# Research Papers — 2026-06 · bundle page 208 of 260 > Full-text markdown of papers #10351–10400 of 12993 published in 2026-06, newest first. AI-readable mirrors of the full 12993-paper 2026-06 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-06/page/208 [← Previous bundle](https://aipod.dev/papers/archive/2026-06/page/207.md) [Next bundle →](https://aipod.dev/papers/archive/2026-06/page/209.md) ### An Integrated Roadside Sensing and Communication Framework for Vulnerable Road User Safety at Signalized Intersections Authors: Parvez Anowar Published: 2026-06-05 Categories: stat.AP, cs.CV arXiv: https://arxiv.org/abs/2606.07016v1 Vulnerable road users (VRUs) account for approximately half of urban traffic deaths globally, with intersections concentrating a disproportionate share of these casualties. Recent reviews of sensing technology for VRU protection have cataloged dozens of single-sensor and dual-sensor deployments, yet none of the surveyed systems couples multi-modal sensing with edge-side near-miss analytics and bidirectional vehicle-to-everything (V2X) and pedestrian-to-everything (P2X) messaging in a single intersection cabinet. This paper presents an integrated framework for VRU protection at signalized intersections, combining LiDAR, radar, RGB camera, and thermal camera at the perception layer, edge-based prediction and surrogate-safety analytics at the computation layer, V2X and P2X messaging at the communication layer, and adaptive signal control at the actuation layer. The framework is grounded in an empirical case study using R-LiViT, the first publicly released roadside LiDAR-Visual-Thermal dataset, which provides 200 multi-modal sequences and 2,400 annotated RGB-T frames at three German intersections. Analysis of 53,319 detection annotations reveals that VRUs comprise approximately 49% of all road-user observations, that day-to-night density drops by 38% for pedestrians and 45% for vehicles while the night distribution shows a higher close-proximity share, that per-frame close-proximity event counts vary approximately 10-fold across the eight unique locations at three intersections, and that 83% of pedestrian bounding boxes are small in image space, indicating that VRUs are typically far from any single sensor. These findings support multi-modal sensing, edge-side analytics, and adaptive context-sensitive deployment rather than uniform single-sensor solutions. --- ### Towards Unified Song Generation and Singing Voice Conversion with Accompaniment Co-Generation Authors: Ziyu Zhang, Chunyu Qiang, Xiaopeng Wang, Yuxin Guo, Kang Yin, Wenjie Tian, Jingbin Hu, Tianlun Zuo, Zhao Guo, Teng Ma, Yuzhe Liang, Chen Zhang et al. Published: 2026-06-05 Categories: cs.SD, cs.AI arXiv: https://arxiv.org/abs/2606.07015v2 While song generation and singing voice conversion (SVC) have evolved significantly, they have long been developed isolated: the former lacks zero-shot speaker cloning, while the latter overlooks vocal-accompaniment synergy. To bridge this gap, we propose UniSinger, the first end-to-end framework unifying speaker cloning song generation and accompaniment co-generation SVC. Building on the multimodal diffusion transformer, we construct a unified speaker embedding space transferring speaker representation from SVC to song generation, endowing fine-grained cross-task timbre control. To mitigate multi-task optimization conflicts, we design a curriculum learning strategy using task-specific modality masking to guide the model to gradually master the generative mechanisms among semantic content, vocal timbre, and accompaniment. Experiments show state-of-the-art performance on both tasks and realizes complementary benefits, offering new possibilities for intelligent music production. --- ### A Geometric View for Understanding Concept Learning and Neuron Interpretation in Sparse Autoencoders Authors: Chenhao Zhang, Chris Lin, Su-In Lee Published: 2026-06-05 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.07007v1 We propose a unified mathematical framework for a geometric understanding of concept learning and neuron interpretation in sparse autoencoders (SAEs). While SAEs improve interpretability of neural networks by learning sparse feature representations, a principled definition of ''concept'' and ''learning'' remains unclear. We formalize concepts as sets of data points and cast concept learning as a set-alignment problem between human-defined and model-induced concepts. This formulation distinguishes three increasingly strong notions of learning -- detection, separation, and approximation -- and yields geometric conditions, error bounds, and capacity constraints for when concepts can be represented by individual neurons or multi-neuron units. It also provides a set-theoretic account for common SAE phenomena, including feature splitting, feature absorption, feature families, and hierarchical concepts. Finally, we connect concept learning and neuron interpretation through formal concept analysis, showing that the two directions need not agree and that their many-to-many structure can be organized by concept lattices. Experiments on synthetic data with ReLU and Top-$K$ SAEs illustrate the theory and reveal the effects of SAE size and sparsity on concept learning. --- ### RASFT: Rollout-Adaptive Supervised Fine-Tuning for Reasoning Authors: Yongliang Miao, Fengyuan Liu, Wei Shi, Yanguang Liu, Fei Sun, Na Zou, Mengnan Du Published: 2026-06-05 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2606.07006v1 Supervised fine-tuning (SFT) is a prevailing method for adapting large language models to reasoning tasks by imitating offline expert demonstrations, often treating a single expert trajectory as the target behavior. However, reasoning is not simple path imitation: rigidly following one demonstrated solution may overfit to surface forms and suppress the model's own reasoning distribution. We propose Rollout-Adaptive Supervised Fine-Tuning (RASFT), a policy-aware SFT framework that calibrates expert supervision according to problem-level solvability estimated from verified on-policy rollouts. For each problem, RASFT strengthens expert guidance when the current policy struggles, while relaxing rigid imitation and incorporating correct self-generated trajectories when the model already exhibits reliable reasoning behavior. To preserve useful reasoning priors, RASFT further introduces a clipped inverse ratio between the frozen reference model and the current policy to constrain excessive policy drift. Experiments across multiple models on six mathematical reasoning benchmarks and two code reasoning benchmarks show that RASFT achieves better overall performance than SFT, SFT variants, and representative RL methods. The code is available at https://github.com/zjd1sq/RASFT. --- ### DataEvolver: Automatic Data Preparation for Large Language Models through Multi-Level Self-Evolving Authors: Chao Deng, Shaolei Zhang, Ju Fan, Xiaoyong Du Published: 2026-06-05 Categories: cs.DB, cs.AI arXiv: https://arxiv.org/abs/2606.07001v2 High-quality training data is essential to large language models (LLMs) and typically requires extensive and costly manual curation. Existing automatic data preparation methods rely on predefined pipelines or customized human instructions, which limits their adaptability to diverse data distributions and lacks principled guidance from high-quality examples. In this paper, we introduce DataEvolver, the first self-evolving data preparation system that automatically constructs pipelines to transform raw data into high-quality data. DataEvolver employs a multi-level mechanism to ensure both pipeline executability and effectiveness. At the operator level, it incrementally expands the operator set to construct a logical plan while resolving dependency conflicts. At the pipeline level, it instantiates logical plans into executable code and iteratively refines pipeline orchestration through a feedback loop that reduces the distribution gap between prepared data and high-quality examples. Experiments on seven benchmarks show that DataEvolver substantially improves data quality and achieves an average 10\% gain in downstream LLM performance compared with training on original data, highlighting new opportunities for the iterative co-evolution of LLMs and data. --- ### Teaching the Way, Not the Answer: Privileged Tutoring Distillation for Multimodal Policy Optimization Authors: Shizhe Xiang, Ke An, Wenlong Yu, Yue Liu, Jian Luan, Pei Fu, Qilong Wang Published: 2026-06-05 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.07000v1 Recent post-training methods, particularly Reinforcement Learning with Verifiable Rewards (RLVR), have significantly enhanced the reasoning ability of Large Vision-Language Models (LVLMs). However, the sparse nature of verifiable rewards provides little token-level supervision for failed rollouts, often leading to inefficient exploration in complex multimodal reasoning tasks. Although policy distillation can offer dense guidance, external teacher based methods introduce substantial computational overhead, while answer conditioned tuning methods may expose answer-level information and induce shortcut-like generation behavior. To address these limitations, we propose PTD-PO, a Privileged Tutoring Distillation Policy Optimization framework for RLVR that provides dense guidance without exposing the answer to the student policy. Specifically, PTD-PO constructs structured privileged hints from spatial attention guidance and intermediate textual reasoning steps, and uses them through in-context learning to produce step-wise token-distribution supervision. The student is still optimized under the original answer-free context, and its failed rollouts are aligned with the hint-augmented reference model at the token-distribution level. To further stabilize distillation under the distribution shift between guided and unguided contexts, we introduce a Top-K Jensen-Shannon divergence objective that focuses alignment on informative token probabilities while reducing memory overhead. Experiments on LVLMs ranging from 2B to 8B parameters show that PTD-PO consistently outperforms RLVR and distillation baselines, mitigates entropy collapse, and improves complex multimodal reasoning performance. --- ### Principles of Concept Representation in Sentence Encoders Authors: Isabelle Mohr, John Dujany, Jonathan Souquet, Andre Freitas Published: 2026-06-05 Categories: cs.CL, cs.DB arXiv: https://arxiv.org/abs/2606.06994v1 What makes a sentence encoder produce good concept representations? We approach this through the lens of representational compositionality: an encoder supports a concept family only when its latent space admits a low-distortion realization of the corresponding semantic operator. This framing predicts both where current encoders succeed and where they are structurally mismatched to their supervision. Through a controlled ablation over encoder conditions trained on 3.3 million synonym and definition pairs from WordNet and Wiktionary, evaluated on three decontaminated splits and a modifier-labeled noun-phrase benchmark, we identify four principles. Fine-tuning recalibrates the latent geometry rather than expanding it (P1). Semantic signal concentrates in the final transformer layer before concept-specific training begins, making cross-layer pooling redundant (P2). Hard negatives improve discrimination and stress-test robustness without improving retrieval ranking, showing that calibration and ranking are independently addressable (P3). Finally, the effectiveness of supervision depends on the composition type of the target concept. Extensional training helps intersective and subsective families while degrading relational and intensional ones, exposing a structural limitation of current training paradigms (P4). We release two new evaluation datasets: a DBpedia semantic-gap benchmark and a modifier-labeled NP paraphrase suite. --- ### Don't Pause: Streaming Video-Language Synchrony for Online Video Understanding Authors: Zhenyu Yang, Kairui Zhang, Shengsheng Qian, Weiming Dong, Changsheng Xu Published: 2026-06-05 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.06991v1 Online Video Large Language Models (Video-LLMs) have advanced toward seamless human-AI interaction through frame-by-frame processing and proactive responding. However, a critical challenge remains in streaming scenarios: existing models typically pause video perception while generating responses, breaking real-time video-language synchrony and causing stutters. To address this, we introduce a novel paradigm for online video understanding: Streaming Video-Language Synchrony (SVLS), and present LyraV, a live streaming assistant built upon a hierarchical control framework with two core innovations. First, the Frame-Driven Transition Controller (FDTC), a training-free verification-based finite-state machine, makes high-level semantic decisions on when to continue speaking, start a new response, or stay silent. Second, the Streaming Token Pacer (SToP), a plug-and-play lightweight predictive module, dynamically adapts the language generation rate to match the pace of the visual content. Concretely, LyraV performs \emph{per-frame incremental, sub-budget decoding}: within each frame interval it emits only a small chunk of tokens that fits the real-time budget, so perception is never blocked for a full sentence. Together, these components enable LyraV to seamlessly interleave incoming video frames with generated word tokens, achieving a fine-grained synchrony. Extensive experiments conducted on five online and three offline benchmarks demonstrate that LyraV preserves the backbone's general understanding ability while substantially improving streaming synchrony and narrative fluency, delivering a 98.29\% synchrony with video playback and a real-time processing speed of 3.89 FPS. Interestingly, we observe an empirical capability in LyraV: dynamic reasoning over streaming tokens, enabling continuous interpretation and "thinking" alongside visual input. --- ### Accelerating Reproducible Research in Synthetic EHR Generation Authors: Jalen Jiang, Chufan Gao, Ethan Rasmussen, Stephen Z. Xie, Jimeng Sun Published: 2026-06-05 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.06990v1 The generation of high-fidelity synthetic Electronic Health Records (EHR) is crucial for advancing medical research while preserving patient privacy. However, head-to-head comparison of existing generative models is hindered by disjointed codebases, incompatible data loaders, conflicting library dependencies, and inconsistent evaluation protocols. To address these gaps, we introduce a lightweight, end-to-end benchmarking framework for reproducible synthetic EHR evaluation, organized as a unified pipeline spanning data ingestion, standardized model training, and architecture-agnostic evaluation. Our current implementation targets the generation of longitudinal ICD diagnosis codes -- the most commonly studied modality in this literature -- and is built on the community-maintained PyHealth library. We reimplement and unify strong baselines (MedGAN, CorGAN, PromptEHR, HALO) under full ICD-9 vocabulary granularity, and add a lightweight GPT-2 baseline from the general-purpose sequence-modeling literature. We contribute a rigorous, architecture-agnostic privacy-utility evaluation suite that applies identically to GAN- and transformer-based generators, and report bootstrapped confidence intervals across all metrics. We further analyze the poor long-tailed performance of existing models and discuss the extensibility of our framework beyond diagnosis codes. By lowering the engineering barrier to running, extending, and evaluating under a single pipeline, we introduce a starting point for community-driven reproducibility and benchmarking synthetic EHR models. --- ### Heterogeneous Effects of Green Finance on Urban Decarbonization: Evidence from 285 Cities in China Authors: Xueyang Li, Jinlei Ma Published: 2026-06-05 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.06986v1 While green finance has become a key instrument for low-carbon city transitions, its actual decarbonization effects and transmission mechanisms remain unclear. This study employs econometric models and machine learning-based analysis to examine whether and how green finance reduces city-level carbon intensity. Results show that green finance significantly lowers carbon intensity, with green bonds and green investment having the strongest impacts and evident spatial spillovers. The effects vary by development level, being most pronounced in Fourth- and Fifth-tier cities. Mediation analysis reveals that green finance operates mainly through energy structure optimization, followed by industrial upgrading, foreign direct investment, and technological innovation. SHAP analysis confirms substantial differences across financial instruments, with green bonds, funds, and credit contributing most to decarbonization. Moreover, the marginal impact is stronger in cities with low technological capacity, high industrial dependency, and coal-based energy mixes. These findings provide theoretical support and policy guidance for building a multi-level, regionally differentiated green finance system to promote inclusive low-carbon transitions. Keywords: Green Finance; Carbon Intensity; Decarbonization Effect; Machine Learning; City --- ### Contrastive Training with LLM-generated Near-Misses for Robust Code-Switching Speech Recognition Authors: Tung X. Nguyen, Hieu Minh Truong, Giang Son Nguyen, Nhu Vo, Wray Buntine, Dung D. Le Published: 2026-06-05 Categories: cs.CL, eess.AS arXiv: https://arxiv.org/abs/2606.06985v2 Code-switching (CS), the alternation between multiple languages within a single utterance, remains challenging for Automatic Speech Recognition (ASR). To address this issue, we propose a Point-of-Interest (POI)-aware contrastive training framework that improves recognition at CS-critical regions. We first identify CS spans by adopting POI detection method from literature, then construct acoustically plausible near-miss hypotheses by perturbing POIs in ASR N-best outputs and expanding candidates with a large language model. Hard but plausible negatives are retained through filtering with acoustic, phonemic, and textual constraints. Finally, we fine-tune Whisper-small with LoRA using a POI-weighted cross-entropy anchor objective together with a multi-negative contrastive ranking loss. Experiments on CS-FLEURS (cmn-eng) and ViMedCSS (vie-eng) show consistent reductions of over 2% in both general and CS-aware error rates compared to standard LoRA fine-tuning. --- ### Accelerating Multi-Objective Bayesian Optimisation via Predictive-Gradient Catalysts Authors: Alma Rahat, Tinkle Chugh, Jonathan Fieldsend, Richard Allmendinger Published: 2026-06-05 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.06984v1 This paper presents a general acceleration mechanism for multi-objective Bayesian optimisation (MOBO) that leverages Gaussian process predictive gradients as auxiliary signals. Rather than replacing existing Pareto-compliant acquisition functions, the proposed approach augments them with local stationarity information derived from surrogate-derived gradients, enabling faster convergence toward the global Pareto set under limited evaluation budgets. Two catalyst instantiations are investigated: an adaptive Multiple-Gradient Descent Algorithm-Based Catalyst (MGDA) and a predefined-weight variant that enables focused exploration when budgets are tight. Experiments on the DTLZ benchmark suite (using 2 objectives and 10 decision variables) show that predictive gradient catalysis can deliver significant acceleration compared to other acquisition functions (EHVI, AugTch, tMPoI, SAF) when surrogates are accurate, particularly for stationary problems. --- ### DaX: Learning General Pathology Representations Across Scales Authors: Bokai Zhao, Yiyang Zhang, Long Bai, Tai Ma, Hanqing Chao, Minfeng Xu Published: 2026-06-05 Categories: eess.IV, cs.AI, cs.CV arXiv: https://arxiv.org/abs/2606.06983v1 Computational pathology requires visual representations that transfer across diverse clinical endpoints and remain robust to variation in magnification, staining, scanner type, slide preparation, and input resolution. We present DaX, a pathology vision foundation model that adapts DINOv3-style self-supervised learning to whole-slide histopathology. DaX is initialized from natural-image DINOv3 weights and incorporates continuous magnification training, cross-scale tissue views, orientation-agnostic and acquisition-robust augmentation, multi-input-size training, and Gram-anchored dense consistency. These designs aim to connect local cellular morphology with global tissue architecture while stabilizing dense token-level representations across input scales. We further construct a WSI-level benchmark comprising 161 clinically meaningful tasks from 44 public datasets, covering 28,182 patients and 34,394 slides across four clinical domains and nine task categories. All models are evaluated under a fixed patient-level cross-validation protocol with fold-level statistical ranking, enabling reproducible comparisons that are less sensitive to split-dependent variation. Across this benchmark, DaX achieves the highest mean performance across tasks and consistently strong task-level ranking scores, with gains spanning diagnostic pathology, biomarker and molecular profiling, tissue/specimen context, and risk, response, and prognosis. These results support DaX as a transferable visual encoder for computational pathology and provide a standardized evaluation framework for future pathology foundation models. Project page: https://alibaba-damo-academy.github.io/DaX/benchboard/. --- ### CL-CLIP: CLIP-Based Continual Learning Framework with Cost-Volume Category Decoupling for Object Detection Authors: Zihan Liu, Yuguang Yang, Shengjie Su, Jianing Pang, Linlin Yang, Chunyu Xie, Nikolai Yu. Zolotykh, Baochang Zhang Published: 2026-06-05 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.06978v1 Continual Object Detection (COD) requires a detector to acquire new categories over time while preserving previously learned ones. This goal is closely related to open-vocabulary detection, since both settings require reasoning over categories that are not fully covered by the annotations available at the current training stage. Recent CLIP-based open-vocabulary detectors have shown strong zero-shot generalization, and frameworks such as F-ViT demonstrate that vision-language pretraining can provide powerful zero-shot detection ability for unseen categories. However, real-world deployments cannot remain purely zero-shot: once these detectors are continually updated on newly introduced categories, they suffer severe catastrophic forgetting and quickly lose their previously calibrated detection ability. We therefore propose CL-CLIP, a CLIP-based COD framework that equips open-vocabulary detectors with better continual learning ability through cost-volume-guided category decoupling. Specifically, following CAT-Seg, we compute a CLIP image-text similarity cost volume, defined as dense category-wise response maps between visual tokens and class text embeddings. This zero-shot spatial prior decomposes shared region features into class-specific pathways, which are then processed by a Multi-Expert RoI head. Extensive experiments on PASCAL VOC and MS-COCO show that CL-CLIP substantially improves the F-ViT baseline under continual fine-tuning and achieves competitive performance with existing continual object detectors, especially in adapting to newly introduced categories while preserving competitive base-class performance. --- ### Exploring Agentic Tool-Calling Decisions via Uncertainty-Aligned Reinforcement Learning Authors: Yijin Zhou, Linqian Zeng, Xiaoya Lu, Wenyuan Xie, Dongrui Liu, Junchi Yan, Jing Shao Published: 2026-06-05 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.06976v1 Large language model (LLM)-based agents often make suboptimal tool-use decisions, including unsupported tool invocation and hallucinated direct responses, which may accumulate errors throughout multi-step interactions. Existing approaches mainly improve these behaviors through inference-time correction or coarse-grained reward signals based on decision outcomes and structured checklists, leaving the uncertainty characteristics of agent decisions underexplored. We observe that decision-oriented reinforcement learning tends to weaken the uncertainty separation between correct and incorrect actions, resulting in overconfident mistakes and weaker exploration signals. Therefore, we propose TRUST, which incorporates uncertainty quantification into reward design as a repulsive force for maintaining uncertainty separation, and labels lightweight key-turn annotations for unified post-training of multi-turn trajectories. Experimental results across diverse tool-use benchmarks show that TRUST consistently enhances both decision quality and agent performance while maintaining more reliable uncertainty estimates during optimization. --- ### Accounting for Context: Shaping Moral Credences for Value Alignment Authors: Jazon Szabo, Sanjay Modgil Published: 2026-06-05 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.06972v1 Ensuring that agent behaviours are aligned with human moral values inevitably raises the problem of how to account for the plurality of moral perspectives that societies -- and even individuals -- typically adopt. Work on moral uncertainty proposes mechanisms to fairly and democratically aggregate evaluations of actions across different moral theories. However, this paper argues that one needs to account for contextual factors when aggregating moral evaluations. For example, consequentialist perspectives assume an ability to accurately determine how an agent's actions change the world; an assumption that often does not hold in real world settings. We, therefore, formalise agent decision making under moral uncertainty, while also accounting for these kinds of contextual factors. We thereby show that a seemingly commonsensical property -- the weak Pareto principle -- is violated. We argue that this apparent problem is, in fact, a variation of Simpson's paradox, and hence reveals the limitations of aggregation mechanisms that ignore the impact of contextual factors. --- ### GenPO++: Generative Policy Optimization with Jacobian-free Likelihood Ratios Authors: Ke Hu, Shutong Ding, Panxin Tao, Jingya Wang, Ye Shi Published: 2026-06-05 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.06967v1 Generative policies provide expressive and multimodal action distributions, making them attractive for reinforcement learning (RL) in complex continuous-control tasks. Among them, flow-based policies are especially appealing because they generate actions through deterministic transport maps. However, applying such generative policies to likelihood-based on-policy learning remains limited by the difficulty of evaluating the probability of executed actions. Existing flow RL methods either replace the true action-density ratio with approximate surrogates, which can introduce biased updates, or recover exact likelihoods through dummy-action augmentation, which enlarges the policy space and increases computation. In this work, we propose GenPO++, a reversible generative policy optimization framework that uses history states as auxiliary memory in a high-order reversible ODE solver, yielding exact inversion without changing the original action dimension. The resulting generative policy map has a log-determinant determined only by fixed solver coefficients, enabling exact and Jacobian-free likelihood-ratio computation. This design preserves the expressiveness of generative flow policies while avoiding both action ratio bias and dummy-action overhead. We evaluate GenPO++ on large-scale simulated control, fine-tuning, and real-world robotic manipulation tasks, where it achieves competitive or superior performance over state-of-the-art on-policy RL methods, while improving training stability and computational efficiency. --- ### From Vision to Text: A Compact Multimodal Approach for Robust, Cross-Domain Presentation Attack Detection on ID Cards Authors: Qingwen Zeng, Juan E. Tapia, Sneha Das, Christoph Busch Published: 2026-06-05 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.06966v1 Cross-domain shifts challenge Presentation Attack Detection (PAD) on ID Cards, given the restricted data available due to privacy concerns. This work proposes a compact multimodal model, based on new generative and discriminative blocks, which combines visual and textual data for PAD on genuine and synthetic ID images. While multimodal models exhibit strong generalisation after supervised fine-tuning, they fail in zero-shot settings. Our findings underscore that model capacity and real-world data are essential for reliable PAD, while existing synthetic datasets may not reflect real-world challenges. We argue for a re-evaluation of synthetic data as a benchmark and emphasise the need for more realistic, diverse datasets to advance PAD research. --- ### FinEvolveBench: A Benchmark for Self-Evolving Agents on Low-Repetition Tasks with Implicit Rewards Authors: Yining Zhu, Zihao Deng, Leiming Wang, Jingfei Lu, Junbo Wang, Yuke Li, Jikun Shen, Chuncheng Ran Published: 2026-06-05 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.06960v4 Large language model (LLM) agents increasingly rely on external experience to continually adapt to changing environments without modifying their underlying models. Recent experience mechanisms have demonstrated promising results across diverse tasks. However, their effectiveness is typically evaluated within individual benchmark settings, and how experience mechanisms generalize across different scenarios remains insufficiently explored. In this work, we present a scenario-oriented analysis of experience mechanisms for LLM agents. We characterize existing evaluation scenarios along four dimensions: outcome observability, credit assignment complexity, environmental dynamics, and experience reusability. Our analysis shows that existing benchmarks often evaluate experience mechanisms under scenarios where at least one dimension is comparatively favorable, leaving more challenging combinations of scenario properties underexplored. To address this gap, we introduce \textsc{FinEvolveBench}, a reproducible benchmark built on a chronological stream of rich financial news and market data that enables systematic evaluation of experience-based self-evolution under challenging experience regimes characterized by noisy feedback, ambiguous credit assignment, environmental non-stationarity, and limited experience reusability. Experiments show that existing approaches exhibit substantially reduced or inconsistent gains in this setting, highlighting the scenario-dependent nature of experience mechanisms and the challenge of maintaining valid experience under changing environments. --- ### Tree-of-Experience: A Structured Experience-Management Solution for Self-Evolving Agents under Low-Repetition and Implicit-Reward Environments Authors: Zihao Deng, Yining Zhu, Leiming Wang, Jingfei Lu, Junbo Wang, Chuncheng Ran, Yu Yang, Dixuan Yang, Jikun Shen Published: 2026-06-05 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.06960v1 Experience-based self-evolution is crucial for LLM agents, but existing benchmarks often assume explicit goals, stable task patterns, and clear feedback. We study a more challenging setting: low-repetition tasks with implicit rewards, where past experience is difficult to reuse and feedback is delayed, noisy, and outcome-level. We introduce \textsc{FinEvolveBench}, a temporally controlled benchmark for financial sentiment prediction that links daily news-driven predictions to future excess returns. We further propose Tree-of-Experience (ToE), a structured experience-management method that organizes, retrieves, validates, and updates agent experience. Experiments show that general-purpose experience mechanisms do not consistently outperform no-experience baselines, while ToE achieves stronger overall performance. These results highlight the importance of structured experience management for self-evolving agents in implicit-reward environments. --- ### OpenHalDet: A Unified Benchmark for Hallucination Detection across Diverse Generation Scenarios Authors: Xinyi Li, Zhen Fang, Yongxin Deng, Jinyuan Luo, Hongnan Ma, Changdae Oh, Zijing Shi, Shanshan Ye, Hanchen Wang, Shu-Lin Chen, Yadan Luo, Mengyue Yang et al. Published: 2026-06-05 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2606.06959v1 Hallucination detection is essential for the reliable deployment of large language models (LLMs). However, existing evaluations face two core challenges: inconsistent inference configuration and evaluation, and limited coverage of downstream domains and tasks. Consequently, reported detector performance is often difficult to compare, reproduce, and generalize beyond specific experimental settings. We introduce OpenHalDet, a unified benchmark for hallucination detection across diverse generation scenarios. OpenHalDet standardizes the evaluation pipeline, from prompt construction and response generation to truthfulness annotation, detector scoring, and metric computation. It supports heterogeneous detector families under different access settings, including black-box methods that use only generated outputs, gray-box methods that rely on probability-based signals, and white-box methods that exploit internal model signals. By bringing diverse tasks, models, and detectors into a shared framework, OpenHalDet enables controlled comparison and provides a systematic view of how different detection paradigms behave in LLM applications. We release OpenHalDet as an open and extensible codebase to facilitate reproducible evaluation and future development of hallucination detection methods. The code and datasets are available at https://github.com/Nellie179/Hallucination-Detection. --- ### MVSegNet: A Lightweight Boundary-Aware Network for Fetal Lateral Ventricle Segmentation and Atrial Width Estimation in Prenatal Ultrasound Authors: Arafat Hossain Sayem Published: 2026-06-05 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.06958v1 Fetal ventriculomegaly is assessed by measuring the atrial width of the lateral ventricle in prenatal ultrasound. Accurate segmentation is essential for this measurement, but acoustic shadowing, speckle noise, and poor contrast make it difficult. We developed MVSegNet, a lightweight encoder-decoder network combining multi-scale feature extraction and boundary-aware refinement. The model was trained and evaluated on 584 expert-annotated transventricular ultrasound frames using a 70/15/15 split. Performance was compared against six segmentation baselines using overlap, boundary, and measurement metrics. MVSegNet achieved a Dice score of 80.79%, IoU of 68.47%, Hausdorff distance of 4.07 mm, and atrial width mean absolute error of 3.40 mm. The model contains 2.31 million parameters and runs at 165.6 frames per second on an NVIDIA T4 GPU. MVSegNet outperformed all evaluated baselines on boundary and measurement metrics while maintaining low computational cost, supporting its use in automated fetal ultrasound analysis. --- ### Deep Single-Index Fréchet Regression Authors: Muqing Cui, Yidong Zhou, Su I Iao, Hans-Georg Müller Published: 2026-06-05 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2606.06957v1 Predicting outputs that are located in non-Euclidean spaces, such as probability distributions, networks, and symmetric positive-definite matrices, is becoming increasingly important in modern data analysis, particularly when inputs are high-dimensional. We propose DeSI (Deep Single-Index Fréchet Regression), a semiparametric framework for regression with metric space-valued outputs and multivariate inputs that assumes a single-index structure for the conditional Fréchet mean. DeSI estimates an interpretable index direction, which quantifies the relative importance of inputs, using a deep neural network, and performs Fréchet regression along the resulting one-dimensional index in the target metric space. This structure mitigates the curse of dimensionality while retaining interpretability, which stands in contrast to standard deep neural networks. We establish theoretical guarantees for DeSI, including uniform approximation and convergence rates, and demonstrate its strong predictive performance through simulations on distributions, networks, and symmetric positive-definite matrices, as well as an application to compositional mood data from New Jersey. --- ### When is 3D Worth It? A Resource-Performance Frontier for CNNs and Transformers in Lung CT Authors: Md Enamul Hoq, Sharafat Hossain, Imraul Emmaka, Linda Larson-Prior, Lawrence Tarbox, Jonathan Bona, Donald Johann Jr. and Fred Prior Published: 2026-06-05 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.06950v1 Three-dimensional models are widely assumed preferable for volumetric medical imaging, yet their practical value depends on whether performance gains justify added computational cost and complexity. Rather than proposing a new architecture, we study how input dimensionality (2D, 2.5D, 3D) affects model behavior across convolutional neural networks (CNNs) and Vision Transformers (ViTs) under a fixed training protocol. Using a leakage-free NLST cohort (n = 1,977) with supporting LIDC-IDRI data, we find that the 2.5D CNN offers the most favorable discrimination-stability trade-off in our comparison (ROC-AUC 0.682, 95% CI (0.546, 0.799)) with a stable operating point. In contrast, 3D CNNs show threshold instability, and transformers exhibit degenerate predictions, such as all-positive predictions. Confidence intervals are wide and overlapping, so we present these results as a controlled resource-performance frontier and a failure-mode taxonomy rather than as definitive superiority claims. For class-imbalanced lung cancer screening classification, 2D and 2.5D inputs provide a more reliable trade-off between performance, stability, and computational efficiency than full 3D representations. --- ### Auditing Training Data in Domain-adapted LLMs: LoRA-MINT Authors: Gonzalo Mancera, Daniel DeAlcala, Aythami Morales, Julian Fierrez, Ruben Tolosana, Francisco Jurado Published: 2026-06-05 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2606.06946v1 We present LoRA-MINT, a new methodology for Membership Inference Test (MINT) applied to recent Large Language Models (LLMs) fine-tuned for specific Natural Language Processing (NLP) tasks through Low-Rank Adaptation (LoRA). The primary goal is to assess whether individual samples were part of the training data of these adapted models, providing a useful auditing tool for the management of intellectual property and sensitive data. Our analysis explores the relationship between model perplexity and membership status, providing a systematic framework for estimating data exposure in fine-tuned LLMs. We conducted experiments on four models and three benchmark datasets, obtaining precision values in determining if given data were used for training ranging from 0.77 to 0.92, which outperform state-of-the-art baselines and demonstrate the robustness and generality of the proposed method. In general, our findings underscore the potential of LoRA-MINT as an effective and scalable framework for auditing LLMs, improving transparency, and fostering the ethical and responsible deployment of AI and NLP technologies. For the sake of concreteness and current relevance, our discussion and experiments are centered on LoRAadjusted LLMs, but note that most of the presented methodology is easily applicable for auditing training data given any other technique for adapting LLMs or, more generally, any other domain-adapted AI models. --- ### SS-TPT: Stability and Suitability-Guided Test-Time Prompt Tuning for Adversarially Robust Vision-Language Models Authors: Sunoh Kim, Daeho Um Published: 2026-06-05 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.06943v1 Vision-language models (VLMs) such as CLIP achieve strong zero-shot recognition but remain highly fragile under adversarial perturbations. Recent test-time adaptation defenses improve robustness by leveraging many augmented views, but this leads to impractical slowdown and a clear robustness-throughput trade-off. To address this challenge, we present Stability and Suitability-guided Test-time Prompt Tuning (SS-TPT), evaluating the quality of each augmented view via two complementary scores: (1) stability, measuring prediction invariance to weak augmentations, and (2) suitability, measuring feature-space density among views. These stability and suitability (SS) scores guide both adaptation and inference through an SS-guided consistency loss and an SS-weighted prediction, amplifying trustworthy views while suppressing corrupted ones. Extensive experiments demonstrate that SS-TPT significantly outperforms prior state-of-the-art methods, achieving superior robustness-throughput trade-offs across diverse datasets and varying numbers of views, thereby demonstrating both strong practicality and generality. Our code is available at https://github.com/sunoh-kim/SS-TPT. --- ### Didact: A Cross-Domain Capability Discovery System for Defence Authors: Aarya Bodhankar, Aditya Joshi, Bao Gia Doan, Thomas Marchant, Oscar Leslie, Flora Salim Published: 2026-06-05 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2606.06942v1 Policymakers in defence and defence-aligned sectors must monitor rapidly evolving research alongside sector priorities relevant to operational and strategic needs. In practice, these sources are fragmented across heterogeneous formats, disjoint repositories, and siloed update streams, making capability discovery slow and difficult to audit. We present Didact, a prototype that integrates publicly available defence reports and policy documents from Australia with a purpose-built knowledge graph derived from Australian research publications. Didact provides natural language conversations for policy-oriented workflows, and leverages a composite retrieval-augmented generation (RAG) pipeline. A key feature of Didact is an interactive Evidence Rail that visualises retrieved evidence and source relationships. Our evaluation of the output quality and runtime of Didact highlights its utility. While Didact has been co-developed as an academia-industry project for the Australian context, it is adaptable to other domains where knowledge is similarly fragmented. A demonstration video is available here: --- ### Quantum-Inspired Trace-Augmented Evidence Selection for Reasoning over Structured Hypothesis Spaces Authors: Laura Wynter, Nirvik Sahoo, Paul Griffin Published: 2026-06-05 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.06941v1 Large language models (LLMs) now solve a wide range of expert-level exams at or above human level, yet remain brittle on specialised, evidence-intensive domains such as law. On these tasks, errors arise not only from gaps in world knowledge but also from subtle distinctions between pieces of evidence and inconsistent use of supporting evidence. The most common aggregator over sampled chain-of-thought (CoT) traces, majority vote, returns the most popular answer regardless of whether its evidence is actually strongest. We propose to treat the selection of CoT reasoning fragments into a set of evidence as an explicit combinatorial optimisation problem, allowing well-supported but minority hypotheses to override noisy majorities, and to evaluate the approach on legal-reasoning benchmarks that are particularly sensitive to evidence quality. We introduce EP-HUBO (Evidence Pool Higher-Order Binary Optimisation), which generates multiple CoT traces with a small local model, parses fragments into per-hypothesis evidence pools, solves a higher-order unconstrained binary optimisation per pool with quality-derived weights (relevance, specificity, distinctiveness), and delegates a single adjudication call per question to a frontier model. We evaluate EP-HUBO on two evidence-intensive legal benchmarks using both simulated annealing on classical hardware and the Dirac-3 photonic entropy-quantum machine from Quantum Computing Inc. HUBO-style optimisation gives a principled way to aggregate reasoning fragments while preserving minority-but-correct hypotheses, and is most valuable in low-contamination domains where frontier models have not already absorbed the benchmark material. --- ### When CLIP Sees More, It Fights Back Harder: Multi-View Guided Adaptive Counterattacks for Test-Time Adversarial Robustness Authors: Sunoh Kim, Daeho Um Published: 2026-06-05 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.06938v1 Vision-language models such as CLIP have achieved remarkable zero-shot recognition capabilities, yet their robustness against adversarial perturbations remains limited. Test-time counterattack (TTC) was recently proposed to improve CLIP's robustness by perturbing an input image to steer it away from a corrupted state during inference. However, TTC remains fragile under strong attacks because its counterattack relies on a directly corrupted original view and employs a noise-driven hard-gating scheme that cannot adapt to varying corruption severity. To address these limitations, we introduce Multi-view guided Adaptive Counterattack (MAC), which performs counterattacks for multi-view with corruption-aware soft weighting. Specifically, MAC first constructs augmented views of an input image to obtain diverse embeddings. It then performs counterattacks to refine corrupted embeddings of views. Next, MAC adaptively scales the counterattack intensity for each view based on its estimated corruption degree. Finally, the adaptively counterattacked views are aggregated to yield a robust final prediction. Extensive experiments across 20 datasets and diverse attack scenarios demonstrate that MAC substantially improves robustness while preserving high inference speed and memory efficiency with its tuning-free design. Our code is available at https://github.com/sunoh-kim/MAC. --- ### Uniform Stability and Generalization Error of GD and SGD on Fixed-Point Parameters Authors: Jonghyun Shin, Sejun Park Published: 2026-06-05 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.06934v1 We analyze generalization error, uniform stability, and uniform argument stability of gradient descent (GD) and stochastic gradient descent (SGD) over discrete parameter spaces, where each update involves deterministic or stochastic rounding. We show that deterministic rounding degrades the generalization error of GD on convex, Lipschitz, and smooth loss functions, increasing the rate from $O(T/n)$ to $O(T/\sqrt{n})$, and establish matching lower bounds. We further prove that uniform stability of GD becomes $Ω(T)$, showing that stability-based generalization bounds are vacuous in this setting. In contrast, for the same losses, stochastic gradient descent with deterministic rounding admits nontrivial uniform stability guarantees, which differ qualitatively from the real-valued case and exhibit distinct dependencies on the number of iterations and the dimension: we prove tight bounds $O(T/n)$ for one dimension and $O(T^2/n)$ for higher dimensions. We also show that stochastic rounding can introduce generalization error that increases with the dimension; such a phenomenon is absent in standard real-valued optimization and in the deterministic rounding case. Finally, we provide upper bounds on uniform argument stability for stochastic rounding schemes and show that these bounds are tight when the loss can be represented as a sum of coordinate-wise functions. --- ### SVHighlights: Towards Extremely Long Sport Video Highlight Detection Authors: Donggyu Lee, Youngbin Ki, Jeonghun Kang, Taehwan Kim Published: 2026-06-05 Categories: cs.CV, cs.MM arXiv: https://arxiv.org/abs/2606.06926v2 While highlight detection for long-form videos is of great practical importance, most existing methods remain limited to short-form content, largely due to the absence of a suitable benchmark. To bridge this gap, we introduce SVHighlights, to the best of our knowledge, the first benchmark for highlight detection in extremely long sports videos, each exceeding one hour in duration, across multiple sports categories. SVHighlights is constructed from pairs of full-length sports videos and their corresponding official highlight videos using a dataset generation pipeline, enabling scalable label generation without conventional per-clip saliency annotation. The benchmark comprises 320 videos with an average duration of 2.00 hours and a total of 640.18 hours, substantially exceeding previous datasets. Existing methods also face fundamental challenges on long videos: models trained on short clips fail to generalize to hour-long content, and their clip-level scoring lacks the broader context needed to identify highlights. To address this and provide a strong baseline, we present TF-SELECTOR, a training-free segment-based approach that divides each video into context-aware segments by merging adjacent shots sharing the same semantic content, and predicts segment-level saliency scores using a large language model with multimodal inputs including visual captions, transcripts, and audio volume. Experiments demonstrate that TF-SELECTOR achieves superior performance across most metrics compared to Video Temporal Grounding (VTG)-tuned baselines, with improvements of +2.50 in HIT@1, +4.04 in HIT@K, and +2.95 in IoU. These results establish SVHighlights as a challenging testbed for long-form highlight detection and demonstrate that a simple segment-based strategy can effectively scale to hour-long videos. --- ### From Sampled Outcomes to Capability Distributions: Rethinking Supervision for LLM Routing Authors: Guannan Lai, Haoran Hu, Long Chen, Zhenguo Li, Han-Jia Ye Published: 2026-06-05 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.06924v2 Existing LLM routing methods often construct supervision from a single sampled response for each query--model pair. Because LLM generation is stochastic, however, such an observation can be an unstable estimate of model capability: semantically equivalent query formulations and repeated decoding may yield different scores and even different model preferences. We show that this instability can further propagate from routing labels to learned routing policies. To address this issue, we propose DARS (Distribution-Aware Routing Supervision), which estimates query-level model capability from repeated observations spanning semantics-preserving query rewrites and stochastic decoding. DARS summarizes expected quality, expected cost, and performance variability to construct risk-aware supervision without changing the downstream router architecture. Experiments across diverse tasks and routing methods show that DARS generally improves routing utility and cost--quality trade-offs over single-shot supervision. Further analyses show that its benefits persist under moderate sampling budgets and different decoding temperatures. These results suggest that reliable LLM routing should move beyond individual sampled outcomes and instead model query-level capability distributions. --- ### From Sampled Outcomes to Capability Distributions: Rethinking Supervision for LLM Routing Authors: Guannan Lai, Haoran Hu, Long Chen, Zhenguo Li, Han-Jia Ye Published: 2026-06-05 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.06924v1 Existing LLM routing methods typically treat a model's single response to a query as its capability label for training routers. However, because LLM generation is inherently stochastic, such single-shot supervision provides only a noisy observation of a query-model pair's behavior rather than a reliable capability estimate. We show that this assumption introduces systematic noise into routing supervision, making learned routing policies less reliable. To address this issue, we propose DARS (Distribution-Aware Routing Supervision), a framework that constructs routing supervision from a distributional view of model behavior. Instead of relying on a single generated response, DARS considers uncertainty from both the input side and the output side, capturing how semantically equivalent query formulations and stochastic generations affect model performance. Based on these distribution-aware observations, DARS builds more reliable supervision signals for routing. Experiments across diverse tasks show that single-shot labels can be misleading for model selection, while distribution-aware supervision provides more stable labels and improves learned routing behavior. Our results suggest that reliable LLM routing should move beyond single-response observations and be grounded in query-level model capability distributions. --- ### Declarative Skills for AI Agents in Knowledge-Grounded Tool-Use Workflows Authors: M. Danish Lim, I. Danial Bin Sharudin, Wen Han Chen, Cedric Lim, Laura Wynter Published: 2026-06-05 Categories: cs.AI, cs.SE arXiv: https://arxiv.org/abs/2606.06923v1 We study orchestration mechanisms for tool-using AI agents in realistic customer-service workflows over an unstructured knowledge base. We argue that declarative agents -- AI agents equipped with natural-language skill files appended to the system prompt -- are an effective orchestration paradigm. Concretely, we compare (i) a DeclarativeAgent that reads three domain-specific skill files at inference time and decides its own control flow, (ii) an ImperativeAgent based on a programmatic state machine with explicit phases, and (iii) an unscaffolded baseline agent modeled after the $τ$-Knowledge benchmark agent. Our ImperativeAgent is motivated by externalised-control inference as in Recursive Language Models and graph-based orchestration frameworks. We formalise the three agents as policy classes within a decentralised partially-observable Markov decision process and analyse their information-theoretic and structural properties; we then test the predicted differences empirically on five language models and two retrieval regimes. Our results show that retrieval quality is a dominant bottleneck for AI agents: when evidence is incomplete or skewed, all agents degrade substantially, and skill files cannot recover lost performance. Under high-quality retrieval, however, declarative skills consistently improve accuracy on procedural tasks and reduce orchestration errors, while the imperative state machine's brittleness does not reliably improve task success or compliance. --- ### The Fine-Tuning Trap: Evaluating Negative Transfer and the Role of PEFT in Sub-1B Mathematical Reasoning Authors: Rahul Nair, Chun Tao Published: 2026-06-05 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.06920v1 Deploying Small Language Models (SLMs) on edge devices requires efficient fine-tuning strategies that adapt models to new tasks without degrading their general capabilities. In this study, we benchmark five sub-1B models (135M-1B) on mathematical reasoning tasks and uncover a critical vulnerability: Full Fine-Tuning (Full FT) actively harms performance in models under 300M parameters, often dropping accuracy below zero-shot baselines. This "negative transfer" makes Parameter-Efficient Fine-Tuning (PEFT) not just an efficiency preference, but a stability requirement. We find that while Low-Rank Adaptation (LoRA) and Weight-Decomposed LoRA (DoRA) perform comparably, their strengths vary by task; DoRA excels in complex reasoning (GSM8K), while LoRA dominates pattern matching (OrcaMath). In particular, Full FT is outperformed by LoRA on aligned models (Qwen2.5-0.5B) and even by simple 5-shot In-Context Learning on the smallest architectures (SmolLM2-135M). Based on these findings, we recommend defaulting to PEFT for all aligned sub-1B models and caution against Full FT for any architecture smaller than 500M parameters to prevent catastrophic forgetting. Reproduction of this work can be found at https://github.com/gulguluu/tiny-slm-finetune-compare. --- ### DRIFT: From Robustness Gaps to Invariance Manifolds for AI-Generated Image Detection Authors: Abhishek Ameta, Sayan Banerjee, Shreyas Pandith, Harshit, Ankita Chatterjee, Akshay Janardan Bankar, Amit Satish Unde Published: 2026-06-05 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.06918v1 The rapid evolution of generative image models challenges existing AI-generated image detectors, particularly in open-world settings with unseen generators. Recent training-free approaches measure robustness gaps in frozen vision foundation models (VFMs), detecting fakes via perturbation-induced embedding drift. However, these methods rely on fixed invariance geometry inherited from pretraining and lack principled adaptation to the detection task. We instead formulate AI-generated image detection as learning a structured invariance manifold of real images under one-class supervision. Building upon a frozen VFM, we introduce lightweight projection heads that decompose representation space into complementary robust and fragile subspaces. The robust subspace is explicitly trained to suppress variations induced by physically plausible imaging transformations, approximating tangent directions of a real-image manifold, while the fragile subspace retains sensitivity to edit-like perturbations. A structured ordering margin enforces hierarchical separation between physical invariance and edit-induced variability, enabling detection as a margin-violation test relative to the learned manifold. At inference, multi-scale patch-wise drift under both transformation families yields a dual-channel invariance signature and interpretable localization. Extensive experiments demonstrate strong open-world generalization across unseen generators and resolutions, consistently outperforming training-free robustness-based baselines while providing interpretable invariance-violation maps. --- ### ThinkBooster: A Unified Framework for Seamless Test-Time Scaling of LLM Reasoning Authors: Vladislav Smirnov, Chieu Nguyen, Sergey Senichev, Minh Ngoc Ta, Ekaterina Fadeeva, Artem Vazhentsev, Daria Galimzianova, Nikolai Rozanov, Viktor Mazanov, Jingwei Ni, Tianyi Wu, Igor Kiselev et al. Published: 2026-06-05 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2606.06915v2 Test-time compute (TTC) scaling has emerged as a powerful paradigm for improving large language model (LLM) reasoning by allocating additional compute during inference, e.g., via multi-sample generation and verifier-based reranking. Existing TTC scaling strategies and reasoning scorers remain fragmented, evaluated under inconsistent protocols, and are rarely analyzed through the lens of quality-cost trade-offs. We introduce ThinkBooster, a unified framework for seamless test-time compute scaling of LLM reasoning, which consists of (i) a modular Python library implementing state-of-the-art TTC scaling strategy and scorer families, (ii) a benchmark that jointly evaluates performance and computational efficiency, and (iii) a deployable OpenAI-compatible proxy service that enables drop-in integration of adaptive reasoning into real-world applications. We further provide a demo visual debugger for inspecting the reasoning trajectories, intermediate selection decisions, and alternative reasoning paths. Empirical results on mathematical and coding tasks reveal the performance-compute trade-offs of TTC scaling strategies and scoring methods and demonstrate that ThinkBooster provides practical gains in real-world tasks. The code is available online under an MIT license. --- ### polyDAG: Polynomial Acyclicity Constraints for Efficient Continuous Causal Discovery in Visual Semantic Graphs Authors: Wenhao Zhang, Ramin Ramezani, Tao Han, Kai Hwang, Minyi Guo Published: 2026-06-05 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.06908v2 Modern image-analysis pipelines often convert images into structured semantic variables, such as facial attributes, object concepts, and scene descriptors. Learning directed dependencies among these variables can produce interpretable visual semantic graphs, but continuous directed acyclic graph learning is limited by the cost of enforcing acyclicity. We present polyDAG, a polynomial acyclicity framework for efficient continuous causal discovery in visual semantic graphs. polyDAG replaces the matrix-exponential acyclicity constraint with a finite polynomial trace constraint and proves that the new constraint is zero exactly for acyclic graphs. We further derive a geometric-series implementation that avoids the explicit summation loop while preserving the same acyclicity condition. Experiments on synthetic Erdos-Renyi graphs and CelebA facial visual attributes show that polyDAG improves efficiency and structure recovery. Averaged over the revised synthetic protocol with d in {100, 200, 500}, polyDAG reduces mean structural Hamming distance from 318.4 to 285.4 and improves mean F1 score from 0.725 to 0.756. At 100 nodes, the geometric variant runs in 3.44 seconds compared with 5.16 seconds for the exponential baseline, corresponding to a 33.4 percent speedup. Code and data are publicly available at https://github.com/wenhaoz-fengcai/polyDAG. --- ### SpectCount: Spectrotemporal Counting via Synthetic Signals Improves Large Audio Language Models Authors: Seonuk Kim, Yonghyeon Jun, Ju Yeon Kang, Jimin Hong, Yoonhyeong Lee, Nam Soo Kim Published: 2026-06-05 Categories: eess.AS, cs.AI, cs.SD arXiv: https://arxiv.org/abs/2606.06907v1 Large audio language models (LALMs) extend large language models with an audio encoder and large-scale audio data. However, the scarcity of high-quality annotated audio data remains a fundamental bottleneck for scaling. Through probing signal detectability analysis, we identify fine-grained spectrotemporal perceptual weaknesses in a foundation LALM. To address these challenges, we propose Spectrotemporal Counting (SpectCount), a data-efficient fine-tuning approach based on fully synthetic audio signals generated on-the-fly, without relying on real-world audio, annotations, or pretrained generative models. SpectCount not only resolves the observed weaknesses but also improves performance on diverse auditory benchmarks spanning sound, music, and speech, unseen during fine-tuning. These results suggest that weakness-targeted synthetic signals provide a data-efficient path toward enhanced auditory understanding capabilities in LALMs. --- ### EASE-TTT: Evidence-Aligned Selective Test-Time Training for Long-Context Question Answering Authors: Xiaopeng Yuan, Zebin Wang, Suwen Wang, Zongxin Yang, Haohan Wang, Yushun Dong Published: 2026-06-05 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2606.06906v1 Long-context question answering (QA) remains challenging for smaller language models even when answer-bearing evidence is already present in the input. Existing within-context retrieval methods localize and expose candidate evidence chunks for the question, but they stop at input-level evidence exposure rather than adapting the query-side attention parameters that control how the model allocates attention over full-context positions. In contrast, lightweight test-time adaptation methods, such as query-only test-time training (qTTT), leave evidence localization unresolved because their generic span-level self-supervised objectives do not identify which context positions support the current answer. In this paper, we propose Evidence-Aligned SElective Test-Time Training (EASE-TTT), a within-context retrieval-augmented test-time training framework that converts selected evidence chunks into a soft attention supervision target over their token positions. Instead of replacing the full context with retrieved chunks, EASE-TTT uses the resulting attention target to guide query-side adaptation, with the adapted model generating the final answer from the original full context. Experiments on six LongBench QA tasks and three small decoder-only language models show that EASE-TTT achieves the strongest macro-average performance among full-context inference, retrieval-only baselines, and qTTT, supporting evidence-aligned test-time adaptation in long-context QA. --- ### ActionMap: Robot Policy Learning via Voxel Action Heatmap Authors: Pei Yang, Hai Ci, Yanzhe Chen, Qi Lv, Han Cai, Mike Zheng Shou Published: 2026-06-05 Categories: cs.RO, cs.CV arXiv: https://arxiv.org/abs/2606.06904v2 Vision-language-action (VLA) models have advanced rapidly across backbones, training recipes, and data scale, yet the action decoder, which converts the backbone's hidden state into a continuous control signal, has barely changed and remains a single-point predictor across the majority of current VLAs. Whether implemented via autoregressive token bins, L1 regression, or flow-matching denoising, the resulting decoder treats the action space as unstructured, leaving the geometric proximity of neighboring actions unexploited during training. To advance this, we introduce ActionMap, a voxel heatmap action head that drops into an existing VLA in place of its native action decoder. For each new action, the head predicts a voxel heatmap over the action space, where each voxel directly stores the probability of the corresponding action. Across LIBERO simulation and real-world Franka manipulation, our heatmap head surpasses two architecturally distinct backbones at matched training steps (e.g., +8.2% over OpenVLA-OFT's L1 regression head on the LIBERO four-suite average), converges at comparable or faster rates on both backbones, and remains markedly more data-efficient at low training data. The cross-backbone consistency indicates that action representation is a real lever for VLA performance, distinct from further backbone or recipe scaling. Project Page: https://showlab.github.io/ActionMap/. --- ### Beyond Skeletons: Learning Animation Directly from Driving Videos with Same2X Training Strategy Authors: Yuan Zeng, Yujia Shi, Yuhao Yang, Dongxia Liu, Zongqing Lu, Wenming Yang, Qingmin Liao Published: 2026-06-05 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.06903v1 Human image animation aims to generate a video from a static reference image, guided by pose information extracted from a driving video. Existing approaches often rely on pose estimators to extract intermediate representations, but such signals are prone to errors under occlusion or complex poses. Building on these observations, we present DirectAnimator, a framework that bypasses pose extraction and directly learns from raw driving videos. We introduce a Driving Cue Triplet consisting of pose, face, and location cues that captures motion, expression, and alignment in a semantically rich yet stable form, and we fuse them through a CueFusion DiT block for reliable control during denoising. To make learning dependable when the driving and reference identities differ, we devise a Same2X training strategy that aligns cross-ID features with those learned from same-ID data, regularizing optimization and accelerating convergence. Extensive experiments demonstrate that DirectAnimator attains state-of-the-art visual quality and identity preservation while remaining robust to occlusions and complex articulation, and it does so with fewer computational resources. Our project page is at https://directanimator.github.io/. --- ### TALAN: Task-Aligned Latent Adaptation Networks for Targeted Post-Training of Large Language Models Authors: Chengkai Zhang, Ziteng Liu, Junpu Wang, Zeyi Tao, Yang Wang, Sagar Chordia, Qin Huang Published: 2026-06-05 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.06902v1 Targeted post-training aims to improve reasoning, math, and code without degrading strengths. Low-rank adapters are efficient but task-global; activation interventions are input-aware but often require separate probes, vectors, or inference-time steering. We introduce TALAN (Task-Aligned Latent Adaptation Networks), a sequence-conditioned latent side path inserted into a transformer's residual stream and co-trained with a low-rank adapter in one SFT loop. TALAN compresses the active sequence into latent memory, remixes it into token-level perturbations, and writes them back through a controlled residual update. It is configured along six axes: insertion location, memory size, mixer, writeback rule, trainability scope, and gradient scale. Across four Qwen3-family backbones and four STEM/code benchmarks, TALAN improves matched LoRA and DoRA baselines. With LoRA, it yields a +1.41 point cross-model mean gain, positive on all four backbones and non-negative on all 16 model-benchmark cells. With DoRA, it yields a +1.85 point mean gain, positive on all backbones and on 13 of 16 cells. Paired seed checks support positive average effects but show nontrivial variance, so we treat them as sensitivity checks. Cost is small: <1% trainable parameters relative to the backbone and 1.01-1.02x inference overhead versus matched LoRA. A Llama-3.2-1B transfer probe is also positive under LoRA and rsLoRA across seven paired seeds, supporting a transfer beyond Qwen. Internal-state analyses suggest TALAN is a small complementary activation intervention. The matched adapter update is 80-1,700x larger than the TALAN perturbation, yet their directions have near-zero cosine; per-layer measurements show this small orthogonal perturbation propagates and amplifies through depth. TALAN offers a practical platform for studying steerable activation-level adaptation within standard adapter-based post-training. --- ### LUCID: Learning Unified Control for Image Deflaring and Exposure Mastery in Nighttime Photography Authors: Tingyu Yang, Yuan Cheng, Xiaoyun Yuan Published: 2026-06-05 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.06901v1 Photography is the art of painting with light, yet nighttime scenes are shaped by competing degradations: intense flares obscure scene structure, while photon-limited regions collapse into noise. Conventional approaches address these factors in isolation, overlooking the fact that these degradations are fundamentally entangled. To bridge this gap, we introduce LUCID, a unified framework that reframes nighttime restoration as a continuous and controllable process rather than a fixed correction. We decompose nighttime restoration into two cooperative components: a flare disentanglement module that lifts the 'curtain' of optical artifacts to provide reliable structural guidance, and a diffusion-driven module that leverages generative priors to reconstruct clean and well-exposed imagery. Crucially, LUCID introduces explicit controllability through a novel four-mode training strategy, enabling users to steer the restoration process via classifier-free guidance (CFG) and allowing selective control over light sources and their associated flare and ghosting artifacts, while also supporting high dynamic range (HDR) reconstruction through continuous exposure control. Extensive experiments demonstrate that LUCID consistently outperforms state-of-the-art methods across diverse real-world nighttime scenarios. --- ### Lighting-Aware Representation Learning under Controllable Lighting Variation Authors: Lizhen Zhu, Charantej Reddy Pochimireddy, James Z Wang, Brad Wyble Published: 2026-06-05 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2606.06899v1 Variations in illumination remain a major challenge for visual representation learning, as they induce substantial appearance changes both across and within environments. While existing approaches typically address this issue through data augmentations that encourage models to become invariant to lighting changes, such strategies do not explicitly model lighting information during learning. Inspired by theories of human vision, we propose a lighting-aware representation learning framework that incorporates illumination variation as an explicit training signal rather than a nuisance factor to be suppressed. Our method extends contrastive learning by introducing an auxiliary objective that captures illumination-dependent variation in rendered scenes, enabling the model to jointly learn representations that preserve semantic consistency while remaining sensitive to lighting-dependent visual structure. We evaluate the proposed model on image classification and object detection tasks across the ImageNet, ExDark, and PASCAL VOC benchmarks. Results demonstrate that the proposed lighting-aware training consistently improves downstream performance over standard contrastive learning baselines, while maintaining the same architecture and training budget. Furthermore, our approach shows promising performance in supervised learning frameworks and under settings involving simpler lighting variation, suggesting broad applicability beyond complex illumination scenarios. These results indicate its potential to enhance model robustness and adaptability in complex visual environments as well as in more conventional image processing tasks. --- ### Blockchain Infrastructure for Intelligent Cyber--Physical--Social Systems:Post-Quantum Security, Interoperability, and Trustworthy Data Economies in the Era of Embodied AI Authors: Song Guo, Huawei Huang, Dongping Liu, Aoyu Zhang, Luyao Zhang Published: 2026-06-05 Categories: cs.CR, cs.AI, cs.CY, cs.ET arXiv: https://arxiv.org/abs/2606.06895v1 The deployment of embodied artificial intelligence via world-model-based robotics presents a transformative opportunity for blockchain infrastructure, establishing urgent demand for trustworthy data provenance, cross-organizational governance, and incentive-compatible sharing across decentralized ecosystems. Simultaneously, quantum computing advances recognized by the 2025 Nobel Prize in Physics and the Turing Award threaten the cryptographic primitives securing these data economies, creating an interdependent imperative: long-lived verification for embodied AI depends on crypto-agile architectures capable of withstanding quantum adversaries. This tutorial examines blockchain as the coordination layer bridging this dual transition, from financial substrate to foundational Cyber-Physical-Social Systems infrastructure that simultaneously secures against quantum cryptanalysis and enables scalable, trustworthy data economies. The session opens with an immersive AWS Braket demonstration engaging participants with superconducting, trapped-ion, and neutral-atom hardware to assess cryptographic threat timelines and witness ECDSA-to-post-quantum signature transitions. Five integrated modules progress from embodied AI and world-model requirements through quantum hardware reality and evidence-based security migration, to scalable cross-shard architectures via BrokerChain protocols, trustworthy data economies implementing Croissant metadata standards and robotic learning provenance, and industry ecosystem integration for multi-modal cloud deployment. By bridging quantum hardware realities with embodied AI data requirements, this tutorial charts blockchain as unified infrastructure for next-generation decentralized intelligent environments, providing open-source frameworks and roadmaps for architecting quantum-resistant, interoperable, and data-trustworthy systems. --- ### Workflow-to-Skill: Skill Creation via Routing-Workflow-Semantics-Attachments Decomposition Authors: Yuyang Zhang, Xinyuan Han, Xudong Jiang, Run Wang Published: 2026-06-05 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.06893v1 Large language model agents increasingly rely on Skills to encode procedural knowledge, yet high-quality Skills remain costly to hand-write. This paper studies automatic Skill construction from heterogeneous interaction evidence, including demonstrations, agent trajectories, tool traces, and execution logs. We argue that trace-to-skill construction is not simple summarization tasks, because traces are fragmented, redundant, and may miss rare but safety-critical behaviors. To address this, we introduce RWSA, a workflow-oriented intermediate representation that decomposes Skills into Workflow structure, execution Semantics, and runtime Attachments, capturing task decomposition, control flow, verification, safety, rollback, and state management. Building on RWSA, we propose W2S, a framework that segments traces, induces local Skill drafts, aligns shared structures, reconciles branches, and compresses redundancy while preserving evidence and confidence annotations. Experiments on 70 Skills show that W2S improves behavioral replay consistency by 10.5% over summarization- and prompting-based baselines, highlighting the need to treat traces as executable runtime specifications rather than compressible text. --- ### GRASP: Geometry-aware Residual Alignment for Scalable Pretraining Data Attribution Authors: Yue Min, Ruining Chen, Yujun Li Published: 2026-06-05 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.06892v2 Scalable data attribution methods typically assign isolated utility scores to individual training examples. This prevalent additive assumption fundamentally fails to capture critical subset dynamics, including data redundancy and complementary coverage. In this work, we reframe attribution as subset-level counterfactual utility prediction and introduce GRASP, an interaction-aware surrogate. Grounded in a theoretical smoothness lower bound, GRASP explicitly models subset interactions through a quadratic geometric penalty. To achieve pretraining-scale efficiency without relying on hidden oracle tuning, we couple low-dimensional feature sketches with a strictly finite lower-confidence bound selection protocol. Extensive subset-retraining evaluations demonstrate that GRASP decisively outperforms existing scalable baselines. It more than doubles the task-level rank correlation for counterfactual subset fidelity while reducing upfront artifact construction costs by nearly an order of magnitude. Downstream diagnostics further show that this scoring mechanism transfers to language model curation and cross-domain vision selection, establishing a robust foundation for optimizing massive pretraining corpora. --- ### GRASP: Geometry-aware Residual Alignment for Scalable Pretraining Data Attribution Authors: Yue Min, Ruining Chen, Yujun Li Published: 2026-06-05 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.06892v1 Scalable data attribution methods typically assign isolated utility scores to individual training examples. This prevalent additive assumption fundamentally fails to capture critical subset dynamics, including data redundancy and complementary coverage. In this work, we reframe attribution as subset-level counterfactual utility prediction and introduce GRASP, an interaction-aware surrogate. Grounded in a theoretical smoothness lower bound, GRASP explicitly models subset interactions through a quadratic geometric penalty. To achieve pretraining-scale efficiency without relying on hidden oracle tuning, we couple low-dimensional feature sketches with a strictly finite lower-confidence bound selection protocol. Extensive subset-retraining evaluations demonstrate that GRASP decisively outperforms existing scalable baselines. It more than doubles the task-level rank correlation for counterfactual subset fidelity while reducing upfront artifact construction costs by nearly an order of magnitude. Downstream diagnostics further show that this scoring mechanism transfers to language model curation and cross-domain vision selection, establishing a robust foundation for optimizing massive pretraining corpora. --- ### Stream3D-VLM: Online 3D Spatial Understanding with Incremental Geometry Priors Authors: Hanxun Yu, Xuan Qu, Lei Ke, Boqiang Zhang, Yuxin Wang, Jianke Zhu, Dong Yu Published: 2026-06-05 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.06891v1 Despite advances in 3D scene understanding, existing 3D Large Multimodal Models operate in offline settings, requiring complete scene observations or predefined video clips. In this paper, we present an online 3D vision-language model that enables real-time spatial understanding from streaming video. Our approach adopts an autoregressive streaming control modeling based on the LLM's next-token prediction objective to learn when to respond, and employs a lightweight Visual-Spatial Feature Integration (VSFI) module to incrementally inject temporally aligned geometry priors into the visual stream. To alleviate long-context decoding overhead, we propose a plug-and-play Geometry-Adaptive Voxel Compression (GAVC) module for efficient visual token compression. To address the scarcity of streaming 3D-language data, we further develop a scalable data generation pipeline that curates over 1M online spatio-temporal 3D QA pairs and establishes a comprehensive benchmark spanning 29 tasks. Extensive experiments show that our approach significantly outperforms both proprietary and open-source models across online and offline 3D spatial understanding, reasoning, and grounding tasks. The project page is available at https://stream3d-vlm.github.io/ ---