Artificial Intelligence · Cross-listed from cs.CL
Who Should Lead Decoding Now? Tracking Reliable Trajectories for Ensembling Masked Diffusion Language Models
Heecheol Yun, Joonhyung Park, Joowon Kim et al. (+1 more)
Masked
Diffusion Language Models (MDLMs) have emerged as a distinct paradigm for sequence generation. As MDLMs become diverse in capabilities and knowledge coverage, an important question is how to combine their knowledge. Toward this, we first investigate the unique decoding dynamics of MDLMs. We find that successful generations exhibit stable confidence dynamics over answer-relevant positions, while unreliable trajectories can often be corrected by injecting promising intermediate states from other models. Guided by this observation, we propose $\textbf{TIE}$ ($\textbf{T}$rajectory-based $\textbf{I}$terative $\textbf{E}$nsembling), a knowledge fusion framework in which MDLMs iteratively identify reliable decoding trajectories and relay them across models. TIE tracks confidence dynamics over answer-relevant positions to determine which model currently follows a more reliable trajectory and selectively transfers partially denoised sequences across models. As the model on the more promising trajectory often changes across denoising steps, TIE allows different models to contribute complementary strengths at different stages of generation. Strong performance across diverse
reasoning tasks, along with our analyses, suggests that TIE offers a practical approach to the underexplored problem of MDLM ensembling.
Read the full abstract ▾
All authors Heecheol Yun, Joonhyung Park, Joowon Kim, Eunho Yang
Artificial Intelligence · Cross-listed from cs.CV
RealityBridge: Bridging Editable 3D Gaussian Splatting Driving Simulations and Real-World Videos
Zhenhua Wu, Yun Pang, Mingkun Chang et al. (+4 more)
Long-tail hazardous scenarios are essential for safety-oriented autonomous driving, yet they are difficult to collect at scale. Editable 3D Gaussian Splatting (3DGS) simulation offers a scalable alternative through real-scene reconstruction and controllable editing. However, edited 3DGS-rendered
videos often exhibit a significant Sim-to-Real gap, manifested as rendering artifacts, degraded foreground assets, illumination mismatch, and temporal flickering. Addressing these coupled defects requires jointly restoring local appearance, harmonizing edited content, and maintaining temporal consistency, whereas existing methods typically address only a subset of these requirements. To fill this gap, we propose RealityBridge, a
video restoration and harmonization framework that converts edited 3DGS renderings into realistic driving footage while preserving simulator-defined structure, edits, and dynamics. RealityBridge conditions a
video foundation model on complementary modality signals, with a lightweight GateNet adaptively controlling their injection across backbone blocks. We further develop a task-oriented curation pipeline to construct training data, and design a four-stage supervised training strategy followed by reward-guided post-training. Extensive experiments demonstrate that RealityBridge outperforms existing methods in restoration and harmonization while preserving strong temporal consistency.
Read the full abstract ▾
All authors Zhenhua Wu, Yun Pang, Mingkun Chang, Yuwei Ning, Liangzhi Wang, Yi Xiao, Guanbin Li
Artificial Intelligence
SpecAlign: Efficient Specification-Grounded Alignment of Large Language Models via Synthetic Data
Wenjie Wang, Yue Huang, Zhengqing Yuan et al. (+6 more)
As large
language models (LLMs) are increasingly deployed in real-world applications,
alignment is no longer governed by a single universal notion of safety or helpfulness, but instead by provider- or application-specific model specifications. These specifications are typically long, structured, and frequently updated, yet existing
alignment pipelines lack a systematic mechanism to operationalize them as training signals. In this paper, we propose specification-grounded
alignment, a new
alignment paradigm that treats provider-authored model specifications as the primary
alignment target rather than abstract principles or static
benchmarks. To instantiate this paradigm, we introduce SpecAlign, a framework that synthesizes
alignment data directly from specification documents. SpecAlign combines structured rule annotation, controllable specification instantiation, and multi-
agent adversarial data synthesis to generate fine-grained, boundary-aware preference pairs that capture both compliant behaviors and meaningful specification violations. Experiments across multiple model specifications and backbone models demonstrate that training with SpecAlign consistently improves rule compliance while preserving general capabilities and avoiding over-conservative behavior. These results suggest that grounding
alignment in explicit model specifications enables rapid, precise, and scalable adaptation of LLM behavior to evolving policy requirements.
Read the full abstract ▾
All authors Wenjie Wang, Yue Huang, Zhengqing Yuan, Han Bao, Shiyi Du, Yuchen Ma, Yue Zhao, Yanfang Ye, Xiangliang Zhang
Artificial Intelligence
UXBench: Measuring the Actionability of LLM-Generated UX Critiques
Wenjie Wang, Yue Huang, Zipeng Ling et al. (+11 more)
Large
language models (LLMs) are increasingly deployed as UX judges that inspect interfaces, diagnose usability problems, and propose repairs. Yet no controlled
benchmark measures whether the resulting critiques are reliable and actionable across heterogeneous product surfaces. We introduce UXBench, a
benchmark for evaluating LLMs as interaction-grounded UX judges. UXBench comprises local-first runnable web fixtures spanning ten product-surface families, paired with coverage-gated browser exploration that forces models to collect interaction evidence before reporting. Each judge model produces a structured UX report over seven rubric dimensions; report quality is measured by whether a fixed downstream repair
agent can improve the interface based on the critique. We evaluate eight frontier models under both an automated repair-lift protocol and a blind human validation study. Results show that UX judging is neither saturated nor one dimensional: models differ meaningfully in report actionability, exhibit distinct rubric-level repair signatures, vary in fixture-level reliability, and trade leadership across surface categories
Read the full abstract ▾
All authors Wenjie Wang, Yue Huang, Zipeng Ling, Han Bao, Hang hua, Xiaonan Luo, Yu Jiang, Shiyi Du, Yuexing Hao, Xiaomin Li, Yuchen Ma, Dianzhuo Wang, Yanfang Ye, Xiangliang Zhang
Artificial Intelligence · Cross-listed from cs.LG
Variance Reduction for Non-Log-Concave Sampling with Applications to Inverse Problems
M. Berk Sahin, Ahmet Ege Tanriverdi, Behzad Sharif et al. (+1 more)
Sampling from high-dimensional, non-log-concave distributions with unnormalized densities is a fundamental challenge in machine learning, particularly when the exact gradient of the potential is unavailable and must be approximated via
stochastic gradients that exhibit high variance under a fixed budget of gradient computations per iteration. Although variance reduction techniques such as SGD with momentum, STORM, and PAGE have demonstrated improved convergence properties in non-convex optimization, their implications for sampling from non-log-concave distributions remain largely unexplored. In this work, we develop the first unified analysis of these
estimators for sampling from non-log-concave distributions. We establish improved non-asymptotic convergence rates in $\varepsilon$-relative Fisher information and, under a Poincaré inequality assumption, in squared total variation distance, and further prove weak convergence to the target distribution. We extend our analysis to solving inverse problems with score-based generative priors. We empirically validate our theory and demonstrate that, under a fixed gradient computations per iteration, variance-reduction techniques consistently improve sample quality in two standard imaging applications.
Read the full abstract ▾
All authors M. Berk Sahin, Ahmet Ege Tanriverdi, Behzad Sharif, Abolfazl Hashemi
Artificial Intelligence · Cross-listed from cs.CV
Learned Image Compression for Vision-Language-Action Models
Hyeonjun Kim, Jegwang Ryu, Sangbeom Ha et al. (+4 more)
Vision-
language-action (VLA) models increasingly rely on high-frequency multi-camera observations, making visual communication a major bottleneck for real-time robotic control in bandwidth-constrained or distributed deployment settings. Existing image and
video codecs, however, are designed to preserve generic visual fidelity rather than the control performance of downstream VLA policies. In this work, we introduce SPARC (SPatially Adaptive Rate Control), a learned image compression framework tailored for VLA-driven
robots. Our key observation is that the importance of visual information varies substantially across both camera views and spatial regions within an image. Based on this observation, SPARC employs a lightweight temporal mask selector that adaptively allocates bitrate over latent representations according to task relevance while leveraging temporal context. We further introduce a tilted rate loss that stabilizes training by reducing the tendency of
entropy-based objectives to over-suppress rare yet task-critical visual patterns. Experiments on diverse robotic
benchmarks, including RoboCasa365, VLABench, and LIBERO, show that SPARC consistently achieves stronger control performance than conventional image/
video codecs and recent learned compression methods under the same bitrate budget. We additionally demonstrate real-world deployment benefits in remote-control settings, where our method substantially improves the bitrate-success tradeoff.
Read the full abstract ▾
All authors Hyeonjun Kim, Jegwang Ryu, Sangbeom Ha, Junhyeok Lee, Jun-Hyuk Kim, Hyemin Ahn, Jaeho Lee
Artificial Intelligence · Cross-listed from cs.LG
Demystifying Training-Time Augmentation for Data-Constrained Language Model Pretraining
Michael K. Chen, Xikun Zhang, Fan Bai et al. (+2 more)
As AI labs approach a data ceiling where compute capacity outpaces the rate of new high-quality text generation,
language model pretraining is shifting toward a data-constrained, compute-abundant regime that demands productive multi-epoch training on fixed corpora. Standard autoregressive (AR) pretraining overfits severely in this setting, reaching its optimum early and then continuously deteriorating. We investigate training-time data augmentation as a regularizer to mitigate this overfitting and enable productive training for hundreds of epochs on the same data. We introduce three orthogonal categories of augmentation for AR pretraining:
token-level noise (masking, random replacement), sequence permutations (right-to-left prediction, Fill-in-the-Middle), and target offset prediction ($x_{t+i}$ for $i > 1$). Through systematic ablations, we find that individual augmentations delay overfitting and lower validation loss relative to the baseline, with random
token replacement achieving the best minimum loss among individual methods. Combining augmentation categories further lowers the minimum validation loss. Our experiments demonstrate that data augmentations mitigate AR pretraining's data inefficiency and offer a promising solution to the data-constrained regime~\footnote{All code and data are available at https://github.com/ michaelchen-lab/ data-augmentations-for-pretraining.
Read the full abstract ▾
All authors Michael K. Chen, Xikun Zhang, Fan Bai, Zhengding Hu, Zhen Wang
Artificial Intelligence · Cross-listed from cs.CR
Activating Latent Security Knowledge through LLM-Guided Risk Analysis for Secure Code Generation
Xiaoyun Xu, Lichao Wu, Jona te Lintelo et al. (+3 more)
Large
language models are pretrained on extensive software and security corpora, yet they frequently generate functionally correct code containing well-known vulnerabilities. Existing defenses commonly treat this behavior as a knowledge deficit, addressing it through model-specific fine-tuning or
retrieval from large collections of vulnerability code examples. We argue that insecure generation can arise from a failure to activate task-relevant security knowledge, rather than from knowledge absence alone. We present BRACE, an inference-time, training-free security harness for risk-conditioned activation of security knowledge in black-box code generation. Given only a coding task, BRACE first utilizes an LLM (as a security expert) to identify task-relevant security risks. The predicted risk identifiers are validated against a lightweight canonical catalog and converted into concise, task-specific risk cues. BRACE then supplies these cues to the target code model, prompting it to apply its own secure-coding knowledge while satisfying the original functional requirements. The framework requires no target-model fine-tuning, parameter access, hidden states, mutable logits, learned retriever, or coding-example knowledge base. We evaluate BRACE on six open-weight models and six frontier commercial models with four
benchmarks (CyberNative, HumanEval, CWEval, BaxBench), measuring functional correctness, security, joint functionality-security, and project-level generation. BRACE raises the average Safe Code Rate from 33.9% under Secure
Prompt to 91.2% on CyberNative. It also improves CWEval Func-Sec performance from 49.7% to 69.6%. Mechanism-oriented ablations further show that the gains depend on selecting task-relevant risks.
Read the full abstract ▾
All authors Xiaoyun Xu, Lichao Wu, Jona te Lintelo, Siyu Zhang, Keke Gai, Stjepan Picek
Artificial Intelligence · Cross-listed from cs.CV
Propagating Structural Guidance: Synthesizing Fluorescein Angiography from Fundus Images and Sparse OCT Scans
Tengfei Ma, Ruiqi Wu, Chenran Zhang et al. (+6 more)
Fundus fluorescein angiography (FFA) is critical for assessing retinal vascular abnormalities, but its acquisition is invasive and not always feasible. In contrast, color fundus photography (CFP) is non-invasive and widely accessible, which has motivated studies on CFP-to-FFA synthesis. However, prior works rely solely on CFP surface texture, fundamentally limiting the ability to reconstruct functional vascular information and subtle pathological changes. To address this, we propose a novel framework that synthesizes FFA from CFP with structural guidance provided by optical coherence
tomography (OCT). We construct a multi-modal retinal imaging
dataset with paired CFP, FFA, and OCT from 3,676 patient eyes–the first tri-modally aligned
dataset in retinal imaging. To bridge the spatial gap between OCT and fundus modalities, we propose a Spatially Aligned Cross-Modal Fusion (SACMF) module that projects depth-resolved OCT features onto the fundus plane and injects them into the CFP encoder via adaptive layer normalization. Beyond feature fusion, we further introduce
Token-wise Cross-Modality
Alignment (TCMA), a
token-level contrastive learning strategy that explicitly aligns CFP and FFA representations at corresponding spatial positions. Our method achieves superior synthesis performance compared to state-of-the-art methods. Moreover, extensive experiments demonstrate that the FFA images synthesized by our approach bring greater improvements in downstream disease diagnosis performance than existing methods, highlighting the clinical potential of our approach as a non-invasive decision-support tool in routine workflows. The code is available at https://github.com/while-plus/OCT-guide-FFA-Syn.
Read the full abstract ▾
All authors Tengfei Ma, Ruiqi Wu, Chenran Zhang, Ye Geng, Na Su, Xiangyuan Duanmu, Tao Zhou, Yi Zhou, Wen Fan
Artificial Intelligence · Cross-listed from cs.LG
From Tokens to Regions: CUDA-Sensitive Instruction Tuning for GPU Kernel Generation
Wentao Chen, Jiace Zhu, Xing Zhe Chai et al. (+4 more)
High-performance CUDA kernels are essential for scalable AI systems, while Large
Language Models (LLMs) still struggle to generate correct kernels due to strict and implicit execution constraints. Existing LLM-based approaches either rely on costly agentic or reinforcement-learning (RL) pipelines, or adopt supervised fine-tuning (SFT) objectives that fail to explicitly model CUDA sensitivity, namely code
tokens or regions tightly coupled with execution constraints. In this work, we investigate CUDA sensitivity from the perspective of
token confidence patterns, showing that CUDA sensitivity appears at both
token and region levels, where most CUDA-sensitive
tokens are predicted with high confidence, while a smaller low-confidence subset forms regions corresponding to execution-critical structures. These findings suggest that effective CUDA kernel generation should both leverage high-confidence CUDA-sensitive
tokens and preserve low-confidence CUDA-sensitive regions. Building on these insights, we propose \textbf{\underline{CU}DA-\underline{Se}nsitive Instruction \underline{T}uning (CuSeT)}, a low-cost post-training method within a simple SFT framework. CuSeT follows the principle of “from
tokens to regions” by combining adaptive
token-level masking with region-aware sample reweighting. Experiments show that CuSeT consistently improves functional correctness across multiple model families and scales, outperforming standard SFT and advanced SFT variants, while achieving competitive performance against frontier CUDA kernel generation models with substantially lower inference cost.
Read the full abstract ▾
All authors Wentao Chen, Jiace Zhu, Xing Zhe Chai, Zeng Qu, Qiaoling Xiao, Liucheng Duan, An Zou
Artificial Intelligence
Latent Thought Flow: Efficient Latent Reasoning in Large Language Models
Xiandong Zou, Jing Huang, Jianshu Li et al. (+1 more)
Large
Language Models (LLMs) increasingly rely on intermediate
reasoning, yet explicit Chain-of-Thought (CoT) suffers from a linguistic space bottleneck: each thought must be decoded into
tokens, causing high inference overhead. Latent
reasoning moves deliberation into continuous space, but existing methods mostly learn deterministic or reward-maximizing paths, lacking a principled way to allocate probability across trajectories with different correctness and costs. We propose Latent Thought Flow (LTF), which models
reasoning as variable-length continuous trajectories and trains a sampler to match a reward-induced posterior over answer quality and computation cost. We instantiate this with a continuous GFlowNet using
stochastic latent transitions. To handle sparse answer supervision, we introduce an
Entropy-Weighted Subtrajectory Balance objective for intermediate rewards and a reference-prior regularizer to anchor exploration. Experiments under finetuning and transfer learning settings show that LTF outperforms explicit CoT and latent
reasoning baselines, improving accuracy by 9.5% while reducing
reasoning length by 27.2% on average compared with strong latent
reasoning baselines.
Read the full abstract ▾
All authors Xiandong Zou, Jing Huang, Jianshu Li, Pan Zhou
Artificial Intelligence · Cross-listed from cs.CL
PACT: Privileged Trace Co-Training for Multi-Turn Tool-Use Agents
Zhenbang Du, Jun Luo, Zhiwei Zheng et al. (+8 more)
Multi-turn tool-use
agents must reason, call tools, and adapt to observations across several interaction turns. Post-training such
agents is challenging, as reinforcement learning often suffers from sparse rewards and weak credit assignment despite matching the
prompt-only inference setting, while supervised fine-tuning on expert traces provides dense process supervision but can over-constrain the model to fixed trajectories. To tackle this, we propose PACT, a Privileged trAce Co-Training framework for multi-turn tool-use
agents. The key idea is to use expert traces only as training-time optimization signals rather than rollout-time hints. PACT keeps rollout generation
prompt-only, then uses expert traces to guide optimization through two complementary signals: a trace-conditioned RL surrogate that evaluates
prompt-only rollouts under expert-trace context, and a component-aware SFT loss that provides annealed process supervision over
reasoning and tool-call components. To reduce over-reliance on the training-only trace context, PACT incorporates
prompt-only anchoring with standard
prompt-only RL updates. Experiments on FTRL, BFCL, and ToolHop show that PACT consistently outperforms strong SFT- and RL-based baselines. Further analyses support the complementarity of the two training objectives and the benefits of component-aware supervision and
prompt-only anchoring. These results highlight the effectiveness of privileged trace co-training for multi-turn tool-use learning. The project page is available at https://zhenbangdu.github.io/pact-project-page.
Read the full abstract ▾
All authors Zhenbang Du, Jun Luo, Zhiwei Zheng, Xiangchi Yuan, Kejing Xia, Dachuan Shi, Qirui Jin, Qijia He, Shaofeng Zou, Yingbin Liang, Wenke Lee
Artificial Intelligence · Cross-listed from cs.LG
Calibrated Sampling-Free Uncertainty Estimation in Bayesian Deep Learning
Tobias Jan Wieczorek, Leon de Andrade, Thomas Möllenhoff et al. (+1 more)
Modern deep learning models remain notoriously prone to overconfidence, limiting their reliability in high-stakes applications.
Bayesian methods aim to counter this by learning a distribution over model parameters, and recent advances now make this feasible for large-scale architectures at costs comparable to AdamW. However, a challenge remains at test time: predictions must be averaged across many forward passes with weights sampled from the posterior, which is prohibitively expensive. Variance propagation offers an efficient alternative, computing layer-wise analytical approximations of uncertainty in a single forward pass. While such techniques are effective for MLPs, their extension to modern architectures remains challenging, due to increased depth and diversity of layer types. To fill this gap, we propose Calibrated Variance Propagation (CVP), which introduces a new propagation method for normalization layers, combines it with recent techniques for handling activation functions, and absorbs residual error through a light calibration step. CVP yields comparably accurate uncertainty estimates to MC sampling across
transformers and CNNs, at a fraction of the cost. Against prior variance propagation work, CVP improves coverage at $0.5\%$ risk from $8.2\%$ to $14.6\%$ with BEiT-3 on Visual
Reasoning (NLVR2) and from $2.6\%$ to $10.8\%$ with ViLT on VQAv2, with gains extending to convolutional architectures.
Read the full abstract ▾
All authors Tobias Jan Wieczorek, Leon de Andrade, Thomas Möllenhoff, Marcus Rohrbach
Artificial Intelligence · Cross-listed from cs.CV
LUCID: Learned Undersampling-Adaptive Consistency-Guided Inference with Deterministic Flow Matching for Sparse-View CT Reconstruction
Jigang Duan, Jiayi Wang, Heran Wang et al. (+3 more)
Sparse-view CT reduces radiation dose and scanning time by acquiring fewer projection views, but angular undersampling makes reconstruction severely ill-posed, causing streak artifacts, structural blurring, and loss of fine details. Existing supervised methods are often tied to specific sampling settings, whereas generative methods may introduce anatomically inconsistent
hallucination-like structures under severe undersampling. We propose Lucid, a sparsity-adaptive, consistency-guided reconstruction framework based on a Flow Matching generative prior for sparse-view CT. Lucid is trained only on high-quality CT images to learn a continuous transport between a Gaussian distribution and the high-quality CT image distribution, independent of view sampling. During inference, the sampling sparsity level is explicitly incorporated to adapt the generative trajectory of a single pretrained model. Specifically, Lucid constructs a degradation-matched initial state by sparsity-weighted fusion of the sparse-view FBP image and Gaussian noise, performs sparsity-modulated Flow Matching updates, and applies projection-domain data-consistency correction after each prior update. Experiments under multiple sparse-view settings show that Lucid achieves stable reconstruction performance across different sampling densities, improves image quality and structural fidelity, and reduces the risk of
hallucination-like structures in generative sparse-view CT reconstruction.
Read the full abstract ▾
All authors Jigang Duan, Jiayi Wang, Heran Wang, Ping Yang, Genwei Ma, Xing Zhao
Artificial Intelligence
Sensor-Conditioned Representation Learning via Scene-Relevant Observation Quotients
Yan Jiao, Pin-Han Ho and Limei Peng
Learned representations in intelligent sensing systems are often evaluated by reconstruction fidelity or downstream prediction accuracy, but these criteria do not specify which latent distinctions are justified by the sensing process. In sensor-conditioned environments, nuisance factors can change measurements without changing the scene, while distinct scenes may be indistinguishable under limited sensing capability. This paper formulates sensor-conditioned representation correctness as preserving sensing-supported scene distinctions while suppressing nuisance-induced and sensor-unsupported variation. We introduce the scene-relevant observation quotient, a representation target induced by sensing-supported distinguishability after nuisance canonicalization, and develop Observation-Quotient Tucker-Structured Autoencoding (OQ-TSAE), a scene-nuisance factorized framework with diagnostics for false distinction, false merge, nuisance sensitivity, and latent ordering consistency. Experiments on a controlled
benchmark show that quotient-consistent supervision improves representation-correctness diagnostics over reconstruction-oriented, metric-learning, and contrastive-learning baselines. Sensitivity, perturbation, and ablation studies show the importance of quotient-aligned supervision, reliable quotient relations, and quotient geometry. Complementary real-radar experiments show that a reconstruction-only OQ-TSAE variant retains competitive downstream utility, robustness under observation degradation, and low seed-to-seed variability. These results suggest that sensor-conditioned representations should be evaluated not only by predictive utility, but also by whether their latent geometry preserves sensing-justified scene distinctions.
Read the full abstract ▾
Artificial Intelligence
Beyond Helpfulness: A Teaching-over-Solving Diagnostic for Measuring Educational Impact in LLM Tutors
Junyi Yao, Zihao Zheng and Baichuan Li
Large
language models are increasingly proposed as educational tutors, yet stronger task-solving ability does not necessarily imply stronger learning support. Motivated by recent calls to measure the social impact of NLP systems in practice, we study whether public LLM tutoring
benchmarks distinguish learning-supportive behavior from mere answer production. We propose a lightweight diagnostic based on the gap between solving-oriented and pedagogy-oriented
benchmark performance. Using public MathTutorBench leaderboard results, we show that these dimensions are only partially aligned: across eight publicly reported models, the correlation between solving and pedagogy composites is 0.421, and several models shift meaningfully in rank when evaluation moves from solving to pedagogy. We then analyze the public TutorBench sample and show that agency-relevant behaviors are explicitly encoded in
benchmark rubrics, especially in active-learning settings that reward guiding questions, calibrated hints, and non-disclosive scaffolding. Together, these findings suggest that educational-impact evaluation should not treat task success as a sufficient proxy for learning support. We argue that public tutoring
benchmarks can better support positive-impact evaluation by reporting solving-oriented and pedagogy-oriented scores separately and by making disclosure-sensitive, student-agency-preserving criteria more explicit.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.CV
EgoPhys: Learning Generalizable Physics Models of Deformable Objects from Egocentric Video
Hyunjin Kim, Ri-Zhao Qiu, Guangqi Jiang et al. (+1 more)
Humans naturally understand object physics through everyday interactions, but faithfully predicting complex deformable dynamics, such as elastic materials and fabrics, remains a major challenge for computer
vision and robotics. We present EgoPhys, a framework that constructs deformable physical digital twins from
egocentric RGB-only
video using generalizable priors. EgoPhys overcomes the limitations of existing methods to enable controllable deformable digital twin generation from
egocentric videos by distilling per-object inverse-physics solutions into a compact codebook, enabling prediction of dense spring stiffness fields for unseen objects without per-spring test-time optimization. Trained with generalizable priors from diverse
egocentric interactions, EgoPhys outperforms baselines in reconstruction, future prediction, and zero-shot generalization. To support training and evaluation, we curate an
egocentric interaction
dataset covering diverse deformable objects, scenes, and manipulation styles. We deploy EgoPhys on a real xArm6
robot, demonstrating that a digital twin initialized from a single
egocentric human play
video can serve as an internal world representation to aid in deformable-object planning, highlighting
egocentric RGB observations as a scalable path toward real-to-sim pipelines.
Read the full abstract ▾
All authors Hyunjin Kim, Ri-Zhao Qiu, Guangqi Jiang, Xiaolong Wang
Artificial Intelligence · Cross-listed from cs.CV
SAE++: Cascaded Sparse Autoencoders Learn Multi-Level Visual Concepts in Multimodal LLMs
Yusong Zhao, Hengyi Wang, Tanuja Ganu et al. (+2 more)
Multimodal Large
Language Models (MLLMs) have demonstrated strong performance on
vision-
language tasks, yet their internal visual representations remain difficult to interpret. Sparse Autoencoders (SAEs) provide a scalable way to decompose dense model activations into sparse, interpretable features. However, existing SAE architectures primarily recover flat feature dictionaries and are less suited for explicit multi-level concept organization. In this paper, we introduce a cascaded sparse autoencoder architecture, dubbed SAE++, for learning hierarchical visual concepts in MLLMs. Rather than nesting or stacking SAE sparse activation codes, SAE++ trains a second-level SAE directly on the decoder weights of the first-level SAE, treating learned low-level feature directions as inputs for higher-level abstraction. This design enables SAE++ to learn "concepts of concepts" while avoiding drawbacks from the shared-prefix coupling of nesting, Matryoshka-style hierarchies and the bottlenecks of naively stacked SAEs. Experiments across Qwen3-VL, Gemma-3, and LLaVA on multiple visual
datasets show that SAE++ improves interpretability in terms of hierarchical concept coherence over state-of-the-art SAE baselines. Results on concept steering further demonstrate that the learned concept groups support effective group-level interventions in MLLM outputs. Code is available at https://github.com/Wang-ML-Lab/sae-plus-plus.
Read the full abstract ▾
All authors Yusong Zhao, Hengyi Wang, Tanuja Ganu, Akshay Nambi, Hao Wang
Artificial Intelligence · Cross-listed from cs.AR
Embedded Arena: Iterative Optimization via Hardware Feedback
Zhihan Zhang, Alexander Le Metzger, Jiuyang Lyu et al. (+10 more)
As AI moves into the physical world, from
robots to environmental sensors to medical wearables, it increasingly relies on local inference to meet latency, communication, or
privacy constraints. Optimizing models for heterogeneous edge hardware such as microcontrollers (MCUs) requires simultaneously satisfying hard physical constraints on
memory, power, and temperature while preserving accuracy, a multidimensional optimization that is today performed manually by experts. We ask whether an LLM
agent can autonomously navigate this complex, multi-turn pipeline guided by real hardware feedback, and introduce a hardware-in-the-loop
agent arena in which the
agent iteratively refines both model and firmware by compiling, flashing, and measuring on real hardware, enabling closed-loop optimization. Frontier models, including Claude Opus 4.7 and Gemini 3.1 Pro, fail entirely without hardware feedback (0% deployment success), whereas our hardware-in-the-loop formulation achieves the first successful deployment within three iterations and can surpass human expert results within seven. This agentic co-optimization achieves 250x compression for
vision models with <3.3% accuracy loss and 400x for audio with <6% Feature Error Rate loss, enabling battery-free operation on a commercial MCU via
solar harvesting. We demonstrate practical impact in two real-world systems: an elk-detection camera trap (96.7% accuracy) and a phonetic-transcription wearable (8.37% FER) for child development research.
Read the full abstract ▾
All authors Zhihan Zhang, Alexander Le Metzger, Jiuyang Lyu, Chun-Cheng Chang, Jiayi Shao, Yujia Liu, Emmanuel Azuh Mensah, Edward Wang, Kurtis Heimerl, Gregory D. Abowd, Shwetak Patel, Natasha Jaques, Vikram Iyer
Artificial Intelligence · Cross-listed from cs.LG
LLM-Powered Virtual Population for Demand Simulation and Pricing
Chengpiao Huang and Kaizheng Wang
We develop an LLM-powered virtual population model that simulates demand for pricing decisions, in settings where products are described by rich unstructured information, such as text descriptions and images, and where decision makers need not only mean-demand predictions but also uncertainty estimates for counterfactual prices. Our model represents exposed customers as draws from a finite mixture of customer personas. For each persona, product, and candidate price, an LLM elicits a persona-level purchase probability using both structured persona information and unstructured product information. These probabilities are aggregated through calibrated mixture weights to form a predictive distribution of aggregate demand. The resulting simulator can evaluate counterfactual prices under various pricing objectives, including expected revenue and risk-aware criteria such as conditional value at risk. We test the framework on an online H&M fashion
dataset with product descriptions and images. The calibrated LLM-based simulator achieves the best overall predictive performance among the models considered, and supports sample-efficient pricing decisions. Our framework provides a practical way to use LLMs as demand simulators for products with limited historical demand data but rich product information. By producing a full predictive demand distribution rather than only a point forecast, it enables managers to compare candidate prices, quantify demand uncertainty, and choose prices that target either average-case revenue or risk-aware objectives.
Read the full abstract ▾
Artificial Intelligence
PAL-Bench: Evidence-Grounded Profile Reconstruction from Longitudinal Personal Albums
Qiwei Yan, Zhiqiang Yuan, Zexi Jia et al. (+4 more)
Longitudinal personal albums are weak-schema
multimodal databases: noisy perceptual records whose key facts require joins across faces, text, timestamps, locations, and repeated events. Existing visual,
video, document, and lifelog
benchmarks test sub-problems, but not album-scale profile reconstruction with social identity binding and evidence citation. Benchmarking this task is difficult because the ground truth needed for evaluation–owner profiles, social graphs, face-name maps, and evidence provenance–is private state that real albums cannot safely release. We introduce PAL-Bench, a controlled
benchmark for evidence-grounded reconstruction under a public-record contract. Its Evidence Compiler builds latent private worlds, programs target-level evidence paths, renders album pixels, re-measures them through perception pipelines, and exports audited public/private views.
Agents receive only perception-derived public records; targets, identifier maps, and evidence paths remain hidden. PAL-Bench contains 50 synthetic users, 36,659 public photo records, and 2,799 targets over owner facts, identities, and relations. A
privacy-preserving audit with 10 participants confirms that PAL-Bench evidence structures match real private albums, though equivalent releases remain
privacy-prohibitive. Across seven systems and two compute-matched diagnostics, a seven-metric protocol reveals a gap between plausible profile summarization and faithful social reconstruction: systems recover some owner facts but struggle with recurring identities and evidence citation. PAL-TRACE, a reference framework that freezes identity bindings before owner-fact mining, performs best but leaves hard identity resolution far from solved. PAL-Bench provides a testbed for perceptual entity resolution,
multimodal data integration, temporal evidence aggregation, and provenance-aware structured prediction.
Read the full abstract ▾
All authors Qiwei Yan, Zhiqiang Yuan, Zexi Jia, Nanxing Hu, Kailin Lyu, Jie Zhou, Jinchao Zhang
Artificial Intelligence
TimeVista: Exploring and Exploiting Vision-Language Models as Judges for Time Series Forecasting
Zhi Chen, Yuxuan Wang, Jialong Wu et al. (+5 more)
High-quality time series
forecasting is pivotal for real-world decision-making. However, traditional point-wise metrics often fail to reveal complex temporal patterns and align poorly with human intuitive preferences. While the ”LLM-as-a-Judge” paradigm has revolutionized text evaluation by providing flexible, human-aligned judgment, its application to time series remains largely unexplored. In this paper, we leverage
Vision-
Language Models (VLMs) as judges for time series
forecasting, harnessing their ability to comprehend time series plots grounded in textual information. Specifically, we propose a novel framework integrating micro- and macro-level judgments informed by contextual information to evaluate time series
forecasting. To this end, we introduce TimeVista, a comprehensive VLM-as-a-Judge
benchmark comprising 5563 time series samples paired with detailed evaluation rubrics. Extensive meta-evaluations demonstrate that VLMs are highly reliable judges, achieving significantly higher consistency with human preferences than conventional metrics. Building upon our
benchmark, we comprehensively assess recent Time Series Foundation Models (TSFMs) under the VLM-as-a-Judge paradigm. Our results demonstrate that VLMs serve as robust and interpretable judges, providing a comprehensive, human-aligned standard for evaluating time series models.
Read the full abstract ▾
All authors Zhi Chen, Yuxuan Wang, Jialong Wu, Yong Liu, Haoran Zhang, Xingjian Su, Jianmin Wang, Mingsheng Long
Artificial Intelligence
AI Pluralism and the Worlds It Misses
Rashid Mushkani
AI pluralism is often framed as a problem of representing diverse values, preferences, users, or outputs. This paper argues that this framing is incomplete because AI systems also impose ontologies: they define what counts as an entity, relation, feature, harm, benefit, and valid form of evidence. We define ontological flattening as the conversion of situated, contested, and historically specific meanings into a restricted technical category, proxy, aggregation rule, or
benchmark target that is treated as neutral and difficult to contest. The paper develops a bounded conceptual and qualitative synthesis across value pluralism, pluralistic
alignment, participatory and democratic AI, procedural justice, science and technology studies, accountability research, aggregate themes from 11 expert interviews, and three urban AI companion cases. The cases illustrate how pluralistic methods can improve or structure model behavior while still compressing categories, proxies, aggregation rules, and revision rights before affected actors have procedural standing. We introduce Pluralistic Lifecycle Governance (PLG) as a preliminary qualitative audit scaffold for documenting ontological openness, epistemic inclusion, procedural authority, evaluation pluralism, and lifecycle accountability. PLG is not presented as a validated scoring instrument; it is a framework for making the evidence and governance conditions of pluralistic AI explicit.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.LG
A comparative and critical study of EEGNet for fNIRS-driven cognitive load classification
Mehshan Ahmed Khan, Houshyar Asadi, Li Zhang et al. (+8 more)
Accurately classifying cognitive load from functional near-infrared spectroscopy (fNIRS) signals remains a significant challenge due to temporal variability, inter-subject differences, and sensitivity to preprocessing choices. This study provides a comprehensive evaluation of EEGNet for fNIRS-based cognitive load classification by systematically examining the effects of temporal
segmentation strategies (overlapping vs. non-overlapping), window lengths (10s, 20s, 30s), feature extraction methods (Analysis of Variance (ANOVA), Principal Component Analysis (PCA), Fast Independent Component Analysis (FastICA)), learning rate configurations (fixed and adaptive), and evaluation protocols (random split vs. subject-independent (SI)). Results from random-split experiments show that overlapping
segmentation, combined with smaller fixed learning rates (0.01-0.001), yields the highest accuracies, due to temporal redundancy and dense sampling of hemodynamic transitions. However, SI evaluation reveals a substantial drop in accuracy, demonstrating limited generalization to unseen participants. Under SI evaluation, non-overlapping
segmentation outperformed overlapping windows, with the best accuracy of 56.11% achieved using PCA features with a 20-second window and a 0.1 learning rate. These findings indicate that eliminating temporal redundancy helps the model learn more robust and generalizable representations of cognitive load across individuals. Although adaptive learning rate strategy improved training stability, it did not surpass the performance of optimally selected fixed learning rates. The study highlights the critical role of
segmentation strategy and learning rate selection in improving model generalization and identifies methodological considerations essential for developing reliable, real-time, and SI cognitive load classification systems using fNIRS.
Read the full abstract ▾
All authors Mehshan Ahmed Khan, Houshyar Asadi, Li Zhang, Mohammad reza Chalak Qazani, Ghazal Bargshady, Stefanos gkikas, Christian arzate, Sam Oladazimi, Zoran Najdovsk, Lei Wei, Chee Peng Lim
Artificial Intelligence · Cross-listed from cs.CV
A Comprehensive Survey of Medical Image Segmentation: Challenges, Benchmarks, and Beyond
Pengyu Zhu, Xiaojing Zhang, Kunbo Zhang et al. (+2 more)
Medical image
segmentation plays a critical role in clinical diagnostics, treatment planning, disease monitoring, and neurological disorder identification. This article presents a comprehensive review of its systematic development, covering widely used public
datasets, representative methods built on the U-Net,
Transformer, and SAM architectures, and key evaluation metrics with their differences, followed by an analysis of major challenges from multiple perspectives. Unlike surveys that focus on a single model family or a specific clinical application, this review organizes U-Net-,
Transformer-, and SAM-based methods within a unified analytical framework, with a particular focus on their effectiveness in improving
segmentation accuracy and efficiency. This work aims to guide future research and support clinical translation of medical image
segmentation, with all related resources publicly available in our GitHub repository: https://github.com/andrew-pengyu/Awsome_MedSeg/tree/main.
Read the full abstract ▾
All authors Pengyu Zhu, Xiaojing Zhang, Kunbo Zhang, Chunyan Zhang, Zhenyu Wang
Artificial Intelligence
The Quality-Utility Paradox: Why High-Reward Data Impairs Small Model Mathematical Reasoning
Haolong Qian, Xianliang Yang, Yinuo ma et al. (+6 more)
Knowledge
distillation from powerful
reasoning models is widely used to improve Small
Language Models (SLMs) on mathematical
reasoning, often assuming that traces with higher reward model scores provide more useful supervision. We identify a counterintuitive Quality-Utility Paradox in mathematical
reasoning distillation. Data refined or synthesized by a stronger Oracle obtains higher perceived quality according to reward models, yet consistently underperforms traces generated by the SLM itself and selected through rejection sampling across Qwen2.5, LLaMA-3, and DeepSeek families. Our analysis shows that Oracle refinement couples logical repair with distributional drift away from the SLM's native
reasoning distribution. This drift increases the learner's adaptation cost and can outweigh the benefit of improved
reasoning logic. To test this mechanism, we introduce Style-Aligned Refinement, which preserves the native trajectory of the SLM while retaining logical repair from the Oracle. This intervention lowers adaptation cost and restores downstream utility. These findings suggest that effective mathematical
reasoning distillation should jointly optimize perceived solution quality and learner-data compatibility, rather than relying solely on reward-model scores. The
datasets and code are available at https://github.com/Dracoqhl/Quality-Utility-Paradox.
Read the full abstract ▾
All authors Haolong Qian, Xianliang Yang, Yinuo ma, Lirong Che, Feng Lu, Ye Guo, Lei Song, Jiang Bian, Chun Yuan
Artificial Intelligence
Teaching agentic AI to generalize expert diagnostic reasoning in rare diseases
Minh-Ha Nguyen, Erica Gray, Bryce A. Schuler et al. (+16 more)
Rare disease diagnosis depends on expert
reasoning that is scarce and difficult to transfer. Large
language models rank the correct disease first in only 35.4% of
benchmark cases and often rely on learned phenotype-disease associations rather than reusable diagnostic
reasoning strategies. We developed liteOdyssey through Policy Iteration with Human Feedback, a process in which model failures and expert corrections are iteratively consolidated into a clinician-gated, natural-
language policy executed by a
language model. Across 1,243 public
benchmark cases spanning 722 rare diseases, liteOdyssey ranked the correct disease first in 59.3% of cases versus 26.5% without the policy, with comparable gains in cases involving diseases excluded from policy development. The same policy transferred across model families and sizes without retraining. Adaptation of the policy to the Undiagnosed Diseases Network (UDN) improved diagnostic accuracy among 515 UDN patients, with gains confirmed by blinded physician adjudication. These results show that expert
reasoning can be externalized into an inspectable and revisable natural-
language policy that generalizes across rare diseases, transfers across model backbones, and adapts to a real-world patient cohort.
Read the full abstract ▾
All authors Minh-Ha Nguyen, Erica Gray, Bryce A. Schuler, Kevin W. Byram, Chih-Ting Yang, Fan Ma, Hua Xu, Wu-Chen Su, Chao Yan, Wei-Qi Wei, Adam Wright, Lisa Bastarache, Josh F. Peterson, Lingyao Li, Siyuan Ma, Undiagnosed Diseases Network, Rizwan Hamid, Thomas A. Cassini, Cathy Shyr
Artificial Intelligence
VibeThinker-3B: Exploring the Frontier of Verifiable Reasoning in Small Language Models
Sen Xu, Shixi Liu, Wei Wang et al. (+6 more)
This technical report introduces VibeThinker-3B, a compact dense model with 3B parameters developed to investigate how far verifiable
reasoning can be pushed within a strictly small-model regime. Building upon the Spectrum-to-Signal post-training paradigm, we systematically enhance the model through an optimized pipeline that includes curriculum-based supervised fine-tuning, multi-domain reinforcement learning, and offline self-
distillation. Experimental evaluations demonstrate that VibeThinker-3B achieves frontier-level performance on highly demanding verifiable tasks. Specifically, it attains a score of 94.3 on AIME26 (improving to 97.1 with claim-level test-time scaling), an 80.2 Pass@1 on LiveCodeBench v6, and exhibits strong out-of-distribution generalization with a 96.1% acceptance rate on recent unseen LeetCode contests. This effectively places it in the performance band of first-tier
reasoning systems, matching or exceeding flagship models that are orders of magnitude larger, such as DeepSeek V3.2, GLM-5, and Gemini 3 Pro. Furthermore, a score of 93.4 on IFEval confirms that this extreme
reasoning enhancement does not compromise strict instruction controllability. Extending our previous 1.5B work, these findings motivate the Parametric Compression-Coverage Hypothesis, which views verifiable
reasoning as compressible into compact
reasoning cores, while open-domain knowledge and general-purpose competence require broad parameter coverage over facts, concepts, and long-tail scenarios. This perspective suggests that compact models are not merely deployment-efficient substitutes, but a complementary path toward frontier-level performance in parameter-dense capability regimes.
Read the full abstract ▾
All authors Sen Xu, Shixi Liu, Wei Wang, Jixin Min, Yingwei Dai, Zhibin Yin, Yirong Chen, Xin Zhou, Junlin Zhang
Artificial Intelligence · Cross-listed from cs.CL
XAI-Grounded Explanation Generation for Speech Deepfake Detection with Training-Free Multimodal Large Language Models
Yupei Li, Qiyang Sun, Xiaoliang Wu et al. (+3 more)
Speech deepfake detection (SDD) systems require trustworthy explanations for reliable decision-making. Existing explanation ways mainly fall into two categories. Traditional explainable AI (XAI), such as gradient-based attribution, produces low-level attribution signals tightly coupled with model decisions, and harder to be understood by human than natural
language explanations. Meanwhile, large
language model (LLM)-based explanation generation often produces generic and ungrounded descriptions due to the lack of heuristic evidence and task-specific supervision, stemming from limited grounded explanation
datasets for SDD. We therefore propose a training-free explanation framework that integrates XAI evidence with
multimodal LLMs to generate grounded and specific explanations. Using the PartialSpoof
dataset, we construct a grounded explanation
dataset and show that methods with XAI increase inside accuracy by over 45%, verified through human evaluation and faithfulness checks.
Read the full abstract ▾
All authors Yupei Li, Qiyang Sun, Xiaoliang Wu, Chenxi Wang, Berrak Sisman, Björn W. Schuller
Artificial Intelligence · Cross-listed from cond-mat.mtrl-sci
InvDesMobility: a reliability-gated first-principles feedback framework for closed-loop materials discovery
Wen-Kao Li, Ze-Feng Gao, Peng-Jie Guo et al. (+2 more)
Inverse materials design starts from target functionality and searches for structures that can realize it. Its value in closed-loop discovery depends not only on prediction performance, but also on whether expensive first-principles results are independently validated, provenance-recorded, and admitted as feedback only when evidence is sufficient. This is especially important for composite properties such as carrier mobility, where a final scalar value hides intermediate quantities, fit quality, convergence history, and workflow assumptions. Here we present InvDesMobility, a reliability-gated first-principles feedback framework that integrates multi-
agent automated DFT, evidence stratification, generative structure proposal, acquisition ranking, and auditable release. Using 516 2DMatPedia-derived candidates, the workflow produced 280 QC-passed materials and 573 retained carrier-direction seed channels after channel-level reliability gating. These records were split into two feedback objects: relaxed structures updated the generative model, while retained mobility channels trained the acquisition model and set validation priority. Over multiple iterations, InvDesMobility screened 2.4 x 10^6 structures, submitted 102 candidates for DFT validation, and retained 86 reliability-gated generated channels across 41 formulas. Overall, the main contribution is not a fixed list of high-mobility materials, but a transferable feedback contract that makes closed-loop inverse design both useful and auditable when learning from expensive calculated properties. All source data, retained feedback records, and workflows are available at https://github.com/DreamLufei/invDesMobility, with an accompanying evidence website at https://dreamlufei.github.io/invDesMobility/.
Read the full abstract ▾
All authors Wen-Kao Li, Ze-Feng Gao, Peng-Jie Guo, Wei Ji, Zhong-Yi Lu