arXiv subjects

Hang Xu

Publications and source records attributed to Hang Xu.

At least 91 records · Page 5Linked to original sources

arXiv2025

MaskFocus: Focusing Policy Optimization on Critical Steps for Masked Image Generation

Reinforcement learning (RL) has demonstrated significant potential for post-training language models and autoregressive visual generative models, but adapting RL to masked generative models remains challenging. The core factor is that policy optimization requires accounting for the probability likelihood of each step due to its multi-step and iterative refinement process. This reliance on entire sampling trajectories introduces high computational cost, whereas natively optimizing random steps often yields suboptimal results. In this paper, we present MaskFocus, a novel RL framework that achieves effective policy optimization for masked generative models by focusing on critical steps. Specifically, we determine the step-level information gain by measuring the similarity between the intermediate images at each sampling step and the final generated image. Crucially, we leverage this to identify the most critical and valuable steps and execute focused policy optimization on them. Furthermore, we design a dynamic routing sampling mechanism based on entropy to encourage the model to explore more valuable masking strategies for samples with low entropy. Extensive experiments on multiple Text-to-Image benchmarks validate the effectiveness of our method.

cs.CV↗

arXiv2025

MCTS-EP: Empowering Embodied Planning with Online Preference Optimization

This paper introduces MCTS-EP, an online learning framework that combines large language models (LLM) with Monte Carlo Tree Search (MCTS) for training embodied agents. MCTS-EP integrates three key components: MCTS-guided exploration for preference data collection, efficient multi-modal reasoning mechanism, and iterative training pipeline based on preference optimization. We theoretically prove that MCTS-EP achieves better performance bounds than conventional on-policy algorithms when the loss function is strongly convex, and demonstrate that it can be formulated as a search-enhanced variant of GAIL. MCTS-EP achieves state-of-the-art performace across serval benchmarks. In ALFWorld, it achieves 92% and 87% success rates for textual and visual tasks. In WebShop, it reaches an average reward of 0.81. MTCS-EP also reduces average interaction steps from from 18.7/19.5 to 10.2/9.9 steps in visual ALFWorld.Code available at: https://github.com/xuhang-2/Embodied-Agent-Planning

cs.AI↗

arXiv2025

Highly Efficient Test-Time Scaling for T2I Diffusion Models with Text Embedding Perturbation

Test-time scaling (TTS) aims to achieve better results by increasing random sampling and evaluating samples based on rules and metrics. However, in text-to-image(T2I) diffusion models, most related works focus on search strategies and reward models, yet the impact of the stochastic characteristic of noise in T2I diffusion models on the method's performance remains unexplored. In this work, we analyze the effects of randomness in T2I diffusion models and explore a new format of randomness for TTS: text embedding perturbation, which couples with existing randomness like SDE-injected noise to enhance generative diversity and quality. We start with a frequency-domain analysis of these formats of randomness and their impact on generation, and find that these two randomness exhibit complementary behavior in the frequency domain: spatial noise favors low-frequency components (early steps), while text embedding perturbation enhances high-frequency details (later steps), thereby compensating for the potential limitations of spatial noise randomness in high-frequency manipulation. Concurrently, text embedding demonstrates varying levels of tolerance to perturbation across different dimensions of the generation process. Specifically, our method consists of two key designs: (1) Introducing step-based text embedding perturbation, combining frequency-guided noise schedules with spatial noise perturbation. (2) Adapting the perturbation intensity selectively based on their frequency-specific contributions to generation and tolerance to perturbation. Our approach can be seamlessly integrated into existing TTS methods and demonstrates significant improvements on multiple benchmarks with almost no additional computation. Code is available at \href{https://github.com/xuhang07/TEP-Diffusion}{https://github.com/xuhang07/TEP-Diffusion}.

cs.CV↗

arXiv2025

FR-TTS: Test-Time Scaling for NTP-based Image Generation with Effective Filling-based Reward Signal

Test-time scaling (TTS) has become a prevalent technique in image generation, significantly boosting output quality by expanding the number of parallel samples and filtering them using pre-trained reward models. However, applying this powerful methodology to the next-token prediction (NTP) paradigm remains challenging. The primary obstacle is the low correlation between the reward of an image decoded from an intermediate token sequence and the reward of the fully generated image. Consequently, these incomplete intermediate representations prove to be poor indicators for guiding the pruning direction, a limitation that stems from their inherent incompleteness in scale or semantic content. To effectively address this critical issue, we introduce the Filling-Based Reward (FR). This novel design estimates the approximate future trajectory of an intermediate sample by finding and applying a reasonable filling scheme to complete the sequence. Both the correlation coefficient between rewards of intermediate samples and final samples, as well as multiple intrinsic signals like token confidence, indicate that the FR provides an excellent and reliable metric for accurately evaluating the quality of intermediate samples. Building upon this foundation, we propose FR-TTS, a sophisticated scaling strategy. FR-TTS efficiently searches for good filling schemes and incorporates a diversity reward with a dynamic weighting schedule to achieve a balanced and comprehensive evaluation of intermediate samples. We experimentally validate the superiority of FR-TTS over multiple established benchmarks and various reward models. Code is available at \href{https://github.com/xuhang07/FR-TTS}{https://github.com/xuhang07/FR-TTS}.

cs.CV↗

arXiv2025

Percept-WAM: Perception-Enhanced World-Awareness-Action Model for Robust End-to-End Autonomous Driving

Autonomous driving heavily relies on accurate and robust spatial perception. Many failures arise from inaccuracies and instability, especially in long-tail scenarios and complex interactions. However, current vision-language models are weak at spatial grounding and understanding, and VLA systems built on them therefore show limited perception and localization ability. To address these challenges, we introduce Percept-WAM, a perception-enhanced World-Awareness-Action Model that is the first to implicitly integrate 2D/3D scene understanding abilities within a single vision-language model (VLM). Instead of relying on QA-style spatial reasoning, Percept-WAM unifies 2D/3D perception tasks into World-PV and World-BEV tokens, which encode both spatial coordinates and confidence. We propose a grid-conditioned prediction mechanism for dense object perception, incorporating IoU-aware scoring and parallel autoregressive decoding, improving stability in long-tail, far-range, and small-object scenarios. Additionally, Percept-WAM leverages pretrained VLM parameters to retain general intelligence (e.g., logical reasoning) and can output perception results and trajectory control outputs directly. Experiments show that Percept-WAM matches or surpasses classical detectors and segmenters on downstream perception benchmarks, achieving 51.7/58.9 mAP on COCO 2D detection and nuScenes BEV 3D detection. When integrated with trajectory decoders, it further improves planning performance on nuScenes and NAVSIM, e.g., surpassing DiffusionDrive by 2.1 in PMDS on NAVSIM. Qualitative results further highlight its strong open-vocabulary and long-tail generalization.

cs.CV↗

arXiv2025

Learning to Restore Heisenberg Limit in Noisy Quantum Sensing via Quantum Digital Twin

Quantum sensors leverage nonclassical resources to achieve sensing precision at the Heisenberg limit, surpassing the standard quantum limit attainable through classical strategies. However, a critical issue is that the environmental noise induces rapid decoherence, fundamentally limiting the realizability of the Heisenberg limit. In this Letter, we propose a quantum digital twin protocol to overcome this issue. The protocol first establishes observable-constrained state reconstruction to infer random errors in the decoherence process, and then utilizes reinforcement learning to derive adaptive compensatory control strategies. Demonstrated across discrete, continuous variable and multi-qubit circuit systems, our approach bypasses quantum state tomography's exponential overhead and discovers optimal control schemes to restore the Heisenberg limit. Unlike quantum error correction or mitigation schemes requiring precise noise characterization and ancillary qubits, our autonomous protocol achieves noise-resilient sensing through environment-adaptive control sequencing. This work establishes quantum digital twin as a generic methodology for quantum control, proposing a noise-immune paradigm for next-generation quantum sensors compatible with NISQ-era experimental constraints.

quant-ph↗

arXiv2025

4D-VLA: Spatiotemporal Vision-Language-Action Pretraining with Cross-Scene Calibration

Leveraging diverse robotic data for pretraining remains a critical challenge. Existing methods typically model the dataset's action distribution using simple observations as inputs. However, these inputs are often incomplete, resulting in a dispersed conditional action distribution-an issue we refer to as coordinate system chaos and state chaos. This inconsistency significantly hampers pretraining efficiency. To address this, we propose 4D-VLA, a novel approach that effectively integrates 4D information into the input to mitigate these sources of chaos. Our model introduces depth and temporal information into visual features with sequential RGB-D inputs, aligning the coordinate systems of the robot and the scene. This alignment endows the model with strong spatiotemporal reasoning capabilities while minimizing training overhead. Additionally, we introduce memory bank sampling, a frame sampling strategy designed to extract informative frames from historical images, further improving effectiveness and efficiency. Experimental results demonstrate that our pretraining method and architectural components substantially enhance model performance. In both simulated and real-world experiments, our model achieves a significant increase in success rate over OpenVLA. To further assess spatial perception and generalization to novel views, we introduce MV-Bench, a multi-view simulation benchmark. Our model consistently outperforms existing methods, demonstrating stronger spatial understanding and adaptability.

cs.CV↗

arXiv2025

Deep (Predictive) Discounted Counterfactual Regret Minimization

Counterfactual regret minimization (CFR) is a family of algorithms for effectively solving imperfect-information games. To enhance CFR's applicability in large games, researchers use neural networks to approximate its behavior. However, existing methods are mainly based on vanilla CFR and struggle to effectively integrate more advanced CFR variants. In this work, we propose an efficient model-free neural CFR algorithm, overcoming the limitations of existing methods in approximating advanced CFR variants. At each iteration, it collects variance-reduced sampled advantages based on a value network, fits cumulative advantages by bootstrapping, and applies discounting and clipping operations to simulate the update mechanisms of advanced CFR variants. Experimental results show that, compared with model-free neural algorithms, it exhibits faster convergence in typical imperfect-information games and demonstrates stronger adversarial performance in a large poker game.

cs.LG↗

arXiv2025

Unseen from Seen: Rewriting Observation-Instruction Using Foundation Models for Augmenting Vision-Language Navigation

Data scarcity is a long-standing challenge in the Vision-Language Navigation (VLN) field, which extremely hinders the generalization of agents to unseen environments. Previous works primarily rely on additional simulator data or web-collected images/videos to improve the generalization. However, the simulator environments still face limited diversity, and the web-collected data often requires extensive labor to remove the noise. In this paper, we propose a Rewriting-driven AugMentation (RAM) paradigm for VLN, which directly creates the unseen observation-instruction pairs via rewriting human-annotated training data. Benefiting from our rewriting mechanism, new observation-instruction pairs can be obtained in both simulator-free and labor-saving manners to promote generalization. Specifically, we first introduce Object-Enriched Observation Rewriting, where we combine Vision-Language Models (VLMs) and Large Language Models (LLMs) to derive rewritten object-enriched scene descriptions, enabling observation synthesis with diverse objects and spatial layouts via Text-to-Image Generation Models (T2IMs). Then, we propose Observation-Contrast Instruction Rewriting, which generates observation-aligned rewritten instructions by requiring LLMs to reason the difference between original and new observations. We further develop a mixing-then-focusing training strategy with a random observation cropping scheme, effectively enhancing data distribution diversity while suppressing augmentation data noise during training. Experiments on both the discrete environments (R2R, REVERIE, and R4R datasets) and continuous environments (R2R-CE dataset) show the superior performance and impressive generalization ability of our method. Code is available at https://github.com/SaDil13/VLN-RAM.

cs.CV↗

arXiv2025

Bergman kernels over polarized Kähler manifolds, Bergman logarithmic flatness, and a question of Lu-Tian

Let $M$ be a complete Kähler manifold, and let $(L, h) \to M$ be a positive line bundle inducing a Kähler metric $g$ on $M$. We study two Bergman kernels in this setting: the Bergman kernel of the disk bundle of the dual line bundle $(L^*, h^*)$, and the Bergman kernel of the line bundle $(L^k, h^k)$, $k\geq 1$, twisted by the canonical line bundle of $(M, g)$. We first prove a localization result for the former Bergman kernel. Then we establish a necessary and sufficient condition for this Bergman kernel to have no logarithmic singularity, expressed in terms of the Tian-Yau-Zelditch-Catlin type expansion of the latter Bergman kernel. This result, in particular, answers a question posed by Lu and Tian. As an application, we show that if $(M, g)$ is compact and locally homogeneous, then the circle bundle of $(L^*, h^*)$ is necessarily Bergman logarithmically flat.

math.CV↗

arXiv2025

VTimeCoT: Thinking by Drawing for Video Temporal Grounding and Reasoning

In recent years, video question answering based on multimodal large language models (MLLM) has garnered considerable attention, due to the benefits from the substantial advancements in LLMs. However, these models have a notable deficiency in the domains of video temporal grounding and reasoning, posing challenges to the development of effective real-world video understanding systems. Inspired by how humans use video players to interact with the progress bar for video comprehension, we introduce VTimeCoT, a simple yet effective training-free framework, designed for high-performance video grounding and reasoning. The proposed framework incorporates two novel visual tools of the progress bar: a plug-and-play progress bar integration tool and a high-efficiency highlighting tool. In addition, to address the limitations of conventional text-based chain-of-thought (CoT) approaches, we introduce a visuotemporal CoT process that integrates cross-modality reasoning across both video and text. Our approach demonstrates significant performance improvements on both Qwen2VL-7B and GPT4o baselines in tasks of video temporal grounding and reasoning-based question answering. Finally, we showcase that the proposed framework achieves a compositional and interpretable reasoning process. Project page: https://vtimecot.github.io

cs.CV↗

arXiv2025

Towards Proprioception-Aware Embodied Planning for Dual-Arm Humanoid Robots

In recent years, Multimodal Large Language Models (MLLMs) have demonstrated the ability to serve as high-level planners, enabling robots to follow complex human instructions. However, their effectiveness, especially in long-horizon tasks involving dual-arm humanoid robots, remains limited. This limitation arises from two main challenges: (i) the absence of simulation platforms that systematically support task evaluation and data collection for humanoid robots, and (ii) the insufficient embodiment awareness of current MLLMs, which hinders reasoning about dual-arm selection logic and body positions during planning. To address these issues, we present DualTHOR, a new dual-arm humanoid simulator, with continuous transition and a contingency mechanism. Building on this platform, we propose Proprio-MLLM, a model that enhances embodiment awareness by incorporating proprioceptive information with motion-based position embedding and a cross-spatial encoder. Experiments show that, while existing MLLMs struggle in this environment, Proprio-MLLM achieves an average improvement of 19.75% in planning performance. Our work provides both an essential simulation platform and an effective model to advance embodied intelligence in humanoid robotics. The code is available at https://anonymous.4open.science/r/DualTHOR-5F3B.

cs.RO↗

arXiv2025

DualTHOR: A Dual-Arm Humanoid Simulation Platform for Contingency-Aware Planning

Developing embodied agents capable of performing complex interactive tasks in real-world scenarios remains a fundamental challenge in embodied AI. Although recent advances in simulation platforms have greatly enhanced task diversity to train embodied Vision Language Models (VLMs), most platforms rely on simplified robot morphologies and bypass the stochastic nature of low-level execution, which limits their transferability to real-world robots. To address these issues, we present a physics-based simulation platform DualTHOR for complex dual-arm humanoid robots, built upon an extended version of AI2-THOR. Our simulator includes real-world robot assets, a task suite for dual-arm collaboration, and inverse kinematics solvers for humanoid robots. We also introduce a contingency mechanism that incorporates potential failures through physics-based low-level execution, bridging the gap to real-world scenarios. Our simulator enables a more comprehensive evaluation of the robustness and generalization of VLMs in household environments. Extensive evaluations reveal that current VLMs struggle with dual-arm coordination and exhibit limited robustness in realistic environments with contingencies, highlighting the importance of using our simulator to develop more capable VLMs for embodied tasks. The code is available at https://github.com/ds199895/DualTHOR.git.

cs.RO↗

arXiv2025

TransMamba: Fast Universal Architecture Adaption from Transformers to Mamba

Transformer-based architectures have become the backbone of both uni-modal and multi-modal foundation models, largely due to their scalability via attention mechanisms, resulting in a rich ecosystem of publicly available pre-trained models such as LLaVA, CLIP, and DeiT, etc. In parallel, emerging sub-quadratic architectures like Mamba offer promising efficiency gains by enabling global context modeling with linear complexity. However, training these architectures from scratch remains resource-intensive (e.g., in terms of data and time). Motivated by this challenge, we explore a cross-architecture knowledge transfer paradigm, termed TransMamba, that facilitates the reuse of Transformer pre-trained knowledge. We propose a two-stage framework to accelerate the training of Mamba-based models, ensuring their effectiveness across both uni-modal and multi-modal tasks. The first stage leverages pre-trained Transformer models to initialize critical components of the Mamba architecture. To bridge architectural and dimensional gaps, we develop a selective weight subcloning strategy and a layered initialization scheme that prioritizes the early $n$ layers. Building on this initialization, the second stage introduces an adaptive multi-directional knowledge distillation method. This mechanism employs layer-wise adaptive scaling factors to align Mamba representations with their Transformer counterparts, while accommodating the scanning order variations inherent to multi-modal Mamba architectures. Despite operating with a reduced training dataset and a more compact model architecture, TransMamba consistently outperforms baseline approaches across diverse mamba-based backbones (e.g., PlainMamba, Vmamba, ViM and VideoMamba) and downstream tasks (e.g., image classification, visual question answering, text-video retrieval and multimodal reasoning). All code and implementation details will be released.

cs.CV↗

arXiv2025

SeePhys: Does Seeing Help Thinking? -- Benchmarking Vision-Based Physics Reasoning

We present SeePhys, a large-scale multimodal benchmark for LLM reasoning grounded in physics questions ranging from middle school to PhD qualifying exams. The benchmark covers 7 fundamental domains spanning the physics discipline, incorporating 21 categories of highly heterogeneous diagrams. In contrast to prior works where visual elements mainly serve auxiliary purposes, our benchmark features a substantial proportion of vision-essential problems (75%) that mandate visual information extraction for correct solutions. Through extensive evaluation, we observe that even the most advanced visual reasoning models (e.g., Gemini-2.5-pro and o4-mini) achieve sub-60% accuracy on our benchmark. These results reveal fundamental challenges in current large language models' visual understanding capabilities, particularly in: (i) establishing rigorous coupling between diagram interpretation and physics reasoning, and (ii) overcoming their persistent reliance on textual cues as cognitive shortcuts.

cs.AI↗

arXiv2025

UniPose: Unified Cross-modality Pose Prior Propagation towards RGB-D data for Weakly Supervised 3D Human Pose Estimation

In this paper, we present UniPose, a unified cross-modality pose prior propagation method for weakly supervised 3D human pose estimation (HPE) using unannotated single-view RGB-D sequences (RGB, depth, and point cloud data). UniPose transfers 2D HPE annotations from large-scale RGB datasets (e.g., MS COCO) to the 3D domain via self-supervised learning on easily acquired RGB-D sequences, eliminating the need for labor-intensive 3D keypoint annotations. This approach bridges the gap between 2D and 3D domains without suffering from issues related to multi-view camera calibration or synthetic-to-real data shifts. During training, UniPose leverages off-the-shelf 2D pose estimations as weak supervision for point cloud networks, incorporating spatial-temporal constraints like body symmetry and joint motion. The 2D-to-3D back-projection loss and cross-modality interaction further enhance this process. By treating the point cloud network's 3D HPE results as pseudo ground truth, our anchor-to-joint prediction method performs 3D lifting on RGB and depth networks, making it more robust against inaccuracies in 2D HPE results compared to state-of-the-art methods. Experiments on CMU Panoptic and ITOP datasets show that UniPose achieves comparable performance to fully supervised methods. Incorporating large-scale unlabeled data (e.g., NTU RGB+D 60) enhances its performance under challenging conditions, demonstrating its potential for practical applications. Our proposed 3D lifting method also achieves state-of-the-art results.

cs.CV↗

arXiv2025

GLaVE-Cap: Global-Local Aligned Video Captioning with Vision Expert Integration

Video detailed captioning aims to generate comprehensive video descriptions to facilitate video understanding. Recently, most efforts in the video detailed captioning community have been made towards a local-to-global paradigm, which first generates local captions from video clips and then summarizes them into a global caption. However, we find this paradigm leads to less detailed and contextual-inconsistent captions, which can be attributed to (1) no mechanism to ensure fine-grained captions, and (2) weak interaction between local and global captions. To remedy the above two issues, we propose GLaVE-Cap, a Global-Local aligned framework with Vision Expert integration for Captioning, which consists of two core modules: TrackFusion enables comprehensive local caption generation, by leveraging vision experts to acquire cross-frame visual prompts, coupled with a dual-stream structure; while CaptionBridge establishes a local-global interaction, by using global context to guide local captioning, and adaptively summarizing local captions into a coherent global caption. Besides, we construct GLaVE-Bench, a comprehensive video captioning benchmark featuring 5X more queries per video than existing benchmarks, covering diverse visual dimensions to facilitate reliable evaluation. We further provide a training dataset GLaVE-1.2M containing 16K high-quality fine-grained video captions and 1.2M related question-answer pairs. Extensive experiments on four benchmarks show that our GLaVE-Cap achieves state-of-the-art performance. Besides, the ablation studies and student model analyses further validate the effectiveness of the proposed modules and the contribution of GLaVE-1.2M to the video understanding community. The source code, model weights, benchmark, and dataset will be open-sourced.

cs.CV↗

arXiv2025

G-LLaVA: Solving Geometric Problem with Multi-Modal Large Language Model

Large language models (LLMs) have shown remarkable proficiency in human-level reasoning and generation capabilities, which encourages extensive research on their application in mathematical problem solving. However, current work has been largely focused on text-based mathematical problems, with limited investigation in problems involving geometric information. Addressing this gap, we aim to enable LLMs to solve geometric problems by understanding image input. We first analyze the limitations of current Multimodal Large Language Models (MLLMs) in this area: they struggle to accurately comprehending basic geometric elements and their relationships. To overcome these challenges, we take advantage of the unique characteristics of geometric problems (such as unique geometric logical form, and geometric scalability) and the capacity of the textual LLMs to build an enriched multimodal geometry dataset based on existing data. The augmented dataset, Geo170K, contains more than 170K geometric image-caption and question-answer pairs. Utilizing our constructed Geo170K dataset, we develop G-LLaVA, which demonstrates exceptional performance in solving geometric problems, significantly outperforming GPT-4-V on the MathVista benchmark with only 7B parameters.

cs.CL↗