bayesian search. We evaluate our methodology on 1D singularly perturbed convection-diffusion equations with diffusion coefficients $ν=10^{-4}$. Our method achieves $L_2$ errors up to $7$ orders of magnitude lower than baseline RBF-PIELMs, successfully resolving exponentially thin boundary layers while retaining the orders-of-magnitude speed advantage of the ELM architecture.

---

### Attribute Distribution Modeling and Semantic-Visual Alignment for Generative Zero-shot Learning

Authors: Haojie Pu, Zhuoming Li, Yongbiao Gao, Yuheng Jia
Published: 2026-03-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2603.06281v2

Generative zero-shot learning (ZSL) synthesizes features for unseen classes, leveraging semantic conditions to transfer knowledge from seen classes. However, it also introduces two intrinsic challenges: (1) class-level attributes fails to capture instance-specific visual appearances due to substantial intra-class variability, thus causing the class-instance gap; (2) the substantial mismatch between semantic and visual feature distributions, manifested in inter-class correlations, gives rise to the semantic-visual domain gap. To address these challenges, we propose an Attribute Distribution Modeling and Semantic-Visual Alignment (ADiVA) approach, jointly modeling attribute distributions and performing explicit semantic-visual alignment. Specifically, our ADiVA consists of two modules: an Attribute Distribution Modeling (ADM) module that learns a transferable attribute distribution for each class and samples instance-level attributes for unseen classes, and a Visual-Guided Alignment (VGA) module that refines semantic representations to better reflect visual structures. Experiments on three widely used benchmark datasets demonstrate that ADiVA significantly outperforms state-of-the-art methods (e.g., achieving gains of 4.7% and 6.1% on AWA2 and SUN, respectively). Moreover, our approach can serve as a plugin to enhance existing generative ZSL methods.

---

### Can we Trust Unreliable Voxels? Exploring 3D Semantic Occupancy Prediction under Label Noise

Authors: Wenxin Li, Kunyu Peng, Di Wen, Junwei Zheng, Jiale Wei, Mengfei Duan, Yuheng Zhang, Rui Fan, Kailun Yang
Published: 2026-03-06
Categories: cs.CV, cs.RO, eess.IV
arXiv: https://arxiv.org/abs/2603.06279v2

3D semantic occupancy prediction is a cornerstone of robotic perception, yet real-world voxel annotations are inherently corrupted by structural artifacts and dynamic trailing effects. This raises a critical but underexplored question: can autonomous systems safely rely on such unreliable occupancy supervision? To systematically investigate this issue, we establish OccNL, the first benchmark dedicated to 3D occupancy under occupancy-asymmetric and dynamic trailing noise. Our analysis reveals a fundamental domain gap: state-of-the-art 2D label noise learning strategies collapse catastrophically in sparse 3D voxel spaces, exposing a critical vulnerability in existing paradigms. To address this challenge, we propose DPR-Occ, a principled label-noise-robust framework that constructs reliable supervision through dual-source partial label reasoning. By synergizing temporal model memory with representation-level structural affinity, DPR-Occ dynamically expands and prunes candidate label sets to preserve true semantics while suppressing noise propagation. Extensive experiments on SemanticKITTI demonstrate that DPR-Occ prevents geometric and semantic collapse under extreme corruption. Notably, even at 90% label noise, our method achieves significant performance gains (up to 2.57% mIoU and 13.91% IoU) over existing label noise learning baselines adapted to the 3D occupancy prediction task. By bridging label noise learning and 3D perception, OccNL and DPR-Occ provide a reliable foundation for safety-critical robotic perception in dynamic environments. The benchmark and source code will be made publicly available at https://github.com/mylwx/OccNL.

---

### Artificial Intelligence for Climate Adaptation: Reinforcement Learning for Climate Change-Resilient Transport

Authors: Miguel Costa, Arthur Vandervoort, Carolin Schmidt, João Miranda, Morten W. Petersen, Martin Drews, Karyn Morrisey, Francisco C. Pereira
Published: 2026-03-06
Categories: cs.AI
arXiv: https://arxiv.org/abs/2603.06278v1

Climate change is expected to intensify rainfall and, consequently, pluvial flooding, leading to increased disruptions in urban transportation systems over the coming decades. Designing effective adaptation strategies is challenging due to the long-term, sequential nature of infrastructure investments, deep climate uncertainty, and the complex interactions between flooding, infrastructure, and mobility impacts. In this work, we propose a novel decision-support framework using reinforcement learning (RL) for long-term flood adaptation planning. Formulated as an integrated assessment model (IAM), the framework combines rainfall projection and flood modeling, transport simulation, and quantification of direct and indirect impacts on infrastructure and mobility. Our RL-based approach learns adaptive strategies that balance investment and maintenance costs against avoided impacts. We evaluate the framework through a case study of Copenhagen's inner city over the 2024-2100 period, testing multiple adaptation options, and different belief and realized climate scenarios. Results show that the framework outperforms traditional optimization approaches by discovering coordinated spatial and temporal adaptation pathways and learning trade-offs between impact reduction and adaptation investment, yielding more resilient strategies. Overall, our results showcase the potential of reinforcement learning as a flexible decision-support tool for adaptive infrastructure planning under climate uncertainty.

---

### Spectral and Trajectory Regularization for Diffusion Transformer Super-Resolution

Authors: Jingkai Wang, Yixin Tang, Jue Gong, Jiatong Li, Shu Li, Libo Liu, Jianliang Lan, Yutong Liu, Yulun Zhang
Published: 2026-03-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2603.06275v2

Diffusion transformer (DiT) architectures show great potential for real-world image super-resolution (Real-ISR). However, their computationally expensive iterative sampling necessitates one-step distillation. Existing one-step distillation methods struggle with Real-ISR on DiT. They suffer from fundamental trajectory mismatch and generate severe grid-like periodic artifacts. To tackle these challenges, we propose StrSR, a novel one-step adversarial distillation framework featuring spectral and trajectory regularization. Specifically, we propose an asymmetric discriminative distillation architecture to bridge the trajectory gap. Additionally, we design a frequency distribution matching strategy to effectively suppress DiT-specific periodic artifacts caused by high-frequency spectral leakage. Extensive experiments demonstrate that StrSR achieves state-of-the-art performance in Real-ISR, across both quantitative metrics and visual perception. The code and models will be released at https://github.com/jkwang28/StrSR .

---

### Stem: Rethinking Causal Information Flow in Sparse Attention

Authors: Lin Niu, Xin Luo, Linchuan Xie, Yifu Sun, Guanghua Yu, Jianchen Zhu, S Kevin Zhou
Published: 2026-03-06
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2603.06274v2

The quadratic computational complexity of self-attention remains a fundamental bottleneck for scaling Large Language Models (LLMs) to long contexts, particularly during the pre-filling phase. In this paper, we rethink the causal attention mechanism from the perspective of information flow. Due to causal constraints, tokens at initial positions participate in the aggregation of every subsequent token. However, existing sparse methods typically apply a uniform top-k selection across all token positions within a layer, ignoring the cumulative dependency of token information inherent in causal architectures. To address this, we propose Stem, a novel, plug-and-play sparsity module aligned with information flow. First, Stem employs the Token Position-Decay strategy, applying position-dependent top-k within each layer to retain initial tokens for recursive dependencies. Second, to preserve information-rich tokens, Stem utilizes the Output-Aware Metric. It prioritizes high-impact tokens based on approximate output magnitude. Extensive evaluations demonstrate that Stem achieves superior accuracy with reduced computation and pre-filling latency.

---

### Looking Through Glass Box

Authors: Alexis Kafantaris
Published: 2026-03-06
Categories: cs.NE, cs.AI, cs.LG, cs.SC
arXiv: https://arxiv.org/abs/2603.06272v1

This essay is about a neural implementation of the fuzzy cognitive map, the FHM, and corresponding evaluations. Firstly, a neural net has been designed to behave the same way that an FCM does; as inputs it accepts many fuzzy cognitive maps and propagates them in order to learn causality patterns. Moreover, the network uses langevin differential Dynamics, which avoid overfit, to inverse solve the output node values according to some policy. Nevertheless, having obtained an inverse solution provides the user a modification criterion. Having the modification criterion suggests that information is now according to discretion as a different service or product is a better fit. Lastly, evaluation has been done on several data sets in order to examine the networks performance.

---

### Agentic retrieval-augmented reasoning reshapes collective reliability under model variability in radiology question answering

Authors: Mina Farajiamiri, Jeta Sopa, Saba Afza, Lisa Adams, Felix Barajas Ordonez, Tri-Thien Nguyen, Mahshad Lotfinia, Sebastian Wind, Keno Bressem, Sven Nebelung, Daniel Truhn, Soroosh Tayebi Arasteh
Published: 2026-03-06
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2603.06271v1

Agentic retrieval-augmented reasoning pipelines are increasingly used to structure how large language models (LLMs) incorporate external evidence in clinical decision support. These systems iteratively retrieve curated domain knowledge and synthesize it into structured reports before answer selection. Although such pipelines can improve performance, their impact on reliability under model variability remains unclear. In real-world deployment, heterogeneous models may align, diverge, or synchronize errors in ways not captured by accuracy. We evaluated 34 LLMs on 169 expert-curated publicly available radiology questions, comparing zero-shot inference with a radiology-specific multi-step agentic retrieval condition in which all models received identical structured evidence reports derived from curated radiology knowledge. Agentic inference reduced inter-model decision dispersion (median entropy 0.48 vs. 0.13) and increased robustness of correctness across models (mean 0.74 vs. 0.81). Majority consensus also increased overall (P<0.001). Consensus strength and robust correctness remained correlated under both strategies (\r{ho}=0.88 for zero-shot; \r{ho}=0.87 for agentic), although high agreement did not guarantee correctness. Response verbosity showed no meaningful association with correctness. Among 572 incorrect outputs, 72% were associated with moderate or high clinically assessed severity, although inter-rater agreement was low (\k{appa}=0.02). Agentic retrieval therefore was associated with more concentrated decision distributions, stronger consensus, and higher cross-model robustness of correctness. These findings suggest that evaluating agentic systems through accuracy or agreement alone may not always be sufficient, and that complementary analyses of stability, cross-model robustness, and potential clinical impact are needed to characterize reliability under model variability.

---

### HiPP-Prune: Hierarchical Preference-Conditioned Structured Pruning for Vision-Language Models

Authors: Lincen Bai, Hedi Tabia, Raul Santos-Rodriguez
Published: 2026-03-06
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2603.06270v1

Pruning vision-language models (VLMs) for efficient deployment is challenging because compression can affect not only task utility but also visual grounding, often amplifying object hallucinations even at the same sparsity level. We present HiPP-Prune, a hierarchical preference-conditioned structured pruning framework that treats pruning as conditional resource allocation under multiple objectives. HiPP-Prune makes plan-level decisions: a single policy invocation outputs a global pruning blueprint by factorizing decisions into an overall sparsity budget and a layer-wise allocation, enabling queryable trade-offs via a user-specified preference vector. To account for VLM-specific failure modes, our policy state integrates a visual sensitivity signal derived from attention flow between vision tokens and language hidden states, discouraging over-pruning of vision-critical layers that facilitate cross-modal fusion. We optimize pruning plans with plan-level Group Relative Policy Optimization (GRPO) under a multi-objective return that combines task utility, hallucination robustness (POPE), compression, and a synaptic-flow-inspired stability proxy to reduce unproductive exploration in high-sparsity regimes. Experiments on LLaVA with POPE and ScienceQA demonstrate that HiPP-Prune discovers diverse non-dominated pruning plans and provides controllable robustness--utility trade-offs under matched sparsity budgets.

---

### ODD-SEC: Onboard Drone Detection with a Spinning Event Camera

Authors: Kuan Dai, Hongxin Zhang, Sheng Zhong, Yi Zhou
Published: 2026-03-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2603.06265v1

The rapid proliferation of drones requires balancing innovation with regulation. To address security and privacy concerns, techniques for drone detection have attracted significant attention.Passive solutions, such as frame camera-based systems, offer versatility and energy efficiency under typical conditions but are fundamentally constrained by their operational principles in scenarios involving fast-moving targets or adverse illumination.Inspired by biological vision, event cameras asynchronously detect per-pixel brightness changes, offering high dynamic range and microsecond-level responsiveness that make them uniquely suited for drone detection in conditions beyond the reach of conventional frame-based cameras.However, the design of most existing event-based solutions assumes a static camera, greatly limiting their applicability to moving carriers--such as quadrupedal robots or unmanned ground vehicles--during field operations.In this paper, we introduce a real-time drone detection system designed for deployment on moving carriers. The system utilizes a spinning event-based camera, providing a 360° horizontal field of view and enabling bearing estimation of detected drones. A key contribution is a novel image-like event representation that operates without motion compensation, coupled with a lightweight neural network architecture for efficient spatiotemporal learning. Implemented on an onboard Jetson Orin NX, the system can operate in real time. Outdoor experimental results validate reliable detection with a mean angular error below 2° under challenging conditions, underscoring its suitability for real-world surveillance applications. We will open-source our complete pipeline to support future research.

---

### Mind the Gap: Pitfalls of LLM Alignment with Asian Public Opinion

Authors: Hari Shankar, Vedanta S P, Sriharini Margapuri, Debjani Mazumder, Ponnurangam Kumaraguru, Abhijnan Chakraborty
Published: 2026-03-06
Categories: cs.CL, cs.CY
arXiv: https://arxiv.org/abs/2603.06264v2

Large Language Models (LLMs) are increasingly being deployed in multilingual, multicultural settings, yet their reliance on predominantly English-centric training data risks misalignment with the diverse cultural values of different societies. In this paper, we present a comprehensive, multilingual audit of the cultural alignment of contemporary LLMs including GPT-4o-Mini, Gemini-2.5-Flash, Llama 3.2, Mistral and Gemma 3 across India, East Asia and Southeast Asia. Our study specifically focuses on the sensitive domain of religion as the prism for broader alignment. To facilitate this, we conduct a multi-faceted analysis of every LLM's internal representations, using log-probs/logits, to compare the model's opinion distributions against ground-truth public attitudes. We find that while the popular models generally align with public opinion on broad social issues, they consistently fail to accurately represent religious viewpoints, especially those of minority groups, often amplifying negative stereotypes. Lightweight interventions, such as demographic priming and native language prompting, partially mitigate but do not eliminate these cultural gaps. We further show that downstream evaluations on bias benchmarks (such as CrowS-Pairs, IndiBias, ThaiCLI, KoBBQ) reveal persistent harms and under-representation in sensitive contexts. Our findings underscore the urgent need for systematic, regionally grounded audits to ensure equitable global deployment of LLMs.

---

### Learning to Solve Orienteering Problem with Time Windows and Variable Profits

Authors: Songqun Gao, Zanxi Ruan, Patrick Floor, Marco Roveri, Luigi Palopoli, Daniele Fontanelli
Published: 2026-03-06
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2603.06260v1

The orienteering problem with time windows and variable profits (OPTWVP) is common in many real-world applications and involves continuous time variables. Current approaches fail to develop an efficient solver for this orienteering problem variant with discrete and continuous variables. In this paper, we propose a learning-based two-stage DEcoupled discrete-Continuous optimization with Service-time-guided Trajectory (DeCoST), which aims to effectively decouple the discrete and continuous decision variables in the OPTWVP problem, while enabling efficient and learnable coordination between them. In the first stage, a parallel decoding structure is employed to predict the path and the initial service time allocation. The second stage optimizes the service times through a linear programming (LP) formulation and provides a long-horizon learning of structure estimation. We rigorously prove the global optimality of the second-stage solution. Experiments on OPTWVP instances demonstrate that DeCoST outperforms both state-of-the-art constructive solvers and the latest meta-heuristic algorithms in terms of solution quality and computational efficiency, achieving up to 6.6x inference speedup on instances with fewer than 500 nodes. Moreover, the proposed framework is compatible with various constructive solvers and consistently enhances the solution quality for OPTWVP.

---

### Robust support vector model based on bounded asymmetric elastic net loss for binary classification

Authors: Haiyan Du, Hu Yang
Published: 2026-03-06
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2603.06257v2

In this paper, we propose a novel bounded asymmetric elastic net ($L_{baen}$) loss function and combine it with the support vector machine (SVM), resulting in the BAEN-SVM. The $L_{baen}$ is bounded and asymmetric and can degrade to the asymmetric elastic net hinge loss, pinball loss, and asymmetric least squares loss. BAEN-SVM not only effectively handles noise-contaminated data but also addresses the geometric irrationalities in the traditional SVM. By proving the violation tolerance upper bound (VTUB) of BAEN-SVM, we show that the model is geometrically well-defined. Furthermore, we derive that the influence function of BAEN-SVM is bounded, providing a theoretical guarantee of its robustness to noise. The Fisher consistency of the model further ensures its generalization capability. Since the \( L_{\text{baen}} \) loss is non-convex, we designed a clipping dual coordinate descent-based half-quadratic algorithm to solve the non-convex optimization problem efficiently. Experimental results on artificial and benchmark datasets indicate that the proposed method outperforms classical and advanced SVMs, particularly in noisy environments.

---

### GazeMoE: Perception of Gaze Target with Mixture-of-Experts

Authors: Zhuangzhuang Dai, Zhongxi Lu, Vincent G. Zakka, Luis J. Manso, Jose M Alcaraz Calero, Chen Li
Published: 2026-03-06
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2603.06256v1

Estimating human gaze target from visible images is a critical task for robots to understand human attention, yet the development of generalizable neural architectures and training paradigms remains challenging. While recent advances in pre-trained vision foundation models offer promising avenues for locating gaze targets, the integration of multi-modal cues -- including eyes, head poses, gestures, and contextual features -- demands adaptive and efficient decoding mechanisms. Inspired by Mixture-of-Experts (MoE) for adaptive domain expertise in large vision-language models, we propose GazeMoE, a novel end-to-end framework that selectively leverages gaze-target-related cues from a frozen foundation model through MoE modules. To address class imbalance in gaze target classification (in-frame vs. out-of-frame) and enhance robustness, GazeMoE incorporates a class-balancing auxiliary loss alongside strategic data augmentations, including region-specific cropping and photometric transformations. Extensive experiments on benchmark datasets demonstrate that our GazeMoE achieves state-of-the-art performance, outperforming existing methods on challenging gaze estimation tasks. The code and pre-trained models are released at https://huggingface.co/zdai257/GazeMoE

---

### NOVA: Next-step Open-Vocabulary Autoregression for 3D Multi-Object Tracking in Autonomous Driving

Authors: Kai Luo, Xu Wang, Rui Fan, Kailun Yang
Published: 2026-03-06
Categories: cs.CV, cs.RO, eess.IV
arXiv: https://arxiv.org/abs/2603.06254v2

Generalizing across unknown targets is critical for open-world perception, yet existing 3D Multi-Object Tracking (3D MOT) pipelines remain limited by closed-set assumptions and ``semantic-blind'' heuristics. To address this, we propose Next-step Open-Vocabulary Autoregression (NOVA), an autoregressive association formulation that shifts the data association stage from fragmented distance-based matching toward trajectory-conditioned spatio-semantic modeling. NOVA reformulates 3D trajectories as structured spatio-temporal semantic sequences, enabling the simultaneous encoding of physical motion continuity and deep linguistic priors. By leveraging the autoregressive capabilities of Large Language Models (LLMs), we transform the tracking task into a principled process of next-step sequence completion. This mechanism allows the model to explicitly utilize the hierarchical structure of language space to resolve fine-grained semantic ambiguities and maintain identity consistency across complex long-range sequences through high-level commonsense reasoning. Extensive experiments on nuScenes, V2X-Seq-SPD, and KITTI demonstrate the superior performance of NOVA. Notably, on the nuScenes dataset, NOVA achieves an AMOTA of 22.41% for Novel categories, yielding a significant 20.21% absolute improvement over the baseline. These gains are realized through a compact 0.5B autoregressive model. Code will be available at https://github.com/xifen523/NOVA.

---

### Synthetic Monitoring Environments for Reinforcement Learning

Authors: Leonard Pleiss, Carolin Schmidt, Maximilian Schiffer
Published: 2026-03-06
Categories: cs.LG, stat.ML
arXiv: https://arxiv.org/abs/2603.06252v1

Reinforcement Learning (RL) lacks benchmarks that enable precise, white-box diagnostics of agent behavior. Current environments often entangle complexity factors and lack ground-truth optimality metrics, making it difficult to isolate why algorithms fail. We introduce Synthetic Monitoring Environments (SMEs), an infinite suite of continuous control tasks. SMEs provide fully configurable task characteristics and known optimal policies. As such, SMEs allow for the exact calculation of instantaneous regret. Their rigorous geometric state space bounds allow for systematic within-distribution (WD) and out-of-distribution (OOD) evaluation. We demonstrate the framework's benefit through multidimensional ablations of PPO, TD3, and SAC, revealing how specific environmental properties - such as action or state space size, reward sparsity and complexity of the optimal policy - impact WD and OOD performance. We thereby show that SMEs offer a standardized, transparent testbed for transitioning RL evaluation from empirical benchmarking toward rigorous scientific analysis.

---

### SPPCSO: Adaptive Penalized Estimation Method for High-Dimensional Correlated Data

Authors: Ying Hu, Hu Yang
Published: 2026-03-06
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2603.06251v1

With the rise of high-dimensional correlated data, multicollinearity poses a significant challenge to model stability, often leading to unstable estimation and reduced predictive accuracy. This work proposes the Single-Parametric Principal Component Selection Operator (SPPCSO), an innovative penalized estimation method that integrates single-parametric principal component regression and $L_{1}$ regularization to adaptively adjust the shrinkage factor by incorporating principal component information. This approach achieves a balance between variable selection and coefficient estimation, ensuring model stability and robust estimation even in high-dimensional, high-noise environments. The primary contribution lies in addressing the instability of traditional variable selection methods when applied to high-noise, high-dimensional correlated data. Theoretically, our method exhibits selection consistency and achieves a smaller estimation error bound compared to traditional penalized estimation approaches. Extensive numerical experiments demonstrate that SPPCSO not only delivers stable and reliable estimation in high-noise settings but also accurately distinguishes signal variables from noise variables in group-effect structured data with highly correlated noise variables, effectively eliminating redundant variables and achieving more stable variable selection. Furthermore, SPPCSO successfully identifies disease-associated genes in gene expression data analysis, showcasing strong practical value. The results indicate that SPPCSO serves as an ideal tool for high-dimensional variable selection, offering an efficient and interpretable solution for modeling correlated data.

---

### Hierarchical Collaborative Fusion for 3D Instance-aware Referring Expression Segmentation

Authors: Keshen Zhou, Runnan Chen, Mingming Gong, Tongliang Liu
Published: 2026-03-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2603.06250v1

Generalised 3D Referring Expression Segmentation (3D-GRES) localizes objects in 3D scenes based on natural language, even when descriptions match multiple or zero targets. Existing methods rely solely on sparse point clouds, lacking rich visual semantics for fine-grained descriptions. We propose HCF-RES, a multi-modal framework with two key innovations. First, Hierarchical Visual Semantic Decomposition leverages SAM instance masks to guide CLIP encoding at dual granularities -- pixel-level and instance-level features -- preserving object boundaries during 2D-to-3D projection. Second, Progressive Multi-level Fusion integrates representations through intra-modal collaboration, cross-modal adaptive weighting between 2D semantic and 3D geometric features, and language-guided refinement. HCF-RES achieves state-of-the-art results on both ScanRefer and Multi3DRefer.

---

### Gradient Flow Polarizes Softmax Outputs towards Low-Entropy Solutions

Authors: Aditya Varre, Mark Rofin, Nicolas Flammarion
Published: 2026-03-06
Categories: cs.LG, math.OC, stat.ML
arXiv: https://arxiv.org/abs/2603.06248v1

Understanding the intricate non-convex training dynamics of softmax-based models is crucial for explaining the empirical success of transformers. In this article, we analyze the gradient flow dynamics of the value-softmax model, defined as ${L}(\mathbf{V} σ(\mathbf{a}))$, where $\mathbf{V}$ and $\mathbf{a}$ are a learnable value matrix and attention vector, respectively. As the matrix times softmax vector parameterization constitutes the core building block of self-attention, our analysis provides direct insight into transformer's training dynamics. We reveal that gradient flow on this structure inherently drives the optimization toward solutions characterized by low-entropy outputs. We demonstrate the universality of this polarizing effect across various objectives, including logistic and square loss. Furthermore, we discuss the practical implications of these theoretical results, offering a formal mechanism for empirical phenomena such as attention sinks and massive activations.

---

### DC-Merge: Improving Model Merging with Directional Consistency

Authors: Han-Chen Zhang, Zi-Hao Zhou, Mao-Lin Luo, Shimin Di, Min-Ling Zhang, Tong Wei
Published: 2026-03-06
Categories: cs.LG, cs.CV
arXiv: https://arxiv.org/abs/2603.06242v2

Model merging aims to integrate multiple task-adapted models into a unified model that preserves the knowledge of each task. In this paper, we identify that the key to this knowledge retention lies in maintaining the directional consistency of singular spaces between merged multi-task vector and individual task vectors. However, this consistency is frequently compromised by two issues: i) an imbalanced energy distribution within task vectors, where a small fraction of singular values dominate the total energy, leading to the neglect of semantically important but weaker components upon merging, and ii) the geometric inconsistency of task vectors in parameter space, which causes direct merging to distort their underlying directional geometry. To address these challenges, we propose DC-Merge, a method for directional-consistent model merging. It first balances the energy distribution of each task vector by smoothing its singular values, ensuring all knowledge components are adequately represented. These energy-balanced vectors are then projected onto a shared orthogonal subspace to align their directional geometries with minimal reconstruction error. Finally, the aligned vectors are aggregated in the shared orthogonal subspace and projected back to the original parameter space. Extensive experiments on vision and vision-language benchmarks show that DC-Merge consistently achieves state-of-the-art performance in both full fine-tuning and LoRA settings. The implementation code is available at https://github.com/Tobeginwith/DC-Merge.

---

### TaPD: Temporal-adaptive Progressive Distillation for Observation-Adaptive Trajectory Forecasting in Autonomous Driving

Authors: Mingyu Fan, Yi Liu, Hao Zhou, Deheng Qian, Mohammad Haziq Khan, Matthias Raetsch
Published: 2026-03-06
Categories: cs.CV, cs.AI, cs.RO
arXiv: https://arxiv.org/abs/2603.06231v1

Trajectory prediction is essential for autonomous driving, enabling vehicles to anticipate the motion of surrounding agents to support safe planning. However, most existing predictors assume fixed-length histories and suffer substantial performance degradation when observations are variable or extremely short in real-world settings (e.g., due to occlusion or a limited sensing range). We propose TaPD (Temporal-adaptive Progressive Distillation), a unified plug-and-play framework for observation-adaptive trajectory forecasting under variable history lengths. TaPD comprises two cooperative modules: an Observation-Adaptive Forecaster (OAF) for future prediction and a Temporal Backfilling Module (TBM) for explicit reconstruction of the past. OAF is built on progressive knowledge distillation (PKD), which transfers motion pattern knowledge from long-horizon "teachers" to short-horizon "students" via hierarchical feature regression, enabling short observations to recover richer motion context. We further introduce a cosine-annealed distillation weighting scheme to balance forecasting supervision and feature alignment, improving optimization stability and cross-length consistency. For extremely short histories where implicit alignment is insufficient, TBM backfills missing historical segments conditioned on scene evolution, producing context-rich trajectories that strengthen PKD and thereby improve OAF. We employ a decoupled pretrain-reconstruct-finetune protocol to preserve real-motion priors while adapting to backfilled inputs. Extensive experiments on Argoverse 1 and Argoverse 2 show that TaPD consistently outperforms strong baselines across all observation lengths, delivers especially large gains under very short inputs, and improves other predictors (e.g., HiVT) in a plug-and-play manner. Code will be available at https://github.com/zhouhao94/TaPD.

---

### Low-Latency Event-Based Object Detection with Spatially-Sparse Linear Attention

Authors: Haiqing Hao, Zhipeng Sui, Rong Zou, Zijia Dai, Nikola Zubić, Davide Scaramuzza, Wenhui Wang
Published: 2026-03-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2603.06228v4

Event cameras provide sequential visual data with spatial sparsity and high temporal resolution, making them attractive for low-latency object detection. Existing asynchronous event-based neural networks exploit this low-latency advantage by updating predictions event by event, but still suffer from two bottlenecks: recurrent architectures are difficult to train efficiently on long sequences, and improving accuracy often increases per-event computation and latency. Linear attention is appealing because it enables parallel training and recurrent inference. However, its dense state updates make per-event computation scale with the state size, yielding a poor accuracy-efficiency trade-off for object detection, where accurate localization requires fine-grained spatial states. The key challenge is therefore to introduce sparse state activation that exploits the spatial sparsity of events while preserving efficient parallel training. We propose Spatially-Sparse Linear Attention (SSLA), which introduces a mixture-of-spaces state decomposition and a scatter-compute-gather training procedure, enabling state-level sparsity as well as training parallelism. Building on SSLA, we develop an end-to-end asynchronous linear attention model, SSLA-Det, for low-latency event-based object detection. On Gen1 and N-Caltech101, SSLA-Det achieves state-of-the-art accuracy among asynchronous methods, reaching 0.375 mAP and 0.515 mAP, respectively, while reducing per-event computation by over 20 times compared with the strongest prior asynchronous baseline, demonstrating the potential of linear attention for low-latency event-based vision.

---

### Low-latency Event-based Object Detection with Spatially-Sparse Linear Attention

Authors: Haiqing Hao, Zhipeng Sui, Rong Zou, Zijia Dai, Nikola Zubić, Davide Scaramuzza, Wenhui Wang
Published: 2026-03-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2603.06228v3

Event cameras provide sequential visual data with spatial sparsity and high temporal resolution, making them attractive for low-latency object detection. Existing asynchronous event-based neural networks exploit this low-latency advantage by updating predictions event by event, but still suffer from two bottlenecks: recurrent architectures are difficult to train efficiently on long sequences, and improving accuracy often increases per-event computation and latency. Linear attention is appealing because it enables parallel training and recurrent inference. However, its dense state updates make per-event computation scale with the state size, yielding a poor accuracy-efficiency trade-off for object detection, where accurate localization requires fine-grained spatial states. The key challenge is therefore to introduce sparse state activation that exploits the spatial sparsity of events while preserving efficient parallel training. We propose Spatially-Sparse Linear Attention (SSLA), which introduces a mixture-of-spaces state decomposition and a scatter-compute-gather training procedure, enabling state-level sparsity as well as training parallelism. Building on SSLA, we develop an end-to-end asynchronous linear attention model, SSLA-Det, for low-latency event-based object detection. On Gen1 and N-Caltech101, SSLA-Det achieves state-of-the-art accuracy among asynchronous methods, reaching 0.375 mAP and 0.515 mAP, respectively, while reducing per-event computation by over 20 times compared with the strongest prior asynchronous baseline, demonstrating the potential of linear attention for low-latency event-based vision.

---

### FedSCS-XGB -- Federated Server-centric surrogate XGBoost for continual health monitoring

Authors: Felix Walger, Mehdi Ejtehadi, Anke Schmeink, Diego Paez-Granados
Published: 2026-03-06
Categories: cs.LG
arXiv: https://arxiv.org/abs/2603.06224v1

Wearable sensors with local data processing can detect health threats early, enhance documentation, and support personalized therapy. In the context of spinal cord injury (SCI), which involves risks such as pressure injuries and blood pressure instability, continuous monitoring can help mitigate these by enabling early deDtection and intervention. In this work, we present a novel distributed machine learning (DML) protocol for human activity recognition (HAR) from wearable sensor data based on gradient-boosted decision trees (XGBoost). The proposed architecture is inspired by Party-Adaptive XGBoost (PAX) while explicitly preserving key structural and optimization properties of standard XGBoost, including histogram-based split construction and tree-ensemble dynamics. First, we provide a theoretical analysis showing that, under appropriate data conditions and suitable hyperparameter selection, the proposed distributed protocol can converge to solutions equivalent to centralized XGBoost training. Second, the protocol is empirically evaluated on a representative wearable-sensor HAR dataset, reflecting the heterogeneity and data fragmentation typical of remote monitoring scenarios. Benchmarking against centralized XGBoost and IBM PAX demonstrates that the theoretical convergence properties are reflected in practice. The results indicate that the proposed approach can match centralized performance up to a gap under 1\% while retaining the structural advantages of XGBoost in distributed wearable-based HAR settings.

---

### SPOT: Span-level Pause-of-Thought for Efficient and Interpretable Latent Reasoning in Large Language Models

Authors: Yunlong Chu, Minglai Shao, Yuhang Liu, Bing Hao, Yumeng Lin, Jialu Wang, Ruijie Wang
Published: 2026-03-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2603.06222v1

Explicit Chain-of-Thought improves the reasoning performance of large language models but often incurs high inference cost due to verbose token-level traces. While recent approaches reduce this overhead via concise prompting or step pruning, they largely truncate what the model says rather than internalize what the model thinks. Latent reasoning offers a promising alternative by performing computation in the hidden space, yet prior methods face two critical challenges. Many existing approaches rely on rigid point-to-point alignment, forcing a latent token to approximate the final representation of a reasoning step, which can be insufficient to capture the dense, variable-length semantics of an entire reasoning segment. Furthermore, these methods often suffer from a lack of interpretability: latent states are commonly produced by unconstrained optimization or embedding mixing, yielding vectors that are difficult to decode or audit under the pretrained language head. We propose SPOT, a flexible framework that compresses explicit CoT into compact latent pause tokens without enforcing a fixed response template. At the core of SPOT is Span-level Semantic Alignment, a Sinkhorn optimal-transport objective that softly matches each pause token to the semantics of an entire reasoning segment, overcoming the rigidity of step-end alignment. To further improve interpretability, SPOT introduces a Frozen-Head Decoding Constraint that keeps latent states directly decodable as token distributions under the frozen pretrained LM head, enabling readable keyword interpretations of latent thoughts. Experiments on reasoning benchmarks demonstrate that SPOT improves accuracy by 2.3 points on average while reducing generated tokens by 37.5% and provides faithful semantic interpretations of the latent reasoning process.

---

### Word-Anchored Temporal Forgery Localization

Authors: Tianyi Wang, Xi Shao, Harry Cheng, Yinglong Wang, Mohan Kankanhalli
Published: 2026-03-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2603.06220v1

Current temporal forgery localization (TFL) approaches typically rely on temporal boundary regression or continuous frame-level anomaly detection paradigms to derive candidate forgery proposals. However, they suffer not only from feature granularity misalignment but also from costly computation. To address these issues, we propose word-anchored temporal forgery localization (WAFL), a novel paradigm that shifts the TFL task from temporal regression and continuous localization to discrete word-level binary classification. Specifically, we first analyze the essence of temporal forgeries and identify the minimum meaningful forgery units, word tokens, and then align data preprocessing with the natural linguistic boundaries of speech. To adapt powerful pre-trained foundation backbones for feature extraction, we introduce the forensic feature realignment (FFR) module, mapping representations from the pre-trained semantic space to a discriminative forensic manifold. This allows subsequent lightweight linear classifiers to efficiently perform binary classification and accomplish the TFL task. Furthermore, to overcome the extreme class imbalance inherent to forgery detection, we design the artifact-centric asymmetric (ACA) loss, which breaks the standard precision-recall trade-off by dynamically suppressing overwhelming authentic gradients while asymmetrically prioritizing subtle forensic artifacts. Extensive experiments demonstrate that WAFL significantly outperforms state-of-the-art approaches in localization performance under both in- and cross-dataset settings, while requiring substantially fewer learnable parameters and operating at high computational efficiency.

---

### Conversational Demand Response: Bidirectional Aggregator-Prosumer Coordination through Agentic AI

Authors: Reda El Makroum, Sebastian Zwickl-Bernhard, Lukas Kranzl, Hans Auer
Published: 2026-03-06
Categories: cs.AI, cs.MA, eess.SY
arXiv: https://arxiv.org/abs/2603.06217v1

Residential demand response depends on sustained prosumer participation, yet existing coordination is either fully automated, or limited to one-way dispatch signals and price alerts that offer little possibility for informed decision-making. This paper introduces Conversational Demand Response (CDR), a coordination mechanism where aggregators and prosumers interact through bidirectional natural language, enabled through agentic AI. A two-tier multi-agent architecture is developed in which an aggregator agent dispatches flexibility requests and a prosumer Home Energy Management System (HEMS) assesses deliverability and cost-benefit by calling an optimization-based tool. CDR also enables prosumer-initiated upstream communication, where changes in preferences can reach the aggregator directly. Proof-of-concept evaluation shows that interactions complete in under 12 seconds. The architecture illustrates how agentic AI can bridge the aggregator-prosumer coordination gap, providing the scalability of automated DR while preserving the transparency, explainability, and user agency necessary for sustained prosumer participation. All system components, including agent prompts, orchestration logic, and simulation interfaces, are released as open source to enable reproducibility and further development.

---

### EntON: Eigenentropy-Optimized Neighborhood Densification in 3D Gaussian Splatting

Authors: Miriam Jäger, Boris Jutzi
Published: 2026-03-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2603.06216v1

We present a novel Eigenentropy-optimized neighboorhood densification strategy EntON in 3D Gaussian Splatting (3DGS) for geometrically accurate and high-quality rendered 3D reconstruction. While standard 3DGS produces Gaussians whose centers and surfaces are poorly aligned with the underlying object geometry, surface-focused reconstruction methods frequently sacrifice photometric accuracy. In contrast to the conventional densification strategy, which relies on the magnitude of the view-space position gradient, our approach introduces a geometry-aware strategy to guide adaptive splitting and pruning. Specifically, we compute the 3D shape feature Eigenentropy from the eigenvalues of the covariance matrix in the k-nearest neighborhood of each Gaussian center, which quantifies the local structural order. These Eigenentropy values are integrated into an alternating optimization framework: During the optimization process, the algorithm alternates between (i) standard gradient-based densification, which refines regions via view-space gradients, and (ii) Eigenentropy-aware densification, which preferentially densifies Gaussians in low-Eigenentropy (ordered, flat) neighborhoods to better capture fine geometric details on the object surface, and prunes those in high-Eigenentropy (disordered, spherical) regions. We provide quantitative and qualitative evaluations on two benchmark datasets: small-scale DTU dataset and large-scale TUM2TWIN dataset, covering man-made objects and urban scenes. Experiments demonstrate that our Eigenentropy-aware alternating densification strategy improves geometric accuracy by up to 33% and rendering quality by up to 7%, while reducing the number of Gaussians by up to 50% and training time by up to 23%. Overall, EnTON achieves a favorable balance between geometric accuracy, rendering quality and efficiency by avoiding unnecessary scene expansion.

---

### Cut to the Chase: Training-free Multimodal Summarization via Chain-of-Events

Authors: Xiaoxing You, Qiang Huang, Lingyu Li, Xiaojun Chang, Jun Yu
Published: 2026-03-06
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2603.06213v1

Multimodal Summarization (MMS) aims to generate concise textual summaries by understanding and integrating information across videos, transcripts, and images. However, existing approaches still suffer from three main challenges: (1) reliance on domain-specific supervision, (2) implicit fusion with weak cross-modal grounding, and (3) flat temporal modeling without event transitions. To address these issues, we introduce **CoE**, a training-free MMS framework that performs structured reasoning through a **Chain-of-Events** guided by a Hierarchical Event Graph (HEG). The HEG encodes textual semantics into an explicit event hierarchy that scaffolds cross-modal grounding and temporal reasoning. Guided by this structure, **CoE** localizes key visual cues, models event evolution and causal transitions, and refines outputs via lightweight style adaptation for domain alignment. Extensive experiments on eight diverse datasets demonstrate that **CoE** consistently outperforms state-of-the-art video CoT baselines, achieving average gains of **+3.04 ROUGE**, **+9.51 CIDEr**, and **+1.88 BERTScore**, highlighting its robustness, interpretability, and cross-domain generalization. Our code is available at https://github.com/youxiaoxing/CoE.

---

### Topological descriptors of foot clearance gait dynamics improve differential diagnosis of Parkinsonism

Authors: Jhonathan Barrios, Wolfram Erlhagen, Miguel F. Gago, Estela Bicho, Flora Ferreira
Published: 2026-03-06
Categories: cs.LG, stat.AP
arXiv: https://arxiv.org/abs/2603.06212v1

Differential diagnosis among parkinsonian syndromes remains a clinical challenge due to overlapping motor symptoms and subtle gait abnormalities. Accurate differentiation is crucial for treatment planning and prognosis. While gait analysis is a well established approach for assessing motor impairments, conventional methods often overlook hidden nonlinear and structural features embedded in foot clearance patterns. We evaluated Topological Data Analysis (TDA) as a complementary tool for Parkinsonism classification using foot clearance time series. Persistent homology produced Betti curves, persistence landscapes, and silhouettes, which were used as features for a Random Forest classifier. The dataset comprised 15 controls (CO), 15 idiopathic Parkinson's disease (IPD), and 14 vascular Parkinsonism (VaP). Models were assessed with leave-one-out cross-validation (LOOCV). Betti-curve descriptors consistently yielded the strongest results. For IPD vs VaP, foot clearance variables minimum toe clearance, maximum toe late swing, and maximum heel clearance achieved 83% accuracy and AUC=0.89 under LOOCV in the medicated (On) state. Performance improved in the On state and further when both Off and On states were considered, indicating sensitivity of the topological features to levodopa related gait changes. These findings support integrating TDA with machine learning to improve clinical gait analysis and aid differential diagnosis across parkinsonian disorders.

---

### VG3S: Visual Geometry Grounded Gaussian Splatting for Semantic Occupancy Prediction

Authors: Xiaoyang Yan, Muleilan Pei, Shaojie Shen
Published: 2026-03-06
Categories: cs.CV, cs.RO
arXiv: https://arxiv.org/abs/2603.06210v2

3D semantic occupancy prediction has become a crucial perception task for comprehensive scene understanding in autonomous driving. While recent advances have explored 3D Gaussian splatting for occupancy modeling to substantially reduce computational overhead, the generation of high-quality 3D Gaussians relies heavily on accurate geometric cues, which are often insufficient in purely vision-centric paradigms. To bridge this gap, we advocate for injecting the strong geometric grounding capability from Vision Foundation Models (VFMs) into occupancy prediction. In this regard, we introduce Visual Geometry Grounded Gaussian Splatting (VG3S), a novel framework that empowers Gaussian-based occupancy prediction with cross-view 3D geometric grounding. Specifically, to fully exploit the rich 3D geometric priors from a frozen VFM, we propose a plug-and-play hierarchical geometric feature adapter, which can effectively transform generic VFM tokens via feature aggregation, task-specific alignment, and multi-scale restructuring. Extensive experiments on the nuScenes occupancy benchmark demonstrate that VG3S achieves remarkable improvements of 12.6% in IoU and 7.5% in mIoU over the baseline. Furthermore, we show that VG3S generalizes seamlessly across diverse VFMs, consistently enhancing occupancy prediction accuracy and firmly underscoring the immense value of integrating priors derived from powerful, pre-trained geometry-grounded VFMs.

---

### Point-Supervised Skeleton-Based Human Action Segmentation

Authors: Hongsong Wang, Yiqin Shen, Pengbo Yan, Jie Gui, Yang Zhang
Published: 2026-03-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2603.06201v2

Skeleton-based temporal action segmentation is a fundamental yet challenging task, playing a crucial role in enabling intelligent systems to perceive and respond to human activities. While fully-supervised methods achieve satisfactory performance, they require costly frame-level annotations and are sensitive to ambiguous action boundaries. To address these issues, we introduce a point-supervised framework for skeleton-based action segmentation, where only a single frame per action segment is labeled. We leverage multimodal skeleton data, including joint, bone, and motion information, encoded via a pretrained unified model to extract rich feature representations. To generate reliable pseudo-labels, we propose a novel prototype similarity method and integrate it with two existing methods: energy function and constrained K-Medoids clustering. Multimodal pseudo-label integration is proposed to enhance the reliability of the pseudo-label and guide the model training. We establish new benchmarks on PKU-MMD (X-Sub and X-View), MCFS-22, and MCFS-130, and implement baselines for point-supervised skeleton-based human action segmentation. Extensive experiments show that our method achieves competitive performance, even surpassing some fully-supervised methods while significantly reducing annotation effort.

---

### Point-Supervised Skeleton-Based Human Action Segmentation

Authors: Hongsong Wang, Yiqin Shen, Pengbo Yan, Jie Gui
Published: 2026-03-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2603.06201v1

Skeleton-based temporal action segmentation is a fundamental yet challenging task, playing a crucial role in enabling intelligent systems to perceive and respond to human activities. While fully-supervised methods achieve satisfactory performance, they require costly frame-level annotations and are sensitive to ambiguous action boundaries. To address these issues, we introduce a point-supervised framework for skeleton-based action segmentation, where only a single frame per action segment is labeled. We leverage multimodal skeleton data, including joint, bone, and motion information, encoded via a pretrained unified model to extract rich feature representations. To generate reliable pseudo-labels, we propose a novel prototype similarity method and integrate it with two existing methods: energy function and constrained K-Medoids clustering. Multimodal pseudo-label integration is proposed to enhance the reliability of the pseudo-label and guide the model training. We establish new benchmarks on PKU-MMD (X-Sub and X-View), MCFS-22, and MCFS-130, and implement baselines for point-supervised skeleton-based human action segmentation. Extensive experiments show that our method achieves competitive performance, even surpassing some fully-supervised methods while significantly reducing annotation effort.

---

### Adaptive Language-Aware Image Reflection Removal Network

Authors: Siyan Fang, Yuntao Wang, Jinpu Zhang, Ziwen Li, Yuehuan Wang
Published: 2026-03-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2603.06200v1

Existing image reflection removal methods struggle to handle complex reflections. Accurate language descriptions can help the model understand the image content to remove complex reflections. However, due to blurred and distorted interferences in reflected images, machine-generated language descriptions of the image content are often inaccurate, which harms the performance of language-guided reflection removal. To address this, we propose the Adaptive Language-Aware Network (ALANet) to remove reflections even with inaccurate language inputs. Specifically, ALANet integrates both filtering and optimization strategies. The filtering strategy reduces the negative effects of language while preserving its benefits, whereas the optimization strategy enhances the alignment between language and visual features. ALANet also utilizes language cues to decouple specific layer content from feature maps, improving its ability to handle complex reflections. To evaluate the model's performance under complex reflections and varying levels of language accuracy, we introduce the Complex Reflection and Language Accuracy Variance (CRLAV) dataset. Experimental results demonstrate that ALANet surpasses state-of-the-art methods for image reflection removal. The code and dataset are available at https://github.com/fashyon/ALANet.

---

### FlashPrefill: Instantaneous Pattern Discovery and Thresholding for Ultra-Fast Long-Context Prefilling

Authors: Qihang Fan, Huaibo Huang, Zhiying Wu, Juqiu Wang, Bingning Wang, Ran He
Published: 2026-03-06
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2603.06199v1

Long-context modeling is a pivotal capability for Large Language Models, yet the quadratic complexity of attention remains a critical bottleneck, particularly during the compute-intensive prefilling phase. While various sparse attention mechanisms have been explored, they typically suffer from either significant search latency or insufficient sparsity. In this paper, we propose FlashPrefill, a framework enabling ultra-fast prefilling via instantaneous pattern discovery and thresholding. FlashPrefill leverages a fast block-searching technique to simultaneously locate dynamic vertical, slash, and block-sparse attention patterns. Crucially, it introduces a dynamic thresholding mechanism that bypasses the prohibitive overhead of sorting or accumulating attention scores while effectively eliminating the long-tail distribution to enhance sparsity. Extensive evaluations demonstrate that FlashPrefill achieves a substantial leap in efficiency, delivering an unprecedented 27.78x speedup on 256K sequences. Notably, unlike existing methods that incur efficiency degradation on shorter contexts, FlashPrefill maintains a 1.71x speedup even at a 4K context length, demonstrating its robustness and practical utility across varying sequence scales.

---

### LIT-RAGBench: Benchmarking Generator Capabilities of Large Language Models in Retrieval-Augmented Generation

Authors: Koki Itai, Shunichi Hasegawa, Yuta Yamamoto, Gouki Minegishi, Masaki Otsuki
Published: 2026-03-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2603.06198v2

Retrieval-Augmented Generation (RAG) is a framework in which a Generator, such as a Large Language Model (LLM), produces answers by retrieving documents from an external collection using a Retriever. In practice, Generators must integrate evidence from long contexts, perform multi-step reasoning, interpret tables, and abstain when evidence is missing. However, existing benchmarks for Generators provide limited coverage, with none enabling simultaneous evaluation of multiple capabilities under unified conditions. To bridge the gap between existing evaluations and practical use, we introduce LIT-RAGBench (the Logic, Integration, Table, Reasoning, and Abstention RAG Generator Benchmark), which defines five categories: Integration, Reasoning, Logic, Table, and Abstention, each further divided into practical evaluation aspects. LIT-RAGBench systematically covers patterns combining multiple aspects across categories. By using fictional entities and scenarios, LIT-RAGBench evaluates answers grounded in the provided external documents. The dataset consists of 114 human-constructed Japanese questions and an English version generated by machine translation with human curation. We use LLM-as-a-Judge for scoring and report category-wise and overall accuracy. Across API-based and open-weight models, no model exceeds 90% overall accuracy. By making strengths and weaknesses measurable within each category, LIT-RAGBench serves as a valuable metric for model selection in practical RAG deployments and for building RAG-specialized models. We release LIT-RAGBench, including the dataset and evaluation code, at https://github.com/Koki-Itai/LIT-RAGBench.

---

### Wisdom of the AI Crowd (AI-CROWD) for Ground Truth Approximation in Content Analysis: A Research Protocol & Validation Using Eleven Large Language Models

Authors: Luis de-Marcos, Manuel Goyanes, Adrián Domínguez-Díaz
Published: 2026-03-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2603.06197v1

Large-scale content analysis is increasingly limited by the absence of observable ground truth or gold-standard labels, as creating such benchmarks through extensive human coding becomes impractical for massive datasets due to high time, cost, and consistency challenges. To overcome this barrier, we introduce the AI-CROWD protocol, which approximates ground truth by leveraging the collective outputs of an ensemble of large language models (LLMs). Rather than asserting that the resulting labels are true ground truth, the protocol generates a consensus-based approximation derived from convergent and divergent inferences across multiple models. By aggregating outputs via majority voting and interrogating agreement/disagreement patterns with diagnostic metrics, AI-CROWD identifies high-confidence classifications while flagging potential ambiguity or model-specific biases.

---

### MICA: Multi-granularity Intertemporal Credit Assignment for Long-Horizon Emotional Support Dialogue

Authors: Naifan Zhang, Ruihan Sun, Jinwei Su, Hengjie Yang, Zhengyuan Pan, Zhaohan Chen, Xiaofan Zhang
Published: 2026-03-06
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2603.06194v3

Reinforcement learning (RL) for large language models (LLMs) has shown strong performance in single-turn tasks, but extending it to multi-turn interaction remains challenging due to sparse rewards and poor per-turn credit assignment. In emotional support dialogues, responses shape future user states, so matched-state step-wise comparison is unavailable, while trajectory-level supervision is insufficient. We propose MICA (Multi-granularity Intertemporal Credit Assignment), a critic-free RL framework for multi-turn emotional support tasks. MICA derives both immediate and delayed credit from a shared potential function over the user's structured support state. Incremental Distance Reward measures the per-turn decrease in residual distance to the target state, while its Monte Carlo return captures delayed effects. After scope-specific normalization, the two signals form a mixed advantage for stable per-turn optimization without matched-state comparisons, rollout trees, or a learned critic. On EMPA, EQ-Bench, and EmoBench with Qwen2.5-7B-Instruct and Qwen3-8B/14B/32B, MICA consistently outperforms GRPO and REINFORCE++, achieving up to +42.5 on EMPA, while adding no rollout cost and remaining robust to reward judges. These results show that turn-aware credit assignment enables effective and practical multi-turn RL for interactive LLMs.

---

### Whisper-CD: Accurate Long-Form Speech Recognition using Multi-Negative Contrastive Decoding

Authors: Hoseong Ahn, Jeongyun Chae, Yoonji Park, Kyuhong Shim
Published: 2026-03-06
Categories: cs.SD, cs.AI, eess.AS
arXiv: https://arxiv.org/abs/2603.06193v2

Long-form speech recognition with large encoder-decoder models such as Whisper often exhibit hallucinations, repetition loops, and content omissions. These errors can accumulate and be further amplified when the previous segment's transcription is used as decoding context. We propose Whisper-CD, a training-free contrastive decoding framework that contrasts clean-audio logits against negative logits computed from three acoustically motivated perturbations: Gaussian noise injection, silence signal, and audio temporal shift. We aggregate these negatives via the log-sum-exp operator, building a unified multi-negative objective for token-by-token decoding. Across five English long-form benchmarks, Whisper-CD reduces WER by up to 24.3pp on CORAAL and shows 48% faster token generation throughput than beam search. Because Whisper-CD operates purely at inference time, it can be applied as a drop-in replacement to already-deployed Whisper systems without retraining.

---

### Random Quadratic Form on a Sphere: Synchronization by Common Noise

Authors: Maximilian Engel, Anna Shalova
Published: 2026-03-06
Categories: math.PR, cs.LG, math.DS
arXiv: https://arxiv.org/abs/2603.06187v2

We introduce the Random Quadratic Form (RQF): a stochastic differential equation which formally corresponds to the gradient flow of a random quadratic functional on a sphere. While the one-point dynamics of the system is a Brownian motion and thus has no preferred direction, the two-point motion exhibits nontrivial synchronizing behaviour. In this work we study synchronization of the RQF, namely we give both distributional and path-wise characterizations of the solutions by studying invariant measures and random attractors of the system. The RQF model is motivated by the study of the role of linear layers in transformers and illustrates the synchronization by common noise phenomena arising in the simplified models of transformers. In particular, we provide an alternative (independent of self-attention) explanation of the clustering behaviour in deep transformers and show that tokens cluster even in the absence of the self-attention mechanism.

---