We introduce VisualPRM, an advanced multimodal Process Reward Model (PRM) with 8B parameters, which improves the reasoning abilities of existing Multimodal Large Language Models (MLLMs) across different model scales and families with Best-of-N (BoN) evaluation strategies. Specifically, our model improves the reasoning performance of three types of MLLMs and four different model scales. Even when applied to the highly capable InternVL2.5-78B, it achieves a 5.9-point improvement across seven multimodal reasoning benchmarks. Experimental results show that our model exhibits superior performance compared to Outcome Reward Models and Self-Consistency during BoN evaluation. To facilitate the training of multimodal PRMs, we construct a multimodal process supervision dataset VisualPRM400K using an automated data pipeline. For the evaluation of multimodal PRMs, we propose VisualProcessBench, a benchmark with human-annotated step-wise correctness labels, to measure the abilities of PRMs to detect erroneous steps in multimodal reasoning tasks. We hope that our work can inspire more future research and contribute to the development of MLLMs. Our model, data, and benchmark are released in https://internvl.github.io/blog/2025-03-13-VisualPRM/.

Existing motion generation methods based on mocap data are often limited by data quality and coverage. In this work, we propose a framework that generates diverse, physically feasible full-body human reaching and grasping motions using only brief walking mocap data. Base on the observation that walking data captures valuable movement patterns transferable across tasks and, on the other hand, the advanced kinematic methods can generate diverse grasping poses, which can then be interpolated into motions to serve as task-specific guidance. Our approach incorporates an active data generation strategy to maximize the utility of the generated motions, along with a local feature alignment mechanism that transfers natural movement patterns from walking data to enhance both the success rate and naturalness of the synthesized motions. By combining the fidelity and stability of natural walking with the flexibility and generalizability of task-specific generated data, our method demonstrates strong performance and robust adaptability in diverse scenes and with unseen objects.

Current robot learning algorithms for acquiring novel skills often rely on demonstration datasets or environment interactions, resulting in high labor costs and potential safety risks. To address these challenges, this study proposes a skill-learning framework that enables robots to acquire novel skills from natural language instructions. The proposed pipeline leverages vision-language models to generate demonstration videos of novel skills, which are processed by an inverse dynamics model to extract actions from the unlabeled demonstrations. These actions are subsequently mapped to environmental contexts via imitation learning, enabling robots to learn new skills effectively. Experimental evaluations in the MetaWorld simulation environments demonstrate the pipeline's capability to generate high-fidelity and reliable demonstrations. Using the generated demonstrations, various skill learning algorithms achieve an accomplishment rate three times the original on novel tasks. These results highlight a novel approach to robot learning, offering a foundation for the intuitive and intelligent acquisition of novel robotic skills.

Physical adversarial attacks in driving scenarios can expose critical vulnerabilities in visual perception models. However, developing such attacks remains challenging due to diverse real-world backgrounds and the requirement for maintaining visual naturality. Building upon this challenge, we reformulate physical adversarial attacks as a one-shot patch-generation problem. Our approach generates adversarial patches through a deep generative model that considers the specific scene context, enabling direct physical deployment in matching environments. The primary challenge lies in simultaneously achieving two objectives: generating adversarial patches that effectively mislead object detection systems while determining contextually appropriate placement within the scene. We propose MAGIC (Mastering Physical Adversarial Generation In Context), a novel framework powered by multi-modal LLM agents to address these challenges. MAGIC automatically understands scene context and orchestrates adversarial patch generation through the synergistic interaction of language and vision capabilities. MAGIC orchestrates three specialized LLM agents: The adv-patch generation agent (GAgent) masters the creation of deceptive patches through strategic prompt engineering for text-to-image models. The adv-patch deployment agent (DAgent) ensures contextual coherence by determining optimal placement strategies based on scene understanding. The self-examination agent (EAgent) completes this trilogy by providing critical oversight and iterative refinement of both processes. We validate our method on both digital and physical level, \ie, nuImage and manually captured real scenes, where both statistical and visual results prove that our MAGIC is powerful and effectively for attacking wide-used object detection systems.

Large vision-language models (LVLMs) have shown remarkable capabilities in interpreting visual content. While existing works demonstrate these models' vulnerability to deliberately placed adversarial texts, such texts are often easily identifiable as anomalous. In this paper, we present the first approach to generate scene-coherent typographic adversarial attacks that mislead advanced LVLMs while maintaining visual naturalness through the capability of the LLM-based agent. Our approach addresses three critical questions: what adversarial text to generate, where to place it within the scene, and how to integrate it seamlessly. We propose a training-free, multi-modal LLM-driven scene-coherent typographic adversarial planning (SceneTAP) that employs a three-stage process: scene understanding, adversarial planning, and seamless integration. The SceneTAP utilizes chain-of-thought reasoning to comprehend the scene, formulate effective adversarial text, strategically plan its placement, and provide detailed instructions for natural integration within the image. This is followed by a scene-coherent TextDiffuser that executes the attack using a local diffusion mechanism. We extend our method to real-world scenarios by printing and placing generated patches in physical environments, demonstrating its practical implications. Extensive experiments show that our scene-coherent adversarial text successfully misleads state-of-the-art LVLMs, including ChatGPT-4o, even after capturing new images of physical setups. Our evaluations demonstrate a significant increase in attack success rates while maintaining visual naturalness and contextual appropriateness. This work highlights vulnerabilities in current vision-language models to sophisticated, scene-coherent adversarial attacks and provides insights into potential defense mechanisms.

Existing open-source multimodal large language models (MLLMs) generally follow a training process involving pre-training and supervised fine-tuning. However, these models suffer from distribution shifts, which limit their multimodal reasoning, particularly in the Chain-of-Thought (CoT) performance. To address this, we introduce a preference optimization (PO) process to enhance the multimodal reasoning capabilities of MLLMs. Specifically, (1) on the data side, we design an automated preference data construction pipeline to create MMPR, a high-quality, large-scale multimodal reasoning preference dataset. and (2) on the model side, we explore integrating PO with MLLMs, developing a simple yet effective method, termed Mixed Preference Optimization (MPO), which boosts multimodal CoT performance. Our approach demonstrates improved performance across multiple benchmarks, particularly in multimodal reasoning tasks. Notably, our model, InternVL2-8B-MPO, achieves an accuracy of 67.0 on MathVista, outperforming InternVL2-8B by 8.7 points and achieving performance comparable to the 10x larger InternVL2-76B. We hope this study could inspire further advancements in MLLMs. Code, data, and model shall be publicly released.

Abstract: Misinformation can seriously impact society, affecting anything from public opinion to institutional confidence and the political horizon of a state. Fake News (FN) proliferation on online websites and Online Social Networks (OSNs) has increased profusely. Various fact-checking websites include news in English and barely provide information about FN in regional languages. Thus the Urdu FN purveyors cannot be discerned using fact-checking portals. FND in regional and resourceconstrained languages lags due to the lack of limited-sized datasets and legitimate lexical resources. The previous datasets for Urdu FND are limited-sized, domain-restricted, publicly unavailable and not manually verified where the news is translated from English into Urdu. In this paper, we curate and contribute the first largest publicly available dataset for Urdu FND, "Ax-to-Grind Urdu", to bridge the identified gaps and limitations of existing Urdu datasets in the literature. It constitutes 10,083 fake and real news on fifteen domains collected from leading and authentic Urdu newspapers and news channel websites in Pakistan and India. The dataset contains news items in Urdu from the year 2017 to the year 2023. The selected models are originally trained on multilingual large corpora. The results of the proposed model are based on performance metrics, F1-score, accuracy, precision, recall and MCC value. F1-score of 0.924, accuracy of 0.956, precision of 0.942, recall of 0.940 and an MCC value of 0.902 demonstrate the effectiveness of the proposed approach for Urdu FND. Comparison analysis with SOTA ML and DL models and existing Urdu benchmark datasets exhibit that the ensemble model outperforms them for Urdu FND.

Large multimodal models demonstrate remarkable generalist ability to perform diverse multimodal tasks in a zero-shot manner. Large-scale web-based image-text pairs contribute fundamentally to this success, but suffer from excessive noise. Recent studies use alternative captions synthesized by captioning models and have achieved notable benchmark performance. However, our experiments reveal significant Scalability Deficiency and World Knowledge Loss issues in models trained with synthetic captions, which have been largely obscured by their initial benchmark success. Upon closer examination, we identify the root cause as the overly-simplified language structure and lack of knowledge details in existing synthetic captions. To provide higher-quality and more scalable multimodal pretraining data, we propose CapsFusion, an advanced framework that leverages large language models to consolidate and refine information from both web-based image-text pairs and synthetic captions. Extensive experiments show that CapsFusion captions exhibit remarkable all-round superiority over existing captions in terms of model performance (e.g., 18.8 and 18.3 improvements in CIDEr score on COCO and NoCaps), sample efficiency (requiring 11-16 times less computation than baselines), world knowledge depth, and scalability. These effectiveness, efficiency and scalability advantages position CapsFusion as a promising candidate for future scaling of LMM training.

Layered architectures have been widely used in robot systems. The majority of them implement planning and execution functions in separate layers. However, there still lacks a straightforward way to transit high-level tasks in the planning layer to the low-level motor commands in the execution layer. In order to tackle this challenge, we propose a novel approach to ground the manipulator primitive tasks to robot low-level actions using large language models (LLMs). We designed a program-function-like prompt based on the task frame formalism. In this way, we enable LLMs to generate position/force set-points for hybrid control. Evaluations over several state-of-the-art LLMs are provided.

As a distributed machine learning paradigm, Federated Learning (FL) enables large-scale clients to collaboratively train a model without sharing their raw data. However, due to the lack of data auditing for untrusted clients, FL is vulnerable to poisoning attacks, especially backdoor attacks. By using poisoned data for local training or directly changing the model parameters, attackers can easily inject backdoors into the model, which can trigger the model to make misclassification of targeted patterns in images. To address these issues, we propose a novel data-free trigger-generation-based defense approach based on the two characteristics of backdoor attacks: i) triggers are learned faster than normal knowledge, and ii) trigger patterns have a greater effect on image classification than normal class patterns. Our approach generates the images with newly learned knowledge by identifying the differences between the old and new global models, and filters trigger images by evaluating the effect of these generated images. By using these trigger images, our approach eliminates poisoned models to ensure the updated global model is benign. Comprehensive experiments demonstrate that our approach can defend against almost all the existing types of backdoor attacks and outperform all the seven state-of-the-art defense methods with both IID and non-IID scenarios. Especially, our approach can successfully defend against the backdoor attack even when 80\% of the clients are malicious.