# Research Papers — 2026-03 · bundle page 201 of 243 > Full-text markdown of papers #10001–10050 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/201 [← Previous bundle](https://aipod.dev/papers/archive/2026-03/page/200.md) [Next bundle →](https://aipod.dev/papers/archive/2026-03/page/202.md) ### BlackMirror: Black-Box Backdoor Detection for Text-to-Image Models via Instruction-Response Deviation Authors: Feiran Li, Qianqian Xu, Shilong Bao, Zhiyong Yang, Xilin Zhao, Xiaochun Cao, Qingming Huang Published: 2026-03-06 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2603.05921v1 This paper investigates the challenging task of detecting backdoored text-to-image models under black-box settings and introduces a novel detection framework BlackMirror. Existing approaches typically rely on analyzing image-level similarity, under the assumption that backdoor-triggered generations exhibit strong consistency across samples. However, they struggle to generalize to recently emerging backdoor attacks, where backdoored generations can appear visually diverse. BlackMirror is motivated by an observation: across backdoor attacks, {only partial semantic patterns within the generated image are steadily manipulated, while the rest of the content remains diverse or benign. Accordingly, BlackMirror consists of two components: MirrorMatch, which aligns visual patterns with the corresponding instructions to detect semantic deviations; and MirrorVerify, which evaluates the stability of these deviations across varied prompts to distinguish true backdoor behavior from benign responses. BlackMirror is a general, training-free framework that can be deployed as a plug-and-play module in Model-as-a-Service (MaaS) applications. Comprehensive experiments demonstrate that BlackMirror achieves accurate detection across a wide range of attacks. Code is available at https://github.com/Ferry-Li/BlackMirror. --- ### Design Experiments to Compare Multi-armed Bandit Algorithms Authors: Huiling Meng, Ningyuan Chen, Xuefeng Gao Published: 2026-03-06 Categories: cs.LG, math.ST, stat.ML arXiv: https://arxiv.org/abs/2603.05919v2 Online platforms routinely compare multi-armed bandit algorithms, such as UCB and Thompson Sampling, to select the best-performing policy. Unlike standard A/B tests for static treatments, each run of a bandit algorithm over $T$ users produces only one trajectory, because the algorithm's decisions depend on all past interactions. Reliable inference therefore demands many independent restarts of the algorithm, making experimentation costly and delaying deployment decisions. We propose Artificial Replay (AR) as a new experimental design for this problem. AR first runs one policy and records its trajectory. When the second policy is executed, it reuses a recorded reward whenever it selects an action the first policy already took, and queries the real environment only otherwise. We develop a new analytical framework for this design and prove three key properties of the resulting estimator: it is unbiased; it requires only $T + o(T)$ user interactions instead of $2T$ for a run of the treatment and control policies, nearly halving the experimental cost when both policies have sub-linear regret; and its variance grows sub-linearly in $T$, whereas the estimator from a naïve design has a linearly-growing variance. Numerical experiments with UCB, Thompson Sampling, and $ε$-greedy policies confirm these theoretical gains. --- ### Stock Market Prediction Using Node Transformer Architecture Integrated with BERT Sentiment Analysis Authors: Mohammad Al Ridhawi, Mahtab Haj Ali, Hussein Al Osman Published: 2026-03-06 Categories: cs.LG, cs.AI, q-fin.ST arXiv: https://arxiv.org/abs/2603.05917v3 Stock market prediction presents considerable challenges for investors, financial institutions, and policymakers operating in complex market environments characterized by noise, non-stationarity, and behavioral dynamics. Traditional forecasting methods, including fundamental analysis and technical indicators, often fail to capture the intricate patterns and cross-sectional dependencies inherent in financial markets. This paper presents an integrated framework combining a node transformer architecture with BERT-based sentiment analysis for stock price forecasting. The proposed model represents the stock market as a graph structure where individual stocks form nodes and edges capture relationships including sectoral affiliations, correlated price movements, and supply chain connections. A fine-tuned BERT model extracts sentiment information from social media posts and combines it with quantitative market features through attention-based fusion mechanisms. The node transformer processes historical market data while capturing both temporal evolution and cross-sectional dependencies among stocks. Experiments conducted on 20 S&P 500 stocks spanning January 1982 to March 2025 demonstrate that the integrated model achieves a mean absolute percentage error (MAPE) of 0.80% for one-day-ahead predictions, compared to 1.20% for ARIMA and 1.00% for LSTM. The inclusion of sentiment analysis reduces prediction error by 10% overall and 25% during earnings announcements, while the graph-based architecture contributes an additional 15% improvement by capturing inter-stock dependencies. Directional accuracy reaches 65% for one-day forecasts. Statistical validation through paired t-tests confirms the significance of these improvements (p < 0.05 for all comparisons). The model maintains lower error during high-volatility periods, achieving MAPE of 1.50% while baseline models range from 1.60% to 2.10%. --- ### DeepFact: Co-Evolving Benchmarks and Agents for Deep Research Factuality Authors: Yukun Huang, Leonardo F. R. Ribeiro, Momchil Hardalov, Bhuwan Dhingra, Markus Dreyer, Venkatesh Saligrama Published: 2026-03-06 Categories: cs.AI arXiv: https://arxiv.org/abs/2603.05912v2 Search-augmented LLM agents can produce deep research reports (DRRs), but verifying claim-level factuality remains challenging. Existing fact-checkers are primarily designed for general-domain, factoid-style atomic claims, and there is no benchmark to test whether such verifiers transfer to DRRs. Yet building such a benchmark is itself difficult. We first show that static expert-labeled benchmarks are brittle in this setting: in a controlled study with PhD-level specialists, unassisted experts achieve only 60.8% accuracy on a hidden micro-gold set of verifiable claims. We propose Evolving Benchmarking via Audit-then-Score (AtS), where benchmark labels and rationales are explicitly revisable: when a verifier disagrees with the current benchmark, it must submit evidence; an auditor adjudicates the dispute; and accepted revisions update the benchmark before models are scored. Across four AtS rounds, expert micro-gold accuracy rises to 90.9%, indicating experts are substantially more reliable as auditors than as one-shot labelers. We instantiate AtS as DeepFact-Bench, a versioned DRR factuality benchmark with auditable rationales, and DeepFact-Eval, a document-level verification agent (with a grouped lite variant) that outperforms existing verifiers on DeepFact-Bench and transfers well to external factuality datasets. --- ### CORE-Seg: Reasoning-Driven Segmentation for Complex Lesions via Reinforcement Learning Authors: Yuxin Xie, Yuming Chen, Yishan Yang, Yi Zhou, Tao Zhou, Zhen Zhao, Jiacheng Liu, Huazhu Fu Published: 2026-03-06 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2603.05911v1 Medical image segmentation is undergoing a paradigm shift from conventional visual pattern matching to cognitive reasoning analysis. Although Multimodal Large Language Models (MLLMs) have shown promise in integrating linguistic and visual knowledge, significant gaps remain: existing general MLLMs possess broad common sense but lack the specialized visual reasoning required for complex lesions, whereas traditional segmentation models excel at pixel-level segmentation but lack logical interpretability. In this paper, we introduce ComLesion-14K, the first diverse Chain-of-Thought (CoT) benchmark for reasoning-driven complex lesion segmentation. To accomplish this task, we propose CORE-Seg, an end-to-end framework integrating reasoning with segmentation through a Semantic-Guided Prompt Adapter. We design a progressive training strategy from SFT to GRPO, equipped with an adaptive dual-granularity reward mechanism to mitigate reward sparsity. Our Method achieves state-of-the-art results with a mean Dice of 37.06\% (14.89\% higher than the second-best baseline), while reducing the failure rate to 18.42\%. Project Page: https://xyxl024.github.io/CORE-Seg.github.io/ --- ### The World Won't Stay Still: Programmable Evolution for Agent Benchmarks Authors: Guangrui Li, Yaochen Xie, Yi Liu, Ziwei Dong, Xingyuan Pan, Tianqi Zheng, Jason Choi, Michael J. Morais, Binit Jha, Shaunak Mishra, Bingrou Zhou, Chen Luo et al. Published: 2026-03-06 Categories: cs.AI arXiv: https://arxiv.org/abs/2603.05910v2 LLM-powered tool-calling agents fulfill user requests by interacting with environments, querying data, and invoking tools in a multi-turn process. Yet, most existing benchmarks evaluate these systems under static environment interfaces, with fixed schemas and toolsets, making it difficult to assess how agents behave as environments evolves -- when capabilities are added, reorganized, or deprecated across successive environment versions. In this paper, we study structured environment evolution as a benchmark-construction problem for tool-calling agents. We propose ProEvolve, a graph-based framework that makes environment evolution programmable. At its core, a typed relational graph provides a unified, explicit representation of the environment - data, tools, and schema. Under this formalism, adding, removing, or modifying capabilities are expressed as graph transformations that coherently propagate updates across tools, schemas, and data access. Building on this, ProEvolve supports (1) automatic generation of evolved executable environments through explicit graph transformations, and (2) graph-grounded construction of task sandboxes via subgraph sampling and instantiation. We validate ProEvolve in two tool-calling domains, e-commerce and airline booking, in terms of quality, implementation validity, and failure modes. Finally, we use the generated benchmark as a downstream diagnostic to study how representative agents behave under structured environment evolution. --- ### InfoGatherer: Principled Information Seeking via Evidence Retrieval and Strategic Questioning Authors: Maksym Taranukhin, Shuyue Stella Li, Evangelos Milios, Geoff Pleiss, Yulia Tsvetkov, Vered Shwartz Published: 2026-03-06 Categories: cs.CL arXiv: https://arxiv.org/abs/2603.05909v1 LLMs are increasingly deployed in high-stakes domains such as medical triage and legal assistance, often as document-grounded QA systems in which a user provides a description, relevant sources are retrieved, and an LLM generates a prediction. In practice, initial user queries are often underspecified, and a single retrieval pass is insufficient for reliable decision-making, leading to incorrect and overly confident answers. While follow-up questioning can elicit missing information, existing methods typically depend on implicit, unstructured confidence signals from the LLM, making it difficult to determine what remains unknown, what information matters most, and when to stop asking questions. We propose InfoGatherer, a framework that gathers missing information from two complementary sources: retrieved domain documents and targeted follow-up questions to the user. InfoGatherer models uncertainty using Dempster-Shafer belief assignments over a structured evidential network, enabling principled fusion of incomplete and potentially contradictory evidence from both sources without prematurely collapsing to a definitive answer. Across legal and medical tasks, InfoGatherer outperforms strong baselines while requiring fewer turns. By grounding uncertainty in formal evidential theory rather than heuristic LLM signals, InfoGatherer moves towards trustworthy, interpretable decision support in domains where reliability is critical. --- ### Pano3DComposer: Feed-Forward Compositional 3D Scene Generation from Single Panoramic Image Authors: Zidian Qiu, Ancong Wu Published: 2026-03-06 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.05908v1 Current compositional image-to-3D scene generation approaches construct 3D scenes by time-consuming iterative layout optimization or inflexible joint object-layout generation. Moreover, most methods rely on limited field-of-view perspective images, hindering the creation of complete 360-degree environments. To address these limitations, we design Pano3DComposer, an efficient feed-forward framework for panoramic images. To decouple object generation from layout estimation, we propose a plug-and-play Object-World Transformation Predictor. This module converts the 3D objects generated by off-the-shelf image-to-3D models from local to world coordinates. To achieve this, we adapt the VGGT architecture to Alignment-VGGT by using target object crop, multi-view object renderings and camera parameters to predict the transformation. The predictor is trained using pseudo-geometric supervision to address the shape discrepancy between generated and ground-truth objects. For input images from unseen domains, we further introduce a Coarse-to-Fine (C2F) alignment mechanism for Pano3DComposer that iteratively refines geometric consistency with feedback of scene rendering. Our method achieves superior geometric accuracy for image/text-to-3D tasks on synthetic and real-world datasets. It can generate a high-fidelity 3D scene in approximately 20 seconds on an RTX 4090 GPU. Project page: https://qiuzidian.github.io/pano3dcomposer-page/. --- ### Beyond Geometry: Artistic Disparity Synthesis for Immersive 2D-to-3D Authors: Ping Chen, Zezhou Chen, Xingpeng Zhang, Yanlin Qian, Huan Hu, Xiang Liu, Zipeng Wang, Xin Wang, Zhaoxiang Liu, Kai Wang, Shiguo Lian Published: 2026-03-06 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.05906v1 Current 2D-to-3D conversion methods achieve geometric accuracy but are artistically deficient, failing to replicate the immersive and emotionally resonant experience of professional 3D cinema. This is because geometric reconstruction paradigms mistake deliberate artistic intent, such as strategic zero-plane shifts for pop-out effects and local depth sculpting, for data noise or ambiguity. This paper argues for a new paradigm: Artistic Disparity Synthesis, shifting the goal from physically accurate disparity estimation to artistically coherent disparity synthesis. We propose Art3D, a preliminary framework exploring this paradigm. Art3D uses a dual-path architecture to decouple global depth parameters (macro-intent) from local artistic effects (visual brushstrokes) and learns from professional 3D film data via indirect supervision. We also introduce a preliminary evaluation method to quantify cinematic alignment. Experiments show our approach demonstrates potential in replicating key local out-of-screen effects and aligning with the global depth styles of cinematic 3D content, laying the groundwork for a new class of artistically-driven conversion tools. --- ### CollabOD: Collaborative Multi-Backbone with Cross-scale Vision for UAV Small Object Detection Authors: Xuecheng Bai, Yuxiang Wang, Chuanzhi Xu, Kang Han, Jun Guo, Pengfei Ye Published: 2026-03-06 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.05905v2 Small object detection in unmanned aerial vehicle (UAV) imagery is challenging because high-altitude viewpoints produce severe scale variation, weak structural cues, and tight computational budgets. Existing lightweight detectors usually fuse multi-scale features after downsampling, where boundary and texture details have already been attenuated and heterogeneous feature streams may be spatially misaligned. To address these issues, we propose CollabOD, a collaborative detection framework that preserves structural details, aligns cross-path features before fusion, and keeps the detection head lightweight at inference time. CollabOD combines a Dual-Path Fusion Stem, a Dense Aggregation Block, a Bilateral Reweighting Module, and a Unified Detail-Aware Head to strengthen localization-oriented representation while limiting extra computation. On VisDrone, CollabOD obtains 52.4 AP50, 30.8 AP75, and 29.9 AP50:95 with 65.5 GFLOPs; on UAVDT it reaches 31.2 AP50 and 17.4 AP50:95; and on AI-TOD it reaches 45.4 AP50 and 20.0 AP50:95 at 137 FPS. The code is available at: https://github.com/Bai-Xuecheng/CollabOD. --- ### LUMINA: LLM-Guided GPU Architecture Exploration via Bottleneck Analysis Authors: Tao Zhang, Rui Ma, Shuotao Xu, Yongqiang Xiong, Peng Cheng Published: 2026-03-06 Categories: cs.AR, cs.AI arXiv: https://arxiv.org/abs/2603.05904v2 GPU design space exploration (DSE) for modern AI workloads, such as Large-Language Model (LLM) inference, is challenging because of GPUs' vast, multi-modal design spaces, high simulation costs, and complex design optimization objectives (e.g. performance, power and area trade-offs). Existing automated DSE methods are often prohibitively expensive, either requiring an excessive number of exploration samples or depending on intricate, manually crafted analyses of interdependent critical paths guided by human heuristics. We present LUMINA, an LLM-driven GPU architecture exploration framework that leverage AI to enhance the DSE efficiency and efficacy for GPUs. LUMINA extracts architectural knowledge from simulator code and performs sensitivity studies to automatically compose DSE rules,which are auto-corrected during exploration. A core component of LUMINA is a DSE Benchmark that comprehensively evaluates and enhances LLMs' capabilities across three fundamental skills required for architecture optimization, which provides a principled and reproducible basis for model selection and ensuring consistent architectural reasoning. In the design space with 4.7 million possible samples, LUMINA identifies 6 designs of better performance and area than an A100 GPU efficiently, using only 20 steps via LLM-assisted bottleneck analysis. In comparison, LUMINA achieves 17.5x higher than design space exploration efficiency, and 32.9% better designs (i.e. Pareto Hypervolume) than Machine-Learning baselines, showcasing its ability to deliver high-quality design guidance with minimal search cost. --- ### Reference-guided Policy Optimization for Molecular Optimization via LLM Reasoning Authors: Xuan Li, Zhanke Zhou, Zongze Li, Jiangchao Yao, Yu Rong, Lu Zhang, Bo Han Published: 2026-03-06 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2603.05900v1 Large language models (LLMs) benefit substantially from supervised fine-tuning (SFT) and reinforcement learning with verifiable rewards (RLVR) in reasoning tasks. However, these recipes perform poorly in instruction-based molecular optimization, where each data point typically provides only a single optimized reference molecule and no step-by-step optimization trajectory. We reveal that answer-only SFT on the reference molecules collapses reasoning, and RLVR provides sparse feedback under similarity constraints due to the model's lack of effective exploration, which slows learning and limits optimization. To encourage the exploration of new molecules while balancing the exploitation of the reference molecules, we introduce Reference-guided Policy Optimization (RePO), an optimization approach that learns from reference molecules without requiring trajectory data. At each update, RePO samples candidate molecules with their intermediate reasoning trajectories from the model and trains the model using verifiable rewards that measure property satisfaction under similarity constraints in an RL manner. Meanwhile, it applies reference guidance by keeping the policy's intermediate reasoning trajectory as context and training only the answer in a supervised manner. Together, the RL term promotes exploration, while the guidance term mitigates reward sparsity and stabilizes training by grounding outputs to references when many valid molecular edits exist. Across molecular optimization benchmarks, RePO consistently outperforms SFT and RLVR baselines (e.g., GRPO), achieving improvements on the optimization metric (Success Rate $\times$ Similarity), improving balance across competing objectives, and generalizing better to unseen instruction styles. Our code is publicly available at https://github.com/tmlr-group/RePO. --- ### Mitigating Bias in Concept Bottleneck Models for Fair and Interpretable Image Classification Authors: Schrasing Tong, Antoine Salaun, Vincent Yuan, Annabel Adeyeri, Lalana Kagal Published: 2026-03-06 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2603.05899v1 Ensuring fairness in image classification prevents models from perpetuating and amplifying bias. Concept bottleneck models (CBMs) map images to high-level, human-interpretable concepts before making predictions via a sparse, one-layer classifier. This structure enhances interpretability and, in theory, supports fairness by masking sensitive attribute proxies such as facial features. However, CBM concepts have been known to leak information unrelated to concept semantics and early results reveal only marginal reductions in gender bias on datasets like ImSitu. We propose three bias mitigation techniques to improve fairness in CBMs: 1. Decreasing information leakage using a top-k concept filter, 2. Removing biased concepts, and 3. Adversarial debiasing. Our results outperform prior work in terms of fairness-performance tradeoffs, indicating that our debiased CBM provides a significant step towards fair and interpretable image classification. --- ### InnoAds-Composer: Efficient Condition Composition for E-Commerce Poster Generation Authors: Yuxin Qin, Ke Cao, Haowei Liu, Ao Ma, Fengheng Li, Honghe Zhu, Zheng Zhang, Run Ling, Wei Feng, Xuanhua He, Zhanjie Zhang, Zhen Guo et al. Published: 2026-03-06 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.05898v1 E-commerce product poster generation aims to automatically synthesize a single image that effectively conveys product information by presenting a subject, text, and a designed style. Recent diffusion models with fine-grained and efficient controllability have advanced product poster synthesis, yet they typically rely on multi-stage pipelines, and simultaneous control over subject, text, and style remains underexplored. Such naive multi-stage pipelines also show three issues: poor subject fidelity, inaccurate text, and inconsistent style. To address these issues, we propose InnoAds-Composer, a single-stage framework that enables efficient tri-conditional control tokens over subject, glyph, and style. To alleviate the quadratic overhead introduced by naive tri-conditional token concatenation, we perform importance analysis over layers and timesteps and route each condition only to the most responsive positions, thereby shortening the active token sequence. Besides, to improve the accuracy of Chinese text rendering, we design a Text Feature Enhancement Module (TFEM) that integrates features from both glyph images and glyph crops. To support training and evaluation, we also construct a high-quality e-commerce product poster dataset and benchmark, which is the first dataset that jointly contains subject, text, and style conditions. Extensive experiments demonstrate that InnoAds-Composer significantly outperforms existing product poster methods without obviously increasing inference latency. --- ### Building an Ensemble LLM Semantic Tagger for UN Security Council Resolutions Authors: Hussein Ghaly Published: 2026-03-06 Categories: cs.CL arXiv: https://arxiv.org/abs/2603.05895v1 This paper introduces a new methodology for using LLM-based systems for accurate and efficient semantic tagging of UN Security Council resolutions. The main goal is to leverage LLM performance variability to build ensemble systems for data cleaning and semantic tagging tasks. We introduce two evaluation metrics: Content Preservation Ratio (CPR) and Tag Well-Formedness (TWF), in order to avoid hallucinations and unnecessary additions or omissions to the input text beyond the task requirement. These metrics allow the selection of the best output from multiple runs of several GPT models. GPT-4.1 achieved the highest metrics for both tasks (Cleaning: CPR 84.9% - Semantic Tagging: CPR 99.99% and TWF 99.92%). In terms of cost, smaller models, such as GPT-4.1-mini, achieved comparable performance to the best model in each task at only 20% of the cost. These metrics ultimately allowed the ensemble to select the optimal output (both cleaned and tagged content) for all the LLM models involved, across multiple runs. With this ensemble design and the use of metrics, we create a reliable LLM system for performing semantic tagging on challenging texts. --- ### Lost in Stories: Consistency Bugs in Long Story Generation by LLMs Authors: Junjie Li, Xinrui Guo, Yuhao Wu, Roy Ka-Wei Lee, Hongzhi Li, Yutao Xie Published: 2026-03-06 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2603.05890v1 What happens when a storyteller forgets its own story? Large Language Models (LLMs) can now generate narratives spanning tens of thousands of words, but they often fail to maintain consistency throughout. When generating long-form narratives, these models can contradict their own established facts, character traits, and world rules. Existing story generation benchmarks focus mainly on plot quality and fluency, leaving consistency errors largely unexplored. To address this gap, we present ConStory-Bench, a benchmark designed to evaluate narrative consistency in long-form story generation. It contains 2,000 prompts across four task scenarios and defines a taxonomy of five error categories with 19 fine-grained subtypes. We also develop ConStory-Checker, an automated pipeline that detects contradictions and grounds each judgment in explicit textual evidence. Evaluating a range of LLMs through five research questions, we find that consistency errors show clear tendencies: they are most common in factual and temporal dimensions, tend to appear around the middle of narratives, occur in text segments with higher token-level entropy, and certain error types tend to co-occur. These findings can inform future efforts to improve consistency in long-form narrative generation. Our project page is available at https://picrew.github.io/constory-bench.github.io/. --- ### PixARMesh: Autoregressive Mesh-Native Single-View Scene Reconstruction Authors: Xiang Zhang, Sohyun Yoo, Hongrui Wu, Chuan Li, Jianwen Xie, Zhuowen Tu Published: 2026-03-06 Categories: cs.CV, cs.GR, cs.LG arXiv: https://arxiv.org/abs/2603.05888v1 We introduce PixARMesh, a method to autoregressively reconstruct complete 3D indoor scene meshes directly from a single RGB image. Unlike prior methods that rely on implicit signed distance fields and post-hoc layout optimization, PixARMesh jointly predicts object layout and geometry within a unified model, producing coherent and artist-ready meshes in a single forward pass. Building on recent advances in mesh generative models, we augment a point-cloud encoder with pixel-aligned image features and global scene context via cross-attention, enabling accurate spatial reasoning from a single image. Scenes are generated autoregressively from a unified token stream containing context, pose, and mesh, yielding compact meshes with high-fidelity geometry. Experiments on synthetic and real-world datasets show that PixARMesh achieves state-of-the-art reconstruction quality while producing lightweight, high-quality meshes ready for downstream applications. --- ### Reconstruct! Don't Encode: Self-Supervised Representation Reconstruction Loss for High-Intelligibility and Low-Latency Streaming Neural Audio Codec Authors: Junhyeok Lee, Xiluo He, Jihwan Lee, Helin Wang, Shrikanth Narayanan, Thomas Thebaud, Laureano Moro-Velazquez, Jesús Villalba, Najim Dehak Published: 2026-03-06 Categories: eess.AS, cs.AI arXiv: https://arxiv.org/abs/2603.05887v2 Neural audio codecs optimized for mel-spectrogram reconstruction often fail to preserve intelligibility. While semantic encoder distillation improves encoded representations, it does not guarantee content preservation in reconstructed speech. In this work, we demonstrate that self-supervised representation reconstruction (SSRR) loss fundamentally improves codec training and performance. First, SSRR significantly accelerates convergence, enabling competitive results after 300k training steps on a single H200 GPU. Second, it enhances intelligibility by reconstructing distilled self-supervised representations from codec outputs. Third, SSRR enables high intelligibility without additional lookahead in streaming Transformer-based codecs, allowing a zero-lookahead architecture for real-time deployment. On LibriSpeech test-clean, JHCodec achieves the best WER and CER among the evaluated codecs while maintaining zero lookahead and low end-to-end latency. We open-source the full implementation, training pipeline, and demo on GitHubh ttps://github.com/jhcodec843/jhcodec. --- ### Reconstruct! Don't Encode: Self-Supervised Representation Reconstruction Loss for High-Intelligibility and Low-Latency Streaming Neural Audio Codec Authors: Junhyeok Lee, Xiluo He, Jihwan Lee, Helin Wang, Shrikanth Narayanan, Thomas Thebaud, Laureano Moro-Velazquez, Jesús Villalba, Najim Dehak Published: 2026-03-06 Categories: eess.AS, cs.AI arXiv: https://arxiv.org/abs/2603.05887v1 Neural audio codecs optimized for mel-spectrogram reconstruction often fail to preserve intelligibility. While semantic encoder distillation improves encoded representations, it does not guarantee content preservation in reconstructed speech. In this work, we demonstrate that self-supervised representation reconstruction (SSRR) loss fundamentally improves codec training and performance. First, SSRR significantly accelerates convergence, enabling competitive results using only a single GPU. Second, it enhances intelligibility by reconstructing distilled self-supervised representations from codec outputs. Third, SSRR enables high intelligibility without additional lookahead in streaming Transformer-based codecs, allowing a zero-lookahead architecture for real-time deployment. As a result, our JHCodec achieves state-of-the-art performance while maintaining minimal latency and reduced training cost. We open-source the full implementation, training pipeline, and demo on Github https://github.com/jhcodec843/jhcodec. --- ### Computational Pathology in the Era of Emerging Foundation and Agentic AI -- International Expert Perspectives on Clinical Integration and Translational Readiness Authors: Qian Da, Yijiang Chen, Min Ju, Zheyi Ji, Albert Zhou, Wenwen Wang, Matthew A Abikenari, Philip Chikontwe, Guillaume Larghero, Bowen Chen, Peter Neidlinger, Dingrong Zhong et al. Published: 2026-03-06 Categories: cs.CE, cs.AI arXiv: https://arxiv.org/abs/2603.05884v2 Recent breakthroughs in artificial intelligence through foundation models and agents have accelerated the evolution of computational pathology. Demonstrated performance gains reported across academia in benchmarking datasets in predictive tasks such as diagnosis, prognosis, and treatment response have ignited substantial enthusiasm for clinical application. Despite this development momentum, real world adoption has lagged, as implementation faces economic, technical, and administrative challenges. Beyond existing discussions of technical architectures and comparative performance, this review considers how these emerging AI systems can be responsibly integrated into medical practice by connecting deployable clinical relevance with downstream analytical capabilities and their technical maturity, operational readiness, and economic and regulatory context. Drawing on perspectives from an international group, we provide a practical assessment of current capabilities and barriers to adoption in patient care settings. --- ### VerChol -- Grammar-First Tokenization for Agglutinative Languages Authors: Prabhu Raja Published: 2026-03-06 Categories: cs.CL arXiv: https://arxiv.org/abs/2603.05883v1 Tokenization is the foundational step in all large language model (LLM) pipelines, yet the dominant approach Byte Pair Encoding (BPE) and its variants is inherently script agnostic and optimized for English like morphology. For agglutinative languages a typological class encompassing the Dravidian family (Tamil, Kannada, Telugu, Malayalam), Turkic languages (Turkish, Azerbaijani, Uzbek), Uralic languages (Finnish, Hungarian, Estonian), Korean, Japanese, Swahili, Basque, and others, a single word may encode root, tense, aspect, person, number, gender agreement, case, and postpositions into one orthographic unit. Statistical tokenizers fragment these words into byte pair chunks that sever morpheme boundaries and inflate token counts. --- ### CylinderSplat: 3D Gaussian Splatting with Cylindrical Triplanes for Panoramic Novel View Synthesis Authors: Qiwei Wang, Xianghui Ze, Jingyi Yu, Yujiao Shi Published: 2026-03-06 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.05882v1 Feed-forward 3D Gaussian Splatting (3DGS) has shown great promise for real-time novel view synthesis, but its application to panoramic imagery remains challenging. Existing methods often rely on multi-view cost volumes for geometric refinement, which struggle to resolve occlusions in sparse-view scenarios. Furthermore, standard volumetric representations like Cartesian Triplanes are poor in capturing the inherent geometry of $360^\circ$ scenes, leading to distortion and aliasing. In this work, we introduce CylinderSplat, a feed-forward framework for panoramic 3DGS that addresses these limitations. The core of our method is a new {cylindrical Triplane} representation, which is better aligned with panoramic data and real-world structures adhering to the Manhattan-world assumption. We use a dual-branch architecture: a pixel-based branch reconstructs well-observed regions, while a volume-based branch leverages the cylindrical Triplane to complete occluded or sparsely-viewed areas. Our framework is designed to flexibly handle a variable number of input views, from single to multiple panoramas. Extensive experiments demonstrate that CylinderSplat achieves state-of-the-art results in both single-view and multi-view panoramic novel view synthesis, outperforming previous methods in both reconstruction quality and geometric accuracy. --- ### Confidence Before Answering: A Paradigm Shift for Efficient LLM Uncertainty Estimation Authors: Changcheng Li, Jiancan Wu, Hengheng Zhang, Zhengsu Chen, Guo An, Junxiang Qiu, Xiang Wang, Qi Tian Published: 2026-03-06 Categories: cs.CL arXiv: https://arxiv.org/abs/2603.05881v2 Reliable deployment of large language models (LLMs) requires accurate uncertainty estimation. Existing methods are predominantly answer-first, producing confidence only after generating an answer, which measure the correctness of a specific response and limits practical usability. We study a confidence-first paradigm, where the model outputs its confidence before answering, interpreting this score as the model's probability of answering the question correctly under its current policy. We propose CoCA(Co-optimized Confidence and Answers), a GRPO reinforcement learning framework that jointly optimizes confidence calibration and answer accuracy via segmented credit assignment. By assigning separate rewards and group-relative advantages to confidence and answer segments, CoCA enables stable joint optimization and avoids reward hacking. Experiments across math, code, and factual QA benchmarks show improved calibration and uncertainty discrimination while preserving answer quality, thereby enabling a broader range of downstream applications. --- ### ROSE: Reordered SparseGPT for More Accurate One-Shot Large Language Models Pruning Authors: Mingluo Su, Huan Wang Published: 2026-03-06 Categories: cs.CL, cs.LG arXiv: https://arxiv.org/abs/2603.05878v1 Pruning is widely recognized as an effective method for reducing the parameters of large language models (LLMs), potentially leading to more efficient deployment and inference. One classic and prominent path of LLM one-shot pruning is to leverage second-order gradients (i.e., Hessian), represented by the pioneering work SparseGPT. However, the predefined left-to-right pruning order in SparseGPT leads to suboptimal performance when the weights exhibit columnar patterns. This paper studies the effect of pruning order under the SparseGPT framework. The analyses lead us to propose ROSE, a reordered SparseGPT method that prioritizes weights with larger potential pruning errors to be pruned earlier. ROSE first performs pre-pruning to identify candidate weights for removal, and estimates both column and block pruning loss. Subsequently, two-level reordering is performed: columns within each block are reordered in descending order of column loss, while blocks are reordered based on block loss. We introduce the relative range of block loss as a metric to identify columnar layers, enabling adaptive reordering across the entire model. Substantial empirical results on prevalent LLMs (LLaMA2-7B/13B/70B, LLaMA3-8B, Mistral-7B) demonstrate that ROSE surpasses the original SparseGPT and other counterpart pruning methods. Our code is available at https://github.com/mingluo-su/ROSE. --- ### Systematic Evaluation of Novel View Synthesis for Video Place Recognition Authors: Muhammad Zawad Mahmud, Samiha Islam, Damian Lyons Published: 2026-03-06 Categories: cs.CV, cs.RO arXiv: https://arxiv.org/abs/2603.05876v2 The generation of synthetic novel views has the potential to positively impact robot navigation in several ways. In image-based navigation, a novel overhead view generated from a scene taken by a ground robot could be used to guide an aerial robot to that location. In Video Place Recognition (VPR), novel views of ground locations from the air can be added that enable a UAV to identify places seen by the ground robot, and similarly, overhead views can be used to generate novel ground views. This paper presents a systematic evaluation of synthetic novel views in VPR using five public VPR image databases and seven typical image similarity methods. We show that for small synthetic additions, novel views improve VPR recognition statistics. We find that for larger additions, the magnitude of viewpoint change is less important than the number of views added and the type of imagery in the dataset. --- ### Stochastic Event Prediction via Temporal Motif Transitions Authors: İbrahim Bahadır Altun, Ahmet Erdem Sarıyüce Published: 2026-03-06 Categories: cs.LG, cs.SI arXiv: https://arxiv.org/abs/2603.05874v1 Networks of timestamped interactions arise across social, financial, and biological domains, where forecasting future events requires modeling both evolving topology and temporal ordering. Temporal link prediction methods typically frame the task as binary classification with negative sampling, discarding the sequential and correlated nature of real-world interactions. We introduce STEP (STochastic Event Predictor), a framework that reformulates temporal link prediction as a sequential forecasting problem in continuous time. STEP models event dynamics through discrete temporal motif transitions governed by Poisson processes, maintaining a set of open motif instances that evolve as new interactions arrive. At each step, the framework decides whether to initiate a new temporal motif or extend an existing one, selecting the most probable event via Bayesian scoring of temporal likelihoods and structural priors. STEP also produces compact, temporal motif-based feature vectors that can be concatenated with existing temporal graph neural network outputs, enriching their representations without architectural modifications. Experiments on five real-world datasets demonstrate up to 21% average precision gains over state-of-the-art baselines in classification and 0.99 precision in next $k$ sequential forecasting, with consistently lower runtime than competing motif-aware methods. --- ### Shifting Adaptation from Weight Space to Memory Space: A Memory-Augmented Agent for Medical Image Segmentation Authors: Bowen Chen, Qiaohui Gao, Shaowen Wan, Shanhui Sun, Wei Liu, Xiang Li, Tianming Liu, Lin Zhao Published: 2026-03-06 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.05873v1 Medical image segmentation is fundamental to clinical workflows, yet models trained on a single dataset often fail to generalize across institutions, scanners, or patient populations. While vision foundation models have shown great promise in addressing this challenge, their deployment typically requires task-specific fine-tuning, which introduces substantial communication overhead in federated learning and prevents continuous knowledge evolution during deployment. In this work, we propose a memory-augmented segmentation agent (MemSeg-Agent) that shifts adaptation from weight space to memory space, enabling few-shot learning, federated supervised learning, and test-time adaptation within a unified architecture. MemSeg-Agent conditions a fixed backbone with lightweight static, few-shot, and test-time working memories, which are dynamically composed by an agentic controller. In federated settings, we update compact memory units instead of model parameters, substantially reducing communication overhead. Experiments on four public datasets demonstrate strong performance and robustness to domain shift: Static memory alone matches or surpasses strong supervised baselines with high parameter efficiency, and test-time working memory further improves in-domain and cross-domain performance without fine-tuning. Overall, MemSeg-Agent introduces a new paradigm for scalable and adaptive medical image segmentation in the era of agentic AI. --- ### PatchCue: Enhancing Vision-Language Model Reasoning with Patch-Based Visual Cues Authors: Yukun Qi, Pei Fu, Hang Li, Yuhan Liu, Chao Jiang, Bin Qin, Zhenbo Luo, Jian Luan Published: 2026-03-06 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.05869v2 Vision-Language Models (VLMs) have achieved remarkable progress on a wide range of challenging multimodal understanding and reasoning tasks. However, existing reasoning paradigms, such as the classical Chain-of-Thought (CoT), rely solely on textual information and often underutilize important visual cues. While prior work has incorporated pixel-level visual cues, these representations require precise spatial localization, introducing additional learning complexity. To address this, we propose PatchCue, a novel patch-based visual cue paradigm designed to significantly enhance the visual reasoning capabilities of VLMs. By partitioning images into patches and representing cues at the patch level, PatchCue aligns better with human perceptual habits and leverages the patch-tokenized input of modern VLMs. We train VLMs using a two-stage approach: cold-start supervised fine-tuning to output patch-level cues, followed by reinforcement learning with a process-supervised cue reward that guides intermediate visual reasoning steps. Extensive experiments on multiple VLMs and diverse benchmarks, including general visual question answering, complex reasoning, and document understanding, demonstrate that PatchCue consistently improves overall model performance. Our results show that patch-level cues outperform both pixel-level bounding boxes and point-based cues, providing a more effective and cognitively aligned visual reasoning paradigm. --- ### TumorChain: Interleaved Multimodal Chain-of-Thought Reasoning for Traceable Clinical Tumor Analysis Authors: Sijing Li, Zhongwei Qiu, Jiang Liu, Wenqiao Zhang, Tianwei Lin, Yihan Xie, Jianxiang An, Boxiang Yun, Chenglin Yang, Jun Xiao, Guangyu Guo, Jiawen Yao et al. Published: 2026-03-06 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.05867v2 Accurate tumor analysis is central to clinical radiology and precision oncology, where early detection, reliable lesion characterization, and pathology-level risk assessment guide diagnosis and treatment planning. Chain-of-Thought (CoT) reasoning is particularly important in this setting because it enables step-by-step interpretation from imaging findings to clinical impressions and pathology conclusions, improving traceability and reducing diagnostic errors. Here, we target the clinical tumor analysis task and build a large-scale benchmark that operationalizes a multimodal reasoning pipeline, spanning findings, impressions, and pathology predictions. We curate TumorCoT, a large-scale dataset of 1.5M CoT-labeled VQA instructions paired with 3D CT scans, with step-aligned rationales and cross-modal alignments along the trajectory from findings to impression to pathology, enabling evaluation of both answer accuracy and reasoning consistency. We further propose TumorChain, a multimodal interleaved reasoning framework that tightly couples 3D imaging encoders, clinical text understanding, and organ-level vision-language alignment. Through cross-modal alignment and iterative interleaved causal reasoning, TumorChain grounds visual evidence, aggregates conclusions, and issues pathology predictions after multiple rounds of self-refinement, improving traceability and reducing hallucination risk. Experiments show consistent improvements over strong baselines in lesion detection, impression generation, and pathology classification, and demonstrate strong generalization on the DeepTumorVQA benchmark. These results highlight the potential of multimodal reasoning for reliable and interpretable tumor analysis in clinical practice. Detailed information about our project can be found on our project homepage at https://github.com/ZJU4HealthCare/TumorChain. --- ### ReflexiCoder: Teaching Large Language Models to Self-Reflect on Generated Code and Self-Correct It via Reinforcement Learning Authors: Juyong Jiang, Jiasi Shen, Sunghun Kim, Kang Min Yoo, Jeonghoon Kim, Sungju Kim Published: 2026-03-06 Categories: cs.CL, cs.LG, cs.SE arXiv: https://arxiv.org/abs/2603.05863v2 While Large Language Models (LLMs) have revolutionized code generation, standard ``System 1'' approaches that generate solutions in a single forward pass often hit a performance ceiling on complex algorithmic tasks. Existing iterative refinement strategies attempt to bridge this gap at inference time, yet they predominantly rely on external oracles, execution feedback, or computationally expensive prompt-response cycles. In this work, we propose ReflexiCoder, a novel reinforcement learning (RL) framework that internalizes the structured reasoning trajectory, encompassing initial generation, bug and optimization aware reflection, and self-correction, directly into the model's weights. Unlike prior methods, ReflexiCoder shifts the paradigm from external-dependent refinement to an intrinsic, fully autonomous self-reflection and self-correction capabilities at inference time. We utilize an RL-only training paradigm with granular reward functions to optimize the entire reflection-correction trajectory, teaching the model how to debug without reliance on ground-truth feedback or execution engines at inference time. Extensive experiments across seven benchmarks demonstrate that our ReflexiCoder-8B establishes a new state-of-the-art (SOTA) among leading open-source models in the 1.5B to 14B range, achieving 94.51% (87.20%) on HumanEval (Plus), 81.80% (78.57%) on MBPP (Plus), 35.00% on BigCodeBench, 52.21% on LiveCodeBench, and 37.34% on CodeForces in a single-attempt setting, rivaling or surpassing proprietary models like GPT-5.1. Notably, our framework is significantly more token-efficient than base models, reducing inference-time compute overhead by approximately 40% through disciplined, efficient reasoning and reflection patterns. The source code and data are available at https://github.com/juyongjiang/ReflexiCoder. --- ### Evolving Medical Imaging Agents via Experience-driven Self-skill Discovery Authors: Lin Fan, Pengyu Dai, Zhipeng Deng, Haolin Wang, Xun Gong, Yefeng Zheng, Yafei Ou Published: 2026-03-06 Categories: cs.AI, cs.CV arXiv: https://arxiv.org/abs/2603.05860v1 Clinical image interpretation is inherently multi-step and tool-centric: clinicians iteratively combine visual evidence with patient context, quantify findings, and refine their decisions through a sequence of specialized procedures. While LLM-based agents promise to orchestrate such heterogeneous medical tools, existing systems treat tool sets and invocation strategies as static after deployment. This design is brittle under real-world domain shifts, across tasks, and evolving diagnostic requirements, where predefined tool chains frequently degrade and demand costly manual re-design. We propose MACRO, a self-evolving, experience-augmented medical agent that shifts from static tool composition to experience-driven tool discovery. From verified execution trajectories, the agent autonomously identifies recurring effective multi-step tool sequences, synthesizes them into reusable composite tools, and registers these as new high-level primitives that continuously expand its behavioral repertoire. A lightweight image-feature memory grounds tool selection in a visual-clinical context, while a GRPO-like training loop reinforces reliable invocation of discovered composites, enabling closed-loop self-improvement with minimal supervision. Extensive experiments across diverse medical imaging datasets and tasks demonstrate that autonomous composite tool discovery consistently improves multi-step orchestration accuracy and cross-domain generalization over strong baselines and recent state-of-the-art agentic methods, bridging the gap between brittle static tool use and adaptive, context-aware clinical AI assistance. Code will be available upon acceptance. --- ### VS3R: Robust Full-frame Video Stabilization via Deep 3D Reconstruction Authors: Muhua Zhu, Xinhao Jin, Xinping Wang, Yu Zhang, Yifei Xue, Tie Ji, Yizhen Lao Published: 2026-03-06 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.05851v2 Video stabilization aims to mitigate camera shake but faces a fundamental trade-off between geometric robustness and full-frame consistency. While 2D methods suffer from aggressive cropping, 3D techniques are often undermined by fragile optimization pipelines that fail under extreme motions. Novel view synthesis models suffer from structural artifacts and scale blindness. To bridge this gap, we propose VS3R, a framework that synergizes feed-forward 3D reconstruction with generative video diffusion. Our pipeline jointly estimates camera parameters, depth, and masks to ensure all-scenario reliability, and introduces a Hybrid Stabilized Rendering (HSR) module that fuses semantic and geometric cues to preliminarily address parallax occlusions caused by pose transformations while maintaining dynamic-static consistency. Finally, a Video Stabilization-Driven Diffusion Model (VSDM) leverages contextual information to restore disoccluded regions, jointly optimizing texture and temporal consistency. Collectively, VS3R achieves high-fidelity, full-frame stabilization across diverse camera models and significantly outperforms state-of-the-art methods in robustness and visual quality. --- ### Cog2Gen3D: Sculpturing 3D Semantic-Geometric Cognition for 3D Generation Authors: Haonan Wang, Hanyu Zhou, Haoyue Liu, Tao Gu, Luxin Yan Published: 2026-03-06 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.05845v1 Generative models have achieved success in producing semantically plausible 2D images, but it remains challenging in 3D generation due to the absence of spatial geometry constraints. Typically, existing methods utilize geometric features as conditions to enhance spatial awareness. However, these methods can only model relative relationships and are prone to scale inconsistency of absolute geometry. Thus, we argue that semantic information and absolute geometry empower 3D cognition, thereby enabling controllable 3D generation for the physical world. In this work, we propose Cog2Gen3D, a 3D cognition-guided diffusion framework for 3D generation. Our model is guided by three key designs: 1) Cognitive Feature Embeddings. We encode different modalities into semantic and geometric representations and further extract logical representations. 2) 3D Latent Cognition Graph. We structure different representations into dual-stream semantic-geometric graphs and fuse them via common-based cross-attention to obtain a 3D cognition graph. 3) Cognition-Guided Latent Diffusion. We leverage the fused 3D cognition graph as the condition to guide the latent diffusion process for 3D Gaussian generation. Under this unified framework, the 3D cognition graph ensures the physical plausibility and structural rationality of 3D generation. Moreover, we construct a validation subset based on the Marble World Labs. Extensive experiments demonstrate that our Cog2Gen3D significantly outperforms existing methods in both semantic fidelity and geometric plausibility. --- ### Remote Sensing Image Classification Using Deep Ensemble Learning Authors: Niful Islam, Md. Rayhan Ahmed, Nur Mohammad Fahad, Salekul Islam, A. K. M. Muzahidul Islam, Saddam Mukta, Swakkhar Shatabda Published: 2026-03-06 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2603.05844v1 Remote sensing imagery plays a crucial role in many applications and requires accurate computerized classification techniques. Reliable classification is essential for transforming raw imagery into structured and usable information. While Convolutional Neural Networks (CNNs) are mostly used for image classification, they excel at local feature extraction, but struggle to capture global contextual information. Vision Transformers (ViTs) address this limitation through self attention mechanisms that model long-range dependencies. Integrating CNNs and ViTs, therefore, leads to better performance than standalone architectures. However, the use of additional CNN and ViT components does not lead to further performance improvement and instead introduces a bottleneck caused by redundant feature representations. In this research, we propose a fusion model that combines the strengths of CNNs and ViTs for remote sensing image classification. To overcome the performance bottleneck, the proposed approach trains four independent fusion models that integrate CNN and ViT backbones and combine their outputs at the final prediction stage through ensembling. The proposed method achieves accuracy rates of 98.10 percent, 94.46 percent, and 95.45 percent on the UC Merced, RSSCN7, and MSRSI datasets, respectively. These results outperform competing architectures and highlight the effectiveness of the proposed solution, particularly due to its efficient use of computational resources during training. --- ### Evaluating LLM Alignment With Human Trust Models Authors: Anushka Debnath, Stephen Cranefield, Bastin Tony Roy Savarimuthu, Emiliano Lorini Published: 2026-03-06 Categories: cs.MA, cs.AI arXiv: https://arxiv.org/abs/2603.05839v1 Trust plays a pivotal role in enabling effective cooperation, reducing uncertainty, and guiding decision-making in both human interactions and multi-agent systems. Although it is significant, there is limited understanding of how large language models (LLMs) internally conceptualize and reason about trust. This work presents a white-box analysis of trust representation in EleutherAI/gpt-j-6B, using contrastive prompting to generate embedding vectors within the activation space of the LLM for diadic trust and related interpersonal relationship attributes. We first identified trust-related concepts from five established human trust models. We then determined a threshold for significant conceptual alignment by computing pairwise cosine similarities across 60 general emotional concepts. Then we measured the cosine similarities between the LLM's internal representation of trust and the derived trust-related concepts. Our results show that the internal trust representation of EleutherAI/gpt-j-6B aligns most closely with the Castelfranchi socio-cognitive model, followed by the Marsh Model. These findings indicate that LLMs encode socio-cognitive constructs in their activation space in ways that support meaningful comparative analyses, inform theories of social cognition, and support the design of human-AI collaborative systems. --- ### Architectural Unification for Polarimetric Imaging Across Multiple Degradations Authors: Chu Zhou, Yufei Han, Junda Liao, Linrui Dai, Wangze Xu, Art Subpa-Asa, Heng Guo, Boxin Shi, Imari Sato Published: 2026-03-06 Categories: eess.IV, cs.CV arXiv: https://arxiv.org/abs/2603.05834v1 Polarimetric imaging aims to recover polarimetric parameters, including Total Intensity (TI), Degree of Polarization (DoP), and Angle of Polarization (AoP), from captured polarized measurements. In real-world scenarios, these measurements are frequently affected by diverse degradations such as low-light noise, motion blur, and mosaicing artifacts. Due to the nonlinear dependency of DoP and AoP on the measured intensities, accurately retrieving physically consistent polarimetric parameters from degraded observations remains highly challenging. Existing approaches typically adopt task-specific network architectures tailored to individual degradation types, limiting their adaptability across different restoration scenarios. Moreover, many methods rely on multi-stage processing pipelines that suffer from error accumulation, or operate solely in a single domain (either image or Stokes domain), failing to fully exploit the intrinsic physical relationships between them. In this work, we propose a unified architectural framework for polarimetric imaging that is structurally shared across multiple degradation scenarios. Rather than redesigning network structures for each task, our framework maintains a consistent architectural design while being trained separately for different degradations. The model performs single-stage joint image-Stokes processing, avoiding error accumulation and explicitly preserving physical consistency. Extensive experiments show that this unified architectural design, when trained for specific degradation types, consistently achieves state-of-the-art performance across low-light denoising, motion deblurring, and demosaicing tasks, establishing a versatile and physically grounded solution for degraded polarimetric imaging. --- ### Lexara: A User-Centered Toolkit for Evaluating Large Language Models for Conversational Visual Analytics Authors: Srishti Palani, Vidya Setlur Published: 2026-03-06 Categories: cs.HC, cs.AI arXiv: https://arxiv.org/abs/2603.05832v1 Large Language Models (LLMs) are transforming Conversational Visual Analytics (CVA) by enabling data analysis through natural language. However, evaluating LLMs for CVA remains a challenge: requiring programming expertise, overlooking real-world complexity, and lacking interpretable metrics for multi-format (visualizations and text) outputs. Through interviews with 22 CVA developers and 16 end-users, we identified use cases, evaluation criteria and workflows. We present Lexara, a user-centered evaluation toolkit for CVA that operationalizes these insights into: (i) test cases spanning real-world scenarios; (ii) interpretable metrics covering visualization quality (data fidelity, semantic alignment, functional correctness, design clarity) and language quality (factual grounding, analytical reasoning, conversational coherence) using rule-based and LLM-as-a-Judge methods; and (iii) an interactive toolkit enabling experimental setup and multi-format and multi-level exploration of results without programming expertise. We conducted a two-week diary study with six CVA developers, drawn from our initial cohort of 22. Their feedback demonstrated Lexara's effectiveness for guiding appropriate model and prompt selection. --- ### Test-Time Adaptation via Many-Shot Prompting: Benefits, Limits, and Pitfalls Authors: Shubhangi Upasani, Chen Wu, Jay Rainton, Bo Li, Urmish Thakker, Changran Hu, Qizheng Zhang Published: 2026-03-06 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2603.05829v3 Test-time adaptation enables large language models (LLMs) to modify their behavior at inference without updating model parameters. A common approach is many-shot prompting, where large numbers of in-context learning (ICL) examples are injected as an input-space test-time update. Although performance can improve as more demonstrations are added, the reliability and limits of this update mechanism remain poorly understood, particularly for open-source models. We present an empirical study of many-shot prompting across tasks and model backbones, analyzing how performance varies with update magnitude, example ordering, and selection policy. We further study Dynamic and Reinforced ICL as alternative test-time update strategies that control which information is injected and how it constrains model behavior. We find that many-shot prompting is effective for structured tasks where demonstrations provide high information gain, but is highly sensitive to selection strategy and often shows limited benefits for open-ended generation tasks. Overall, we characterize the practical limits of prompt-based test-time adaptation and outline when input-space updates are beneficial versus harmful. --- ### HART: Data-Driven Hallucination Attribution and Evidence-Based Tracing for Large Language Models Authors: Shize Liang, Hongzhi Wang Published: 2026-03-06 Categories: cs.CL arXiv: https://arxiv.org/abs/2603.05828v1 Large language models (LLMs) have demonstrated remarkable performance in text generation and knowledge-intensive question answering. Nevertheless, they are prone to producing hallucinated content, which severely undermines their reliability in high-stakes application domains. Existing hallucination attribution approaches, based on either external knowledge retrieval or internal model mechanisms, primarily focus on semantic similarity matching or representation-level discrimination. As a result, they have difficulty establishing structured correspondences at the span level between hallucination types, underlying error generation mechanisms, and external factual evidence, thereby limiting the interpretability of hallucinated fragments and the traceability of supporting or opposing evidence. To address these limitations, we propose HART, a fine-grained hallucination attribution and evidence retrieval framework for large language models. HART formalizes hallucination tracing as a structured modeling task comprising four stages: span localization, mechanism attribution, evidence retrieval, and causal tracing. Based upon this formulation, we develop the first structured dataset tailored for hallucination tracing, in which hallucination types, error mechanisms, and sets of counterfactual evidence are jointly annotated to enable causal-level interpretability evaluation. Experimental results on the proposed dataset demonstrate that HART substantially outperforms strong retrieval baselines, including BM25 and DPR, validating the effectiveness and generalization capability of the proposed tracing paradigm for hallucination analysis and evidence alignment. --- ### Self-Auditing Parameter-Efficient Fine-Tuning for Few-Shot 3D Medical Image Segmentation Authors: Son Thai Ly, Hien V. Nguyen Published: 2026-03-06 Categories: cs.LG arXiv: https://arxiv.org/abs/2603.05822v1 Adapting foundation models to new clinical sites remains challenging in practice. Domain shift and scarce annotations must be handled by experts, yet many clinical groups do not have ready access to skilled AI engineers to tune adapter designs and training recipes. As a result, adaptation cycles can stretch from weeks to months, particularly in few-shot settings. Existing PEFT methods either require manual adapter configuration or automated searches that are computationally infeasible in few-shot 3D settings. We propose SEA-PEFT (SElf-Auditing Parameter-Efficient Fine-Tuning) to automate this process. SEA-PEFT treats adapter configuration as an online allocation problem solved during fine-tuning rather than through manual, fixed-topology choices. SEA-PEFT uses a search-audit-allocate loop that trains active adapters, estimates each adapter's Dice utility by momentarily toggling it off, and then reselects the active set under a parameter budget using a greedy knapsack allocator. Exponential Moving Average and Interquartile Range smoothing, together with a Finite-State Ranking controller, stabilize the loop and improve reliability in high-noise few-shot regimes. On TotalSegmentator and FLARE'22, SEA-PEFT improves mean Dice by 2.4--2.8 points over the strongest fixed-topology PEFT baselines across 1/5/10-shot settings while training <1% of parameters. For reproducibility purposes, we made our code publicly available at https://github.com/tsly123/SEA_PEFT --- ### RouteGoT: Node-Adaptive Routing for Cost-Efficient Graph of Thoughts Reasoning Authors: Yuhang Liu, Ruijie Wang, Yunlong Chu, Bing Hao, Yumeng Lin, Shengzhong Liu, Minglai Shao Published: 2026-03-06 Categories: cs.CL arXiv: https://arxiv.org/abs/2603.05818v1 Large Language Models (LLMs) excel at multi-step reasoning, yet increasing the structural complexity of inference does not consistently improve system-level returns. Methods such as Tree of Thoughts (ToT), Graph of Thoughts (GoT), and Adaptive Graph of Thoughts (AGoT) can boost accuracy on some benchmarks, but often introduce substantial overhead in token consumption and latency, and their gains can be unstable across task distributions-sometimes underperforming simpler Chain-of-Thought (CoT) or direct input-output prompting (IO). We attribute this inefficiency to stage-wise and node-wise heterogeneity inside GoT-style reasoning pipelines: high-quality planning and final synthesis are globally coupled and typically benefit from strong models, whereas many intermediate subtasks are localized and can be solved accurately by lighter models with far fewer tokens. Motivated by these observations, we propose RouteGoT, a budget-controllable, node-adaptive routing framework for graph-structured reasoning. RouteGoT performs in-graph routing by prioritizing strong models for planning and synthesis, while dynamically allocating lightweight models and cost-effective strategies to leaf subtasks based on predicted difficulty. It further integrates explicit budget constraints into a global inference scheduler to control graph expansion under a user-specified token budget, enabling predictable performance-cost trade-offs. Experiments across reasoning, retrieval, and multi-hop QA benchmarks show that RouteGoT matching or improving accuracy while substantially reducing token usage; specifically, it achieves an average 8.1 percentage points accuracy improvement and 79.1\% output token reduction compared to AGoT. Furthermore, RouteGoT outperforms existing routing baselines by maintaining a superior cost-accuracy trade-off, demonstrating improved robustness under varying budget targets and tasks. --- ### Margin and Consistency Supervision for Calibrated and Robust Vision Models Authors: Salim Khazem Published: 2026-03-06 Categories: cs.CV, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2603.05812v1 Deep vision classifiers often achieve high accuracy while remaining poorly calibrated and fragile under small distribution shifts. We present Margin and Consistency Supervision (MaCS), a simple, architecture-agnostic regularization framework that jointly enforces logit-space separation and local prediction stability. MaCS augments cross-entropy with (i) a hinge-squared margin penalty that enforces a target logit gap between the correct class and the strongest competitor, and (ii) a consistency regularizer that minimizes the KL divergence between predictions on clean inputs and mildly perturbed views. We provide a unifying theoretical analysis showing that increasing classification margin while reducing local sensitivity formalized via a Lipschitz-type stability proxy yields improved generalization guarantees and a provable robustness radius bound scaling with the margin-to-sensitivity ratio. Across several image classification benchmarks and several backbones spanning CNNs and Vision Transformers, MaCS consistently improves calibration (lower ECE and NLL) and robustness to common corruptions while preserving or improving top-1 accuracy. Our approach requires no additional data, no architectural changes, and negligible inference overhead, making it an effective drop-in replacement for standard training objectives. --- ### Video Compression Meets Video Generation: Latent Inter-Frame Pruning with Attention Recovery Authors: Dennis Menn, Yuedong Yang, Bokun Wang, Xiwen Wei, Mustafa Munir, Feng Liang, Radu Marculescu, Chenfeng Xu, Diana Marculescu Published: 2026-03-06 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.05811v2 Current video generation models suffer from high computational latency, making real-time applications prohibitively costly. In this paper, we address this limitation by exploiting the temporal redundancy inherent in video latent patches. To this end, we propose the Latent Inter-frame Pruning with Attention Recovery (LIPAR) framework, which detects and skips recomputing duplicated latent patches. Additionally, we introduce a novel Attention Recovery mechanism that approximates the attention values of pruned tokens, thereby removing visual artifacts arising from naively applying the pruning method. Empirically, our method increases video editing throughput by $1.53\times$, achieving an average of 19.3 FPS on an NVIDIA RTX 4090 with the 1.3B Self-Forcing model (4-step denoising, FP16). The proposed method does not compromise generation quality and can be seamlessly integrated with the model without additional training. Our approach effectively bridges the gap between traditional compression algorithms and modern generative pipelines. --- ### EventGeM: Global-to-Local Feature Matching for Event-Based Visual Place Recognition Authors: Adam D. Hines, Gokul B. Nair, Nicolás Marticorena, Michael Milford, Tobias Fischer Published: 2026-03-06 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.05807v2 Event cameras are rapidly rising in popularity for robotic and computer vision tasks because their sparse activation delivers energy-efficient, high-dynamic-range, and fast sensing. Event cameras have been used in robotic navigation and localization tasks where positioning must occur in real time with sufficient accuracy. However, current event-based localization methods suffer from poor spatial understanding and are not viewpoint tolerant. In this paper, we address the problem of viewpoint-robust place recognition directly from event streams. We present EventGeM, a global-to-local feature fusion pipeline for event-based visual place recognition that combines whole-image feature detection to shortlist top candidates for 2D homography-based re-ranking with random sample consensus (RANSAC). We also contribute a regional generalized mean pooling (GeM) layer that learns to return the most relevant spatial features using per-row exponents to pool event streams into a compact global descriptor, trained on the NYC-Event-VPR dataset. These contributions overcome shortfalls in currently available event-based localization methods that fail to recognize similar places with large changes in viewpoint. To evaluate viewpoint-robust localization, we contribute a new event-based dataset that includes repeated traverses with a severe lateral shift. EventGeM improves absolute Recall@1 by 7 to 43 percentage points over the strongest baseline in each experiment. We also deploy EventGeM on a robotic platform, demonstrating real-time performance of our hierarchical pipeline. The code for EventGeM is available at https://github.com/AdamDHines/Event-GeM. --- ### EventGeM: Global-to-Local Feature Matching for Event-Based Visual Place Recognition Authors: Adam D. Hines, Gokul B. Nair, Nicolás Marticorena, Michael Milford, Tobias Fischer Published: 2026-03-06 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.05807v1 Dynamic vision sensors, also known as event cameras, are rapidly rising in popularity for robotic and computer vision tasks due to their sparse activation and high-temporal resolution. Event cameras have been used in robotic navigation and localization tasks where accurate positioning needs to occur on small and frequent time scales, or when energy concerns are paramount. In this work, we present EventGeM, a state-of-the-art global to local feature fusion pipeline for event-based Visual Place Recognition. We use a pre-trained vision transformer (ViT-S/16) backbone to obtain global feature patch for initial match predictions embeddings from event histogram images. Local feature keypoints were then detected using a pre-trained MaxViT backbone for 2D-homography based re-ranking with RANSAC. For additional re-ranking refinement, we subsequently used a pre-trained vision foundation model for depth estimation to compare structural similarity between references and queries. Our work performs state-of-the-art localization when compared to the best currently available event-based place recognition method across several benchmark datasets and lighting conditions all whilst being fully capable of running in real-time when deployed across a variety of compute architectures. We demonstrate the capability of EventGeM in a real-world deployment on a robotic platform for online localization using event streams directly from an event camera. Project page: https://eventgemvpr.github.io/ --- ### MoE Lens -- An Expert Is All You Need Authors: Marmik Chaudhari, Idhant Gulati, Nishkal Hundia, Pranav Karra, Shivam Raval Published: 2026-03-06 Categories: cs.LG arXiv: https://arxiv.org/abs/2603.05806v1 Mixture of Experts (MoE) models enable parameter-efficient scaling through sparse expert activations, yet optimizing their inference and memory costs remains challenging due to limited understanding of their specialization behavior. We present a systematic analysis of expert specialization in MoEs through two complementary approaches: domain-specific routing patterns and an early decoding framework that tracks expert contributions to output representations. Our analysis of the DeepSeekMoE model reveals that despite having 64 routed experts with 6 active for each layer's computation, the model predominantly relies on a few specialized experts, with the top-weighted expert's output closely approximating the full ensemble prediction. We quantitatively validate these findings through a systematic analysis of the token routing distribution, demonstrating that very few experts handle over 50\% of routing decisions across different specialized domains. Hidden state similarity between single and ensemble experts for every layer is extremely high, with some layers having cosine similarity as high as 0.95 and perplexity increasing by only 5\% when using a single expert across all three domains. Our results indicate that Mixture of Experts models exhibit concentrated expertise highlighting potential opportunities for inference optimization through targeted expert pruning while maintaining model performance and opening avenues towards studying localization of learned knowledge in these models. --- ### Sparse Crosscoders for diffing MoEs and Dense models Authors: Marmik Chaudhari, Nishkal Hundia, Idhant Gulati Published: 2026-03-06 Categories: cs.LG arXiv: https://arxiv.org/abs/2603.05805v1 Mixture of Experts (MoE) achieve parameter-efficient scaling through sparse expert routing, yet their internal representations remain poorly understood compared to dense models. We present a systematic comparison of MoE and dense model internals using crosscoders, a variant of sparse autoencoders, that jointly models multiple activation spaces. We train 5-layer dense and MoEs (equal active parameters) on 1B tokens across code, scientific text, and english stories. Using BatchTopK crosscoders with explicitly designated shared features, we achieve $\sim 87\%$ fractional variance explained and uncover concrete differences in feature organization. The MoE learns significantly fewer unique features compared to the dense model. MoE-specific features also exhibit higher activation density than shared features, whereas dense-specific features show lower density. Our analysis reveals that MoEs develop more specialized, focused representations while dense models distribute information across broader, more general-purpose features. --- ### Ambiguity Collapse by LLMs: A Taxonomy of Epistemic Risks Authors: Shira Gur-Arieh, Angelina Wang, Sina Fazelpour Published: 2026-03-06 Categories: cs.CY, cs.AI arXiv: https://arxiv.org/abs/2603.05801v1 Large language models (LLMs) are increasingly used to make sense of ambiguous, open-textured, value-laden terms. Platforms routinely rely on LLMs for content moderation, asking them to label text based on disputed concepts like "hate speech" or "incitement"; hiring managers may use LLMs to rank who counts as "qualified"; and AI labs increasingly train models to self-regulate under constitutional-style ambiguous principles such as "biased" or "legitimate". This paper introduces ambiguity collapse: a phenomenon that occurs when an LLM encounters a term that genuinely admits multiple legitimate interpretations, yet produces a singular resolution, in ways that bypass the human practices through which meaning is ordinarily negotiated, contested, and justified. Drawing on interdisciplinary accounts of ambiguity as a productive epistemic resource, we develop a taxonomy of the epistemic risks posed by ambiguity collapse at three levels: process (foreclosing opportunities to deliberate, develop cognitive skills, and shape contested terms), output (distorting the concepts and reasons agents act upon), and ecosystem (reshaping shared vocabularies, interpretive norms, and how concepts evolve over time). We illustrate these risks through three case studies, and conclude by sketching multi-layer mitigation principles spanning training, institutional deployment design, interface affordances, and the management of underspecified prompts, with the goal of designing systems that surface, preserve, and responsibly govern ambiguity. --- ### StreamWise: Serving Multi-Modal Generation in Real-Time at Scale Authors: Haoran Qiu, Gohar Irfan Chaudhry, Chaojie Zhang, Íñigo Goiri, Esha Choukse, Rodrigo Fonseca, Ricardo Bianchini Published: 2026-03-06 Categories: cs.DC, cs.AI arXiv: https://arxiv.org/abs/2603.05800v1 Advances in multi-modal generative models are enabling new applications, from storytelling to automated media synthesis. Most current workloads generate simple outputs (e.g., image generation from a prompt) in batch mode, often requiring several seconds even for basic results. Serving real-time multi-modal workflows at scale is costly and complex, requiring efficient coordination of diverse models (each with unique resource needs) across language, audio, image, and video, all under strict latency and resource constraints. We tackle these challenges through the lens of real-time podcast video generation, integrating LLMs, text-to-speech, and video-audio generation. To meet tight SLOs, we design an adaptive, modular serving system, StreamWise, that dynamically manages quality (e.g., resolution, sharpness), model/content parallelism, and resource-aware scheduling. We leverage heterogeneous hardware to maximize responsiveness and efficiency. For example, the system can lower video resolution and allocate more resources to early scenes. We quantify the trade-offs between latency, cost, and quality. The cheapest setup generates a 10-minute podcast video on A100 GPUs in 1.4 hours (8.4x slower than the real-time) for less than \$25. StreamWise enables high-quality real-time streaming with a sub-second startup delay under $45. --- ### The Coordination Gap: Multi-Agent Alternation Metrics for Temporal Fairness in Repeated Games Authors: Nikolaos Al. Papadopoulos, Ismael Tito Freire, Marti Sanchez-Fibla, Konstantinos E. Psannis Published: 2026-03-06 Categories: cs.MA, cs.GT, cs.LG arXiv: https://arxiv.org/abs/2603.05789v5 Repeated multi-agent interactions require evaluation metrics that capture not only payoff distributions but also their temporal organization. Conventional outcome-based fairness measures can assign similar aggregate scores to temporally distinct coordination patterns, obscuring whether access to a shared resource is genuinely rotating or persistently monopolized. We study this problem in the Honey-Jar Game (HJG), a minimally dynamic repeated threshold-congestion Markov game in which n agents compete for exclusive access to a single high-reward resource. We introduce Perfect Alternation (PA), a reference turn-taking regime corresponding to the n-periodic round-robin picking sequence, together with six novel Alternation (ALT) metrics and a benchmarking methodology mapping ALT values to interpretable PA-equivalent performance. Using Q-learning agents as a minimal adaptive baseline against analytically derived random-policy baselines, we uncover a clear measurement failure: despite deceptively high traditional metrics (e.g., reward fairness often exceeding 0.9), learned policies perform worse than random on every ALT metric, by 34-74% on CALT and up to 92% on EALT, with PA-equivalent coordination falling to roughly one-fifth of the population at n=10. EALT further reveals two distinct failure patterns: without episodic memory (Type-A), the deficit grows from -20% at n=2 to -92% at n=10; with episodic memory (Type-B), it reaches a trough of -76% at n=5 before partially recovering to -10% at n=10. Conventional efficiency and fairness metrics do not reveal these differences. The ALT framework complements the temporal fair division and picking-sequence literature by diagnosing whether temporal fairness emerges spontaneously in decentralized adaptive systems rather than how to enforce it. ---