Large language models (LLMs) and time-series language models (TSLMs) are increasingly applied to time-series question answering (TSQA). Unlike text-only QA, TSQA requires models to ground answers in temporal signals whose patterns may occur at different scales, specific time locations, or across separated intervals. However, existing benchmarks are typically organized by task types or high-level reasoning categories, making it difficult to diagnose the underlying signal-level capabilities driving model performance. We introduce TS-Skill, a controlled benchmark for evaluating three composable analytical skills in TSQA: temporal scale selection (SK1), temporal localization (SK2), and cross-interval integration (SK3). TS-Skill provides timestamp-aware questions, broad domain coverage, and human-validated QA quality. To construct the benchmark at scale, we develop SKEvol, a skill-guided agentic framework that combines domain-aware time-series seed generation, skill-controlled question generation, metadata- and code-assisted answer construction, multi-phase signal-grounded verification, and human-in-the-loop curation. Experiments on ten state-of-the-art LLMs and TSLMs reveal substantial and uneven capability gaps across SK1-SK3. In particular, SK3 remains consistently challenging for non-agent models, whereas tool-augmented agents show a selective advantage on standalone SK3. These findings demonstrate that skill-level evaluation can uncover temporal reasoning failures that are obscured by aggregate TSQA scores.

Reference-free image-to-text evaluators are now standard for scoring image-caption alignment, yet it is unclear whether they respect semantic invariances. We present an invariance probe on five popular evaluators (CLIPScore, PAC-S, UMIC, FLEUR, and a deterministic LLM judge) under semantics-preserving perturbations along three axes -- spatial (flips, context-preserving repositioning, light rotations), object (scale, category), and socio-linguistic framing (cultural/economic adjectives with neutral and length-matched controls). Across curated slices of three detection datasets and three caption evaluation suites, we find consistent non-semantic sensitivities, where benign spatial edits and simple phrasing changes shift scores by $\approx$6--9\% on average, and for systems separated by just 0.7\%, these shifts can cause ranking flips in up to $\sim$37\% of cases, particularly under spatial changes. A small human study also supports this finding and confirms that annotators generally judge perturbed pairs as equally correct, so these shifts reflect metric behavior rather than semantic change. We further propose invariance-calibrated scoring, a post-hoc adjustment that roughly halves median absolute sensitivity while retaining correlation with learned caption evaluators.

Most reported gains on agentic-LLM clinical benchmarks are often attributed to prompt engineering, yet our results suggest that larger improvements can come from architectural and engine-level design. We present MDIA, a Multi-agent Diagnostic Intelligence Agent implemented as a 7-node specialty-routed clinical reasoning graph, on the full HealthBench Professional benchmark (n = 525), on a non-fine-tuned LLM. MDIA achieves 0.6272 under OpenAI's GPT-5.4-2026-03-05, which is +3.72 pp above the performance of OpenAI's ChatGPT for Clinicians. The experimental work shows that performance lift is attributable to system architecture: specialty routing, multi-turn context preservation, drug-state safety gating, site-filtered search, length-aware synthesis, and engine-level reliability. These findings support the view that agentic clinical benchmark performance is shaped both by the underlying foundation model and the orchestration architecture. Nevertheless, we also noticed notable differences when using other models as a grader; in particular, when using Gemini 2.5 Pro, MDIA scored 0.6585, which suggests that the choice of grader is a source of variability. Robust evaluation of LLMs would therefore require assessment across several independent grader models.

Diffusion large language models promise faster generation by refining many token positions in parallel, but this parallelism introduces a hidden control problem: which proposed tokens should be transferred into the partially decoded sequence at each step? We refer to this decision as token commitment. Existing frozen-generator decoders largely rely on hand-designed confidence rules or block-specific acceptance filters. We argue that token commitment can instead be learned as a reusable trace-state policy. We introduce TraceLock, a lightweight plug-in controller that instantiates this policy for a frozen diffusion language model. Since oracle commitment times are unavailable, TraceLock derives self-supervision from future stability: at decoding step t, a proposed token for position i is labeled stable if it matches the final token at position i after the full decoding trace completes. The controller scores variable-length trace states and decides which active token proposals should be committed to the partially decoded sequence. Once trained for a given frozen backbone, the controller can be deployed across local-window widths, generation lengths, and step budgets without retraining or per-setting calibration. Experiments on question answering, mathematical reasoning, and code generation show that TraceLock improves the quality-step tradeoff over heuristic and learned baselines, with particularly stable behavior under cross-setting deployment. Diagnostic analyses show that its decisions are not reducible to scalar confidence, suggesting that frozen diffusion language models expose a learnable space of commitment trajectories beyond confidence-based decoding. Code is available at https://github.com/BobSun98/TraceLock.

Large language models still struggle with contest-level programming, while many agentic remedies rely on massive inference-time sampling or expensive multi-stage post-training. We study when execution feedback reliably helps an LLM CP solver and which mechanisms govern the gains. We model feedback-driven solving as a calibrated stopped process and identify three quantities: false-admission risk, program-level evidence against bad programs, and the active-state success hazard. Under held-out trace calibration and selection from a pre-declared finite controller manifest, the resulting structural certificate lower-bounds the clean success probability before false admission. We instantiate mechanisms targeting these quantities as Dual-Granularity Verification, Test Augmentation, and Experience-Driven Self-Evolving, yielding CP-Agent. Without updating any parameters, CP-Agent raises Pass@1 from 25.8\% to 48.5\% on LiveCodeBench Pro and improves Refine@5 by 11.0\% on ICPC-Eval. Across three LLM backbones, CP-Agent lies on the cost--accuracy efficiency frontier, and ablations show that each component primarily affects its corresponding certificate quantity.

Detecting objects reliably under extreme low-light conditions is an open problem in computer vision, with practical urgency in applications ranging from nighttime surveillance to search-and-rescue robotics. Conventional RGB cameras degrade sharply at low photon flux, while event cameras which record asynchronous per-pixel brightness changes at microsecond resolution and high dynamic range provide complementary structural cues that are largely illumination-invariant. We present AdaFuse-Det, a dual-stream framework that fuses CLAHE-enhanced RGB frames with voxelized event tensors through an Adaptive Cross-Modal Fusion (ACMF) module grounded in minimum-variance linear estimation theory. We formally show that the learned attention map asymptotically recovers the Gauss-Markov optimal fusion weights, and establish event conservation and temporal resolution bounds for the voxelization stage. On the LLE-VOS benchmark, AdaFuse-Det achieves a Recall of $65.54\%$, Precision of $53.85\%$, and F1-Score of $59.12\%$ under severe illumination degradation, outperforming single-modality detectors in recall by a margin that reflects the theoretically predicted illumination-adaptation behavior.

The motion planning problem for robotic manipulation can be addressed through classical or deep learning approaches. Existing methods face significant challenges in generalizing to diverse settings. In this study, we present a method with high generalization capability that generates collision-free trajectories using diffusion models where the denoising process is guided by the gradient of the total collision cost. We are also presenting a dynamic approach for choosing start step of the gradient guidance. Experimental results demonstrate that guiding the diffusion model dynamically with the sum of collision costs offers more robust performance by overcoming the generalization issues faced by competing methods. The proposed model demonstrates its effectiveness by achieving the highest performance on diverse test settings in M$π$nets\ dataset among the compared methods.

Text-to-Image (T2I) models have made significant strides in visual realism and semantic consistency, yet they often perpetuate and amplify societal biases. Existing evaluation methods typically address only single-dimensional biases, lacking perspectives to uncover model biases at social-related deeper semantic levels. We introduce HoloFair, a comprehensive benchmark framework for multidimensional demographic bias analysis. Built upon our large-scale fairness-oriented dataset and the SpaFreq (Spatial-Frequency) attribute classifier, this framework proposes the Multi-attribute, Group-wise Bias Index (MGBI) metric, designed to assess both intrinsic diversity and conditional biases. Beyond evaluation, we further introduce Fair-GRPO, a reinforcement-learning-based debiasing method that alters the distribution of generative models through a designed multi-objective reward function. E.g., experiments on the SD3.5-Medium model demonstrate that Fair-GRPO significantly improves multidimensional fairness while maintaining high image quality. We also analyze potential reward hacking phenomena and provide corresponding mitigation strategies. Code and dataset are available at https://github.com/1059684669/HoloFair

As large language models (LLMs) are increasingly integrated into emotionally sensitive domains, the structural integrity of their emotional intelligence (EI) becomes a critical frontier for safety and alignment. Current benchmarks often conflate superficial politeness with deep affective reasoning, failing to distinguish between perceptual accuracy and interactive efficacy. Here, we introduce FACET (Functional Affective Competence and Empathy Test), a psychometrically grounded framework comprising 480 expert-crafted items. Unlike previous metrics, FACET is theoretically anchored in the Mayer-Salovey-Caruso four-branch ability model, operationalizing EI through perception, facilitation, understanding, and management of emotions. Through an evaluation of nine frontier models (including GPT-5, Claude-Sonnet-4), we demonstrate that emotional intelligence is not a monolithic capability but is fragmented across cognitive and interactive dimensions. While frontier models demonstrate robust proficiency in objective emotion recognition and social reasoning, this does not consistently translate to interactive success. We categorize these discrepancies into three distinct performance profiles: cognitive-dominant, interactive-dominant, and context-dependent. These typologies indicate that emotional skills do not scale uniformly with general intelligence or model size; rather, they are shaped by specific alignment paradigms. Notably, we identify hidden emotion recognition as a universal performance bottleneck across all architectures. Our results suggest that current RLHF processes may optimize for "stochastic empathy", a statistical mimicry of emotional syntax, at the expense of integrated affective reasoning. These findings challenge the assumption of linear emotional scaling and provide a rigorous roadmap for developing socially aware agents capable of genuine clinical resonance.

Multimodal Attributed Graph Learning (MAGL) integrates intrinsic node attributes with structural topology via graph aggregation. However, as pretrained encoders evolve into Large Foundation Models (LFMs), the landscape of MAGL fundamentally shifts: under high-confidence LFM priors, mandatory aggregation introduces topological noise that overwhelms discriminative signals, triggering a counter-intuitive performance inversion where sophisticated MAGL architectures underperform simple topology-agnostic MLPs. Through systematic empirical and theoretical analysis, we identify that this inversion stems from a fundamental aggregation dilemma characterized by two concurrent pathologies: (1) Representational Pathology (SNR Degradation) - mandatory aggregation dilutes robust intrinsic features with topological noise, causing the noise penalty to outweigh its collaborative benefit; and (2) Optimization Pathology (Gradient Starvation) - topological aggregation attenuates gradient flow, while a shared task loss causes dominant modalities to prematurely suppress weaker ones. To resolve this dilemma, we propose SUPRA (Shared-Unique Prior-Retaining Architecture), a decoupled dual-pathway paradigm. SUPRA processes modality-specific features through topology-agnostic MLPs while capturing structural synergy via a lightweight shared GNN, with auxiliary deep supervision counteracting gradient starvation. Extensive evaluations demonstrate that SUPRA achieves state-of-the-art performance while requiring 3.5x lower peak GPU memory and up to 4.4x faster training time than Multimodal Graph Transformers.

Gradient-boosted trees achieve strong performance on tabular data, yet often leave a long tail of poorly predicted instances. We introduce a Trajectory-based Difficulty Score (TDS), an instance-level difficulty estimator for boosted ensembles derived from per-tree cumulative prediction trajectories. For each instance, we compute interpretable trajectory descriptors (e.g., variance, oscillation peaks, sign switches, and tail stability) and train a lightweight regression model to predict held-out loss. An empirical CDF calibrates the resulting signal into a score in $[0,1]$ that supports ranking hard cases. Across diverse tabular benchmarks and ensemble sizes, TDS exhibits strong rank correlation with error and outperforms established instance-hardness and uncertainty baselines on classification, while remaining competitive on regression. We then show how a single difficulty signal improves multiple data mining workflows: difficulty-driven active learning for label-efficient training, difficulty-thresholded selective prediction for improved risk-coverage trade-offs, and TDS-stratified (Mondrian) conformal prediction for more uniform conditional coverage. Finally, clustering high-TDS instances using SHAP attributions reveals coherent failure modes characterized by compact feature-value ranges, supporting error analysis and targeted data acquisition.

Cross-subject brain-to-visual decoding remains a core challenge in brain-computer interfaces due to severe inter-individual variability that induces systematic subject-specific functional misalignment. To address this issue, we propose MindAdapter, a parameter-efficient few-shot calibration framework for pretrained brain-to-visual decoding models. MindAdapter adopts a decoupled linear-residual cascade alignment paradigm by freezing a pretrained explicit brain functional alignment backbone (coarse) and introducing a lightweight nonlinear residual adapter (fine), thereby disentangling global cross-subject correspondence from subject-specific residual corrections for fine-grained spatial and semantic calibration. To further preserve global representational stability, we design a topology-anchored dual-stream manifold constraint, where a small set of shared stimuli serves as topological pins with voxel-level paired supervision, while a semantic stream enforces consistency through a frozen vision-language decoder on unpaired brain data. Together, MindAdapter efficiently injects subject-specific corrections while maintaining the global representational geometry learned during pretraining. Experiments on the Natural Scenes Dataset (NSD) demonstrate that MindAdapter substantially improves cross-subject visual reconstruction and retrieval accuracy using only a few shared stimuli, offering a practical and data-efficient solution for personalized brain-to-visual decoding.

Translating text embedded in Web images is crucial for improving content accessibility and cross-lingual information retrieval, particularly within social media and e-commerce domains. Although Large Vision-Language Models (LVLMs) have advanced multimodal understanding, applying them to Web image translation remains challenging due to the visual representation gap: standard encoders often prioritize high-level semantics over the fine-grained visual details required for recognizing diverse character morphologies. To address this challenge, we propose VaaWIT, an end-to-end framework that adapts Large Language Models for multilingual Web image translation. The framework introduces two key technical contributions: (1) a Dual-Stream Attention Module (DSAM), which facilitates bidirectional interaction between multilingual semantic features and detailed visual representations, thereby synthesizing unified features robust to textual variations; and (2) a Visual-Aware Adapter (VAA), a parameter-efficient fine-tuning strategy that dynamically injects these fused visual cues into the frozen LLM backbone. This design enables the model to align the visual context with linguistic reasoning effectively while minimizing computational costs. Extensive experiments on eight tasks on three public benchmarks demonstrate that VaaWIT significantly outperforms state-of-the-art (SOTA) open-source baselines and achieves competitive performance against proprietary models. These results validate the efficacy of integrating fine-grained visual perception into LLMs for complex Web content analysis.

Instruction-based video editing requires transforming a source video according to a natural-language instruction while preserving irrelevant content and remaining temporally coherent. We argue that existing Diffusion Transformer (DiT) editors struggle with this task for two structural reasons. First, conditioning signals are fed undifferentiated into all transformer blocks, forcing a single token stream to encode both global editing intent and fine-grained visual evidence. Second, the cross-attention patterns that govern the edit are supervised only indirectly through pixel-level reconstruction, leaving the model's internal reasoning process under-constrained. To address both limitations, we propose RVEDiT, an implicit Reasoning Video Editing DiT framework built around two complementary components. The first, Granularity-Routed Token Conditioning, introduces learnable editing tokens distilled from a multimodal LLM and routes them to shallow blocks, while reserving native visual and textual tokens for deeper blocks, thereby inducing a coarse-to-fine editing process inside the backbone. The second, Reference-Anchored Attention Alignment, employs a parameter-sharing reference branch during training and maximizes the mutual information between the attention features of the editing and reference branches, regularizing the model's internal reasoning without incurring any additional inference cost. Experiments on standard instruction-based video editing benchmarks show that RVEDiT consistently outperforms state-of-the-art baselines, with particularly strong gains on localized and compositional edits.

Mean cross-entropy is the standard validation metric for language models, but it can fail to track model quality during training. We examine this in two common scenarios. First, in Qwen2.5-1.5B SFT on synthetic fact-learning, we find that mean CE rises substantially after the initial learning phase while held-out fact-recall accuracy remains near its peak. Second, we find that in top-K distillation on TinyStories, decreasing K improves median CE while worsening mean CE; the Top-5 student attains the highest LLM-judge score and crosses below its teacher on median CE, despite having the worst mean CE. In both cases, median CE correlates much more closely with task performance than does mean CE. Analyzing how bulk and tail percentile CE move during training reveals that training reshapes the empirical per-token CE distribution. In top-K distillation, smaller K yields a distribution with more mass at both extremes, decreasing the median and increasing the mean. In Qwen SFT, the bulk saturates quickly while the tail extends in the latter half of training. In both, the task-evaluation metric appears more sensitive to the bulk than to the tail. Practically, we recommend reporting a small set of percentile CE summaries alongside the mean, and using concordance among them as a tool to keep track of distribution reshaping, as well as a low-cost diagnostic for when mean and median CE disagree on model selection.

LLMs have achieved remarkable success in complex reasoning tasks, yet current evaluation approaches predominantly rely on final-answer correctness, offering limited insight into the underlying reasoning processes that produce those answers. To address this gap, this study proposes a unified multi-dimensional framework for measuring reasoning quality in LLMs from a behavioral perspective, operationalizing six theoretically grounded dimensions: Correctness (CQ), Consistency (CS), Robustness (RS), Logical Coherence (LS), Efficiency (ES), and Stability (SS). Extensive experiments on seven LLMs across 975 items from four benchmarks demonstrate that the framework reveals behaviors invisible to accuracy-only metrics. Notably, logical coherence is orthogonal to correctness (r = -0.172, ns), confirming that correct answers can arise from incoherent reasoning, while Claude-Haiku-4.5 achieves the highest multi-dimensional score (Q_bal = 0.778). Furthermore, the framework exposes critical ranking inversions: DeepSeek-V3 ranks second under accuracy-priority but fifth under legal/compliance weighting, a reversal that single-metric evaluation cannot detect. Discriminant validity confirms 11/15 dimension pairs are independent (|r| < 0.50), providing psychometric support for treating each dimension as a distinct signal. The dimensional profiles produced by the framework directly support three classes of deployment decision: identifying models whose reasoning traces would fail accountability audits despite correct final answers (LS--CQ orthogonality); preventing ranking errors caused by accuracy-only benchmarking; and ensuring that no single metric silently substitutes for the six independent signals the framework captures.

LLM-based agents are increasingly deployed for complex tasks requiring planning, tool use, and interaction with external services. Their reliance on untrusted external content exposes them to indirect prompt injection (IPI), in which adversarial instructions embedded in retrieved data hijack agent behavior. Existing attacks rely on static payloads that cannot adapt to agent-specific defenses; even recent adaptive methods lack structured feedback to guide optimization. We introduce \oursys, a feedback-guided iterative framework that closes the loop between injection, diagnosis, and refinement: a rule-based diagnoser produces structured outcome labels with behavioral descriptions, and an LLM-based optimizer refines payloads conditioned on the full optimization history. A synthesis step generates new disguise seeds from failure patterns, enabling the strategy space to self-evolve. On AgentDojo and InjectAgent, \oursys substantially outperforms static baselines and existing adaptive methods across four victim models. Extension experiments on Claude Code, a production-grade coding agent with layered defenses, show that optimized payloads achieve full success on 5 of 9 targets; even those that resist full exploitation exhibit measurable improvement from iterative refinement. We further present a mechanistic analysis of IPI, identifying an attention-mediated threshold mechanism in mid-to-late layers; three causal interventions validate this finding and point to concrete defense directions.

This work introduces the Wavelet-Laplace Neural Operator (WLNO), a novel neural operator that fuses Haar wavelet multi-scale spatial decomposition with the Laplace-domain pole-residue formulation of the Laplace Neural Operator (LNO). While LNO captures transient and steady-state dynamics through learnable system poles and residues, it lacks an explicit mechanism for extracting spatially localized multi-scale features inherent in complex PDE solutions. WLNO addresses this by augmenting the LNO core with a parallel single-level Haar discrete wavelet transform (DWT) branch that decomposes the lifted feature map into four frequency subbands: approximation (LL), horizontal detail (LH), vertical detail (HL), and diagonal detail (HH) and applies independent learned $1\times1$ convolutions to each subband before reconstruction via the inverse DWT. The two branches are fused through a learnable sigmoid-gated weight $α_\mathrm{wav}$, initialized to give a small initial contribution to the wavelet branch, allowing the model to adaptively balance Laplace-domain dynamics against spatial multi-scale features throughout training. WLNO is evaluated against LNO on five benchmark PDE problems using identical hyperparameters, training data, and evaluation protocols: the diffusion equation, the Burgers equation, the reaction-diffusion system, Darcy flow, and the two-dimensional Navier-Stokes equation. WLNO consistently outperforms LNO on all five problems, with the most pronounced improvement on problems with strong spatial multi-scale structure, such as the Burgers equation with sharp shock fronts and the Navier-Stokes equation with coherent vortical structures, while remaining consistent across smoother and elliptic problems. These results demonstrate that wavelet-based multi-scale spatial decomposition is a principled and effective complement to Laplace-domain operator learning.

Major LLM platforms deploy models in an inference-only configuration: the model serves requests but never updates per-user weights. Users must repeatedly re-teach preferences, corrections, and project context, and context-based workarounds consume context-window space and degrade under cascading compaction. We evaluate an alternative: nightly consolidation of interaction knowledge into model weights via reflection, synthesis, and Low-Rank Adaptation (LoRA) fine-tuning on a single consumer GPU. Across ten realistic software development conversations (n = 10, 1,146 test questions across three memory types), three cycles of cascading compaction retain 36.8 +/- 3.0% of knowledge (between an 11.8% no-context floor and a 90.1% full-context ceiling), while consolidation retains 80.4 +/- 1.3% -- a 43.6 pp gain (paired t(9) = 14.8, p < 0.001) that more than doubles what compaction preserves, with the largest gains on procedural corrections (36.3% -> 74.6%) and episodic project facts (31.5% -> 78.2%). As a methodological aside, mean per-token validation cross-entropy is negatively correlated with LLM-judged accuracy (r = -0.51) while median per-token validation cross-entropy tracks accuracy almost exactly (r = +0.99): under evaluators that tolerate surface-form variation, the mean is misleading and a heavy-tail-robust statistic is the faithful signal. Persistent personalization requires moving beyond inference-only deployment toward architectures that consolidate knowledge into weights.

Rapid advances in audio-video (AV) generation have enabled high-fidelity synthesis with synchronized sound, particularly for human-related scenarios involving speech and interactions. Yet evaluation for AV generation remains at an early stage, with only a few coarse-grained benchmarks for human-related scenarios and relying on limited preset evaluations with generic multimodal LLMs, leading to inaccurate assessments of model capabilities. To address these issues, we introduce AVBench, a fully automated benchmark tailored for human-centric AV generation. AVBench is built on two key designs for comprehensive and accurate evaluation: (i) Human-centric and fine-grained metrics. AVBench integrates ten evaluation dimensions designed for human-centered real-world scenarios, covering visual quality, audio quality, and multi-level consistency across modalities. These practical metrics capture human-related details that existing benchmarks often overlook. (ii) Specialized evaluators via preference learning. To address the lack of specialized training data, we construct large-scale supervision by transforming real-world videos into diverse training pairs with controlled perturbations. After fine-tuning on this high-quality dataset, the evaluators learn to reliably detect subtle cross-modal inconsistencies. Crucially, instead of producing discrete textual judgment, AVBench derives continuous evaluation scores from the model's prediction confidence on binary decisions. This probabilistic scoring mechanism enables a more reliable assessment than traditional VQA-style evaluation and aligns closely with human judgment. Taken together, AVBench offers automated evaluation for AV generation, demonstrates strong potential for data filtering, and serves as a differentiable reward signal for Reinforcement Learning from Human Feedback (RLHF).

Personalized dialogue requires more than recalling explicit user histories: systems also need to infer hidden user states that evolve through interaction and shape appropriate response strategies. Existing memory- and profile-based methods primarily reuse observable user information, offering limited support for modeling user-state dynamics or selecting actions based on how they shape future user states. We propose PUMA (Prospective User-state Modeling for Action selection), a framework grounded in the Free Energy Principle (FEP) that formulates personalization as decision-making under partial observability, centered on an explicit user state model that captures latent user states and their action-conditioned dynamics. At each turn, PUMA maintains a belief over the user's hidden state, refines the user state model for observation generation and action-conditioned state transition, and selects dialogue actions by minimizing expected free energy, balancing epistemic and pragmatic objectives under a unified criterion. This formulation shifts personalization from passive memory retrieval to model-based decision-making over user evolution. We instantiate PUMA on healthcare-oriented counseling and motivational interviewing benchmarks with latent state annotations for rigorous evaluation. Experiments show that PUMA improves long-horizon dialogue outcomes while maintaining strong response quality, and a cross-dataset study demonstrates more reliable user-state estimation and next-state prediction.

This paper presents reinforcement learning (RL) policies for dynamic quadrupedal locomotion in planetary exploration scenarios. Building on a taskoptimized quadruped with a 5-bar leg design, we develop RL policies for walking, vertical jumping, forward jumping, and in-flight attitude control, explicitly tailored to the reduced gravity on Mars. These policies jointly enable such robots to overcome obstacles larger than themselves through coordinated jumping and precise in-flight reorientation for safe landings. We demonstrate Sim2Real transfer of the attitude control policy on the Olympus quadruped through single-axis reorientation tests, while all locomotion policies are validated in simulation. A complete Mars exploration mission scenario demonstrates coordinated policy deployment across challenging terrain. Experimental results show 90° attitude reorientation in 2.6 seconds, with simulations demonstrating 3.1 meter vertical jumps and 3.9 meter forward jumps under Martian gravity conditions. - Supplementary video: https://www.youtube.com/watch?v=qlSJ3P87A4A

Recent work explores new opportunities at the intersection of vision-language-action models (VLAs) and geometric foundation models (GFMs) for 3D reconstruction, such as VGGT. While the resulting geometric VLAs often show improved performance, it remains unclear (i) if modern VLAs already have sufficient geometric understanding to start with, (ii) what is the best architecture to inject geometric understanding into a VLA, and (iii) what is the effect of other design choices that affect geometric VLAs. In this paper we provide a rigorous experimental analysis to shed light on these questions, for a specific choice of VLA (GR00T-N1.5) and GFM (VGGT). Our first contribution is to formalize prior work's intuition that current VLAs lack geometric understanding, by providing a rigorous analysis based on linear probing. The analysis quantifies, for the first time, the "geometric gap" between VLAs and GFMs. Our second contribution is to identify and compare different strategies to bridge GFMs with VLAs. We implement three different architectures, which differ in the way they inject geometry in the VLA, while keeping low-level implementation details as similar as possible, to ensure a fair comparison. Finally, we analyze the impact of non-architectural choices (e.g., training data, number of cameras, reconstruction quality) on the performance of the geometric VLAs.

With the widespread application of drones in recent years, object detection of aerial images has attracted increasing attention, especially open-vocabulary aerial detection which is not restricted to predefined categories. Due to the scarcity of drone's viewpoint images and their significant differences from natural images, it is difficult to achieve satisfying results by directly applying vanilla open-vocabulary detection methods designed for natural scenarios. Some studies propose to transfer knowledge from pre-trained models by using lightweight networks or generating pseudo labels, but they tend to rely on models trained on natural images, neglecting the potential of foundation models specifically tailored for remote sensing and aerial imagery. To address this limitation, we propose DisDop, a unified framework that systematically distills multi-level domain priors from remote sensing foundation models (e.g., RemoteCLIP and DINOv3) into a lightweight detector. Specifically, we first distill visual priors through a teacher fusion strategy that combines RemoteCLIP's cross-modal alignment capability with DINOv3's fine-grained local feature extraction ability, transferring their complementary strengths to the detector's backbone. Second, we distill textual priors embedded in RemoteCLIP's text encoder by explicitly modeling inter-category semantic relationships, while incorporating global contextual priors to enhance local feature representation for small objects. Through this multi-level prior distillation framework, our DisDop achieves new state-of-the-art performance on open-vocabulary aerial detection benchmarks. Extensive ablation analysis also demonstrates the rationality and effectiveness of our proposed modules.

Medical large language models hold promise for reducing healthcare disparities, yet Hindi remains severely underrepresented. While medical LLMs excel in high-resource languages, their performance degrades sharply in Hindi, particularly on Indian systems of medicine. We argue that robust cross-lingual medical transfer requires Hindi reasoning. To this end, we introduce HiMed, a Hindi reasoning medical corpus and benchmark suite covering both Western and Indian medicine. We further propose HiMed-8B, a Hindi-form medical reasoning LLM, through the design of decaying scaffolding reward. Extensive experiments demonstrate improvement in Hindi medical reasoning performance and reduction in the English--Hindi accuracy gap. Ablation studies validate the contribution of each training stage and reward component. All data and code are available on GitHub: https://github.com/FreedomIntelligence/HiMed.

Minority sampling aims to generate low-density instances on a data manifold and is of central importance in applications such as medical diagnosis, anomaly detection, and creative AI. Existing approaches, however, define minority samples relative to generative priors learned from training data, confining rarity to model-specific notions that may poorly reflect real-world semantics. In this work, we propose a world-centric perspective on minority sampling, which defines rarity with respect to real-world priors rather than generator-induced densities. To this end, we introduce JEPA guidance, a diffusion sampling framework guided by a Joint-Embedding Predictive Architecture (JEPA) -- a class of world models that encode broad, semantically rich representations. JEPA guidance steers diffusion trajectories toward low-density regions under the implicit density induced by the JEPA, thereby aligning generated minorities with real-world semantic rarity. To make JEPA guidance computationally practical, we develop principled approximation strategies accompanied by theoretical error bounds, significantly reducing the overhead of guidance computation. Extensive experiments across unconditional, class-conditional, and text-to-image generation demonstrate that JEPA guidance consistently improves the fidelity and semantic validity of minority samples, outperforming generator-centric baselines in capturing real-world notions of rarity. Code is available at https://github.com/soobin-um/jepa-guidance.

Recent progress of video diffusion models have enabled extensive simulation of the physical world. While simulation with hand object interaction has been less explored. We propose DexSIM, a dexterous simulation framework for simulating dexterous manipulation in real-time. While previous works utilizing video diffusion and 3D reconstruction focus on navigation, dexterous manipulation has been limited while it has extensive applications for creating interactive experiences with the simulated world and for generating synthetic data for robotics. Existing methods lack real-time interactivity and long-term spatial consistency and memory. We propose a 2-stage training framework for DexSIM. First we train a bi-directional video diffusion model by jointly embedding the hand action trajectory and video in a unified feature space. We utilize gaussian heatmap hand encoding for more accurate hand representation. Then we conduct a roll-out based autoregressive training with updated spatial cache as attention sink for spatial memory, which improves long-term consistency and 3D aware dexterous manipulation simulation. DexSIM outperforms the baseline on pixel and semantic similarity, motion fidelity, and hand projection accuracy. It also allows new applications such as hand motion transfer and runs at 15.24 FPS real-time interactivity.

Ultra-low-field (ULF) MRI offers portable and accessible neuroimaging but suffers from reduced signal-to-noise ratio and limited spatial resolution compared to high-field (HF) systems. Acquiring paired ULF-HF data for supervised enhancement is often difficult, particularly in resource-limited settings. We introduce ULF-Synth, a framework that combines: (i) acquisition-based synthesis of realistic ULF images from HF volumes to create large-scale paired training data, (ii) a spatial-frequency domain objective that prioritizes recovery of high-frequency anatomical detail. This formulation is architecture-agnostic, consistently improving structural similarity and perceptual fidelity across encoder-decoder, adversarial, and diffusion-based translation models. When trained exclusively on synthetic data, the resulting models generalize effectively to real 64mT ULF acquisitions, improving downstream multiclass brain segmentation and achieving higher radiologist preference and diagnostic acceptability in a blinded reader study. These findings demonstrate that synthetic paired supervision provides a practical and scalable pathway for enhancing ULF MRI without requiring real paired acquisitions. Code, Models and Dataset: https://github.com/toufiqmusah/ULF-Synth

In-context image generation models such as FLUX.2 take a text prompt and an optional reference image as visual conditioning for the output. Internally, all three inputs -- text, reference image, and the noise tokens -- are concatenated and processed through a single attention stream, where all tokens can attend to one another. This leaves open how reference information flows through the model to produce the output image. We show that an implicit cross-modal binding emerges between the text tokens and the reference image: the text tokens absorb visual reference content during the forward pass, and that absorbed content causally influences the generated output. We surface this binding with three causal interventions on FLUX.2: T2I Lens, which decodes intermediate text-token activations through a text-to-image path; Attention Knockout, which severs specific attention edges; and I2I-to-I2I Patching, which copies text token activations between editing runs. Across 2,875 editing tasks on various images, including SUN397 and DreamBench++ datasets and images collected online, we observe a consistent division of labor: properties of the reference image, like color, style, and scene setting, are first written into the text tokens, which carry them to the generated image; pixel-exact properties like a specific face or instance identity bypass the text tokens and flow directly from reference to image through image-to-image attention. We further localize the reference-text binding to the padding tokens of the text sequence. These results show that text tokens in a multimodal DiT are not just prompt holders, but a structured channel for reference image content. More broadly, they suggest that even in unified-attention multimodal generative models, token modality structures how conditioning information is represented and routed across the network.

A robot resolving ``put the cup on that one'' must fuse gesture, language, and scene geometry, yet 3D grounding benchmarks only partially capture this regime: descriptions are written post-hoc, gestures are templated, or pointing is staged for the camera. MM-Conv captures natural co-speech gesture from dyadic VR interaction alongside full-body motion capture and 3D scene graphs. We use it to evaluate pose-language fusion with a decoupled late-fusion architecture in which pose and text pathways share no learned parameters. The two choices together make category, pose, and text contributions easier to isolate through controlled ablations. Fusion with frozen MiniLM category embeddings exceeds pose alone and the best text-only pathway on every reference type, reaching 31.9% top-1. The learned scalar gate flips between opposing policies depending on whether the text pathway has category access. This is a reliability diagnostic: fusion-accuracy claims for semantic grounding systems are indistinguishable from category-representation artifacts unless pathways are architecturally decoupled.