CoVEBench: Can Video Editing Models Handle Complex Instructions?

While recent text-guided video editing models excel at elementary tasks (e.g., style transfer, object insertion), real-world user requests are highly compositional. A single prompt often demands multiple coupled edits, such as modifying subjects, actions, and camera views, while strictly preserving unrelated spatiotemporal content. Existing benchmarks, heavily constrained by isolated edits and coarse global metrics, fail to diagnose how models handle such complex workflows. To address this gap, we introduce CoVEBench, a compositional video editing benchmark comprising 416 curated source videos, 626 multi-point editing instructions, and 9,990 fine-grained checklist items. Covering diverse editing dimensions, CoVEBench evaluates models via MLLM-judged instruction compliance and video fidelity, alongside automated metrics for video quality. Extensive experiments reveal that compositional editing remains a profound challenge: current models frequently omit edits, violate preservation constraints, or introduce artifacts when handling multiple operations simultaneously. CoVEBench provides a challenging, diagnostic testbed to advance video editing toward realistic user workflows.

OmniCap-IF: Benchmarking and Improving Instruction Following Abilities for Omni-Video Captioning

While Omni-modal Large Language Models (OLLMs) have demonstrated impressive capabilities in jointly processing audio and visual streams, their ability to strictly adhere to complex, multi-faceted user instructions remains largely unexplored. Existing benchmarks primarily focus on holistic video understanding or text-only instruction following, failing to capture the intricate interplay between modalities and user constraints. To bridge this gap, we introduce OmniCap-IF, the first comprehensive benchmark specifically designed to evaluate instruction-following capabilities in omni-modal captioning. OmniCap-IF incorporates a systematic framework that assesses captions on two dimensions: format correctness and content correctness. Our benchmark encompasses 50 distinct constraint types across pure visual, pure audio, and audio-visual modalities, while integrating Temporal Grounding to assess spatio-temporal precision. Extensive evaluations of prominent models on 1,920 high-quality samples reveal significant performance disparities. Furthermore, our analysis uncovers a critical "format-content tradeoff", demonstrating that increasing formatting complexity directly degrades models' omni-modal reasoning abilities. Finally, to advance the field, we curate a 54K instruction-tuning dataset, OmniCap-IF-54K and present OmniCaptioner-IF, which achieves notable improvements in both complex instruction adherence and general omni-modal captioning performance.

Mind Your Steps: A General Learning Framework for Accurate Humanoid Foothold Tracking

Enabling humanoid robots to operate in complex, dynamic environments remains a critical challenge, fundamentally limited by the ability to navigate robustly, safely, and accurately. While reinforcement learning with velocity-commanded policies has achieved remarkable robustness in humanoid locomotion, this approach lacks explicit control of the foothold placement, leading to unsafe behavior, such as stepping onto human feet, or imprecise navigation, hindering the following manipulation task. Conversely, explicit foothold-tracking policies offer a promising alternative by directly being commanded with target foot poses. However, existing approaches are often limited by unrealistic state assumptions, compromising real-world deployment, or they are part of staged pipelines, making them tied to specific downstream tasks. In this work, we introduce a novel, lightweight framework for training general-purpose 3D foothold-tracking policies. By dynamically providing footstep support through a goal sampler, this method enables the learned policy to be agnostic to specific terrains. Our new target representation effectively mitigates challenges arising in the real world, such as noisy and inaccurate pose estimation and foot contact estimation. Designed for direct real-world transfer, our policy acts as a standalone low-level controller that can be seamlessly paired with various high-level foothold generators. We demonstrate the effectiveness of our framework through extensive experiments in simulation and in the real world. By coupling our policy with different upstream planners, we achieve natural and accurate locomotion in challenging settings, paving the way for loco-manipulation tasks in complex environments.

MMG2Skill: Can Agents Distill In-the-Wild Guides into Self-Evolving Skills?

Abundant procedural knowledge on the Web holds great potential for helping agents solve long-horizon tasks. However, such knowledge is often multimodal, heterogeneous, noisy, and implicitly assumes human executors, making it difficult to use directly as the skills required by agents. To bridge the gap between human-oriented guides and agent-executable skills, we formalize this problem as guide-to-skill learning: converting in-the-wild guides into executable skills and continuously improving them from trajectories observable to the agent. To evaluate the capability of existing agents on this task, we introduce MMG2Skill-Bench, the first benchmark designed for this problem. We further propose MMG2Skill, a closed-loop framework that compiles guides into editable skills, conditions a fixed vision-language model (VLM) agent on these skills during execution, and revises the skills from trajectory-level root-cause feedback without using benchmark scores. Across GUI control, open-ended gameplay, and strategic card play with six VLM backbones, MMG2Skill consistently outperforms vanilla baseline agents in every model-domain setting, achieving macro-average gains of +12.8 to +25.3 percentage points across backbones. Ablation studies show that directly prompting agents with raw guides can degrade performance, while both structured skill construction and trajectory-driven revision are necessary for the observed improvements. On success-inferable tasks, analyzer-based early stopping further prevents late-stage performance regressions and saves 25%-53% of attempts when the success signal is properly calibrated.

SmoCap: Unified Scale-Pose Canonicalization with Proxy-Mapped Trust-Region QP

Objective: Stage-wise workflows that separate model scaling and inverse kinematics can induce morphology-posture compensation, resulting in anatomically inconsistent yet numerically acceptable solutions, especially in weakly observed directions. We present SmoCap, a leakage-resistant canonicalization framework that estimates morphology and posture jointly in each local trust-region quadratic program (QP) within a sparse control subspace. Methods: SmoCap solves a constrained trust-region QP with analytical proxy-mapped pose and scale Jacobians. The low dimensional proxy map stabilizes weakly observed directions and drives coordinated structures. An optional pre-solve provides warm starts in difficult configurations. The framework is evaluated using cohort fluoroscopy knee motion, anthropometric ground truth, and extreme yoga sequences. Results: SmoCap achieved 2.9 degree knee flexion RMSE against fluoroscopy, and a pooled anthropometric endpoint error around 3%. In the leakage audit against segment wise scaling, SmoCap also reduced marker RMSE, FE error, and anthropometric endpoint error. Proxy coupling preserved expressive and coordinated spine motion with marginal fitting error increase (+0.14 mm, +0.6%) against baseline models in yoga ablation. Median marker RMSE was around 20 mm, and median runtime was 0.204-0.332 ms/frame, achieved with consistently 2-3 iterations. Conclusion: SmoCap provides an externally validated unified coupling-aware scale-pose framework, making externally consistent motion canonicalization practical at dataset scale.

DR$^{3}$-Eval: Towards Realistic and Reproducible Deep Research Evaluation

Deep Research Agents (DRAs) aim to solve complex, long-horizon research tasks involving planning, retrieval, multimodal understanding, and report generation, yet their evaluation remains challenging due to dynamic web environments and ambiguous task definitions. We propose DR$^{3}$-Eval, a realistic and reproducible benchmark for evaluating deep research agents on multimodal, multi-file report generation. DR$^{3}$-Eval is constructed from authentic user-provided materials and paired with a per-task static research sandbox corpus that simulates open-web complexity while remaining fully verifiable, containing supportive documents, distractors, and noise. Moreover, we introduce a multi-dimensional evaluation framework measuring Information Recall, Factual Accuracy, Citation Coverage, Instruction Following, and Depth Quality, and validate its alignment with human judgments. Experiments with our developed multi-agent system DR$^{3}$-Agent based on multiple state-of-the-art language models demonstrate that DR$^{3}$-Eval is highly challenging and reveals critical failure modes in retrieval robustness and hallucination control. Our code and data are publicly available.

A Wearable ECG Device for Differentiating Hypertrophic Cardiomyopathy from Acquired Left Ventricular Hypertrophy

Hypertrophic Cardiomyopathy (HCM) is a genetic heart disease affecting approximately 1 in 500 people and is the leading cause of sudden cardiac death in young athletes. Current diagnostic methods -- cardiovascular magnetic resonance (CMR), echocardiography, and genetic testing -- are limited by high costs, operator dependency, or insufficient accuracy, while standard electrocardiogram (ECG) analysis cannot reliably distinguish HCM from acquired left ventricular hypertrophy (LVH). This paper presents a wearable ECG device paired with a classification algorithm that differentiates HCM from acquired LVH using ECG signals alone. The portable device integrates a 3-lead electrode system, an AD8232 signal conditioning module, an Arduino Nano 33 BLE microcontroller, and a lithium polymer battery. The algorithm extracts two quantitative indices -- HCM Index~1 and HCM Index~2 -- from each heartbeat and classifies patients via dual statistical thresholds. Validation on 483 LVH patients (PhysioNet) and 29 HCM patients (digitized clinical records) yields 75.86\% sensitivity, 99.17\% specificity, and an F1-score of 80.00\%. Leave-one-out cross-validation confirms generalizability, with cross-validated sensitivity of 72.41\%, specificity of 98.96\%, and F1-score of 76.36\% (95\% confidence intervals reported). A digitization confound analysis demonstrates that the classification is driven by physiological cardiac features rather than data source artifacts. A simulated device acquisition chain analysis confirms that the wearable hardware's signal characteristics are compatible with the classification algorithm. The system offers a promising tool for affordable HCM screening in resource-limited settings.

VLM-in-the-Loop: A Plug-In Quality Assurance Module for ECG Digitization Pipelines

ECG digitization could unlock billions of archived clinical records, yet existing methods collapse on real-world images despite strong benchmark numbers. We introduce \textbf{VLM-in-the-Loop}, a plug-in quality assurance module that wraps any digitization backend with closed-loop VLM feedback via a standardized interface, requiring no modification to the underlying digitizer. The core mechanism is \textbf{tool grounding}: anchoring VLM assessment in quantitative evidence from domain-specific signal analysis tools. In a controlled ablation on 200 records with paired ground truth, tool grounding raises verdict consistency from 71\% to 89\% and doubles fidelity separation ($Δ$PCC 0.03 $\rightarrow$ 0.08), with the effect replicating across three VLMs (Claude Opus~4, GPT-4o, Gemini~2.5 Pro), confirming a pattern-level rather than model-specific gain. Deployed across four backends, the module improves every one: 29.4\% of borderline leads improved on our pipeline; 41.2\% of failed limb leads recovered on ECG-Digitiser; valid leads per image doubled on Open-ECG-Digitizer (2.5 $\rightarrow$ 5.8). On 428 real clinical HCM images, the integrated system reaches 98.0\% Excellent quality. Both the plug-in architecture and tool-grounding mechanism are domain-parametric, suggesting broader applicability wherever quality criteria are objectively measurable.

Gradient-Based Data Valuation Improves Curriculum Learning for Game-Theoretic Motion Planning

We demonstrate that gradient-based data valuation produces curriculum orderings that significantly outperform metadata-based heuristics for training game-theoretic motion planners. Specifically, we apply TracIn gradient-similarity scoring to GameFormer on the nuPlan benchmark and construct a curriculum that weights training scenarios by their estimated contribution to validation loss reduction. Across three random seeds, the TracIn-weighted curriculum achieves a mean planning ADE of $1.704\pm0.029$\,m, significantly outperforming the metadata-based interaction-difficulty curriculum ($1.822\pm0.014$\,m; paired $t$-test $p=0.021$, Cohen's $d_z=3.88$) while exhibiting lower variance than the uniform baseline ($1.772\pm0.134$\,m). Our analysis reveals that TracIn scores and scenario metadata are nearly orthogonal (Spearman $ρ=-0.014$), indicating that gradient-based valuation captures training dynamics invisible to hand-crafted features. We further show that gradient-based curriculum weighting succeeds where hard data selection fails: TracIn-curated 20\% subsets degrade performance by $2\times$, whereas full-data curriculum weighting with the same scores yields the best results. These findings establish gradient-based data valuation as a practical tool for improving sample efficiency in game-theoretic planning.

Behavioral Score Diffusion: Model-Free Trajectory Planning via Kernel-Based Score Estimation from Data

Diffusion-based trajectory optimization has emerged as a powerful planning paradigm, but existing methods require either learned score networks trained on large datasets or analytical dynamics models for score computation. We introduce \emph{Behavioral Score Diffusion} (BSD), a training-free and model-free trajectory planner that computes the diffusion score function directly from a library of trajectory data via kernel-weighted estimation. At each denoising step, BSD retrieves relevant trajectories using a triple-kernel weighting scheme -- diffusion proximity, state context, and goal relevance -- and computes a Nadaraya-Watson estimate of the denoised trajectory. The diffusion noise schedule naturally controls kernel bandwidths, creating a multi-scale nonparametric regression: broad averaging of global behavioral patterns at high noise, fine-grained local interpolation at low noise. This coarse-to-fine structure handles nonlinear dynamics without linearization or parametric assumptions. Safety is preserved by applying shielded rollout on kernel-estimated state trajectories, identical to existing model-based approaches. We evaluate BSD on four robotic systems of increasing complexity (3D--6D state spaces) in a parking scenario. BSD with fixed bandwidth achieves 98.5\% of the model-based baseline's average reward across systems while requiring no dynamics model, using only 1{,}000 pre-collected trajectories. BSD substantially outperforms nearest-neighbor retrieval (18--63\% improvement), confirming that the diffusion denoising mechanism is essential for effective data-driven planning.