Beijing Key Laboratory of Digital Media, School of Computer Science and Engineering, Beihang University, Beijing, China
Abstract:Robust deployment of large multimodal models (LMMs) in real-world scenarios requires access to external knowledge sources, given the complexity and dynamic nature of real-world information. Existing approaches such as retrieval-augmented generation (RAG) and prompt engineered search agents rely on rigid pipelines, often leading to inefficient or excessive search behaviors. We present MMSearch-R1, the first end-to-end reinforcement learning framework that enables LMMs to perform on-demand, multi-turn search in real-world Internet environments. Our framework integrates both image and text search tools, allowing the model to reason about when and how to invoke them guided by an outcome-based reward with a search penalty. To support training, We collect a multimodal search VQA dataset through a semi-automated pipeline that covers diverse visual and textual knowledge needs and curate a search-balanced subset with both search-required and search-free samples, which proves essential for shaping efficient and on-demand search behavior. Extensive experiments on knowledge-intensive and info-seeking VQA tasks show that our model not only outperforms RAG-based baselines of the same model size, but also matches the performance of a larger RAG-based model while reducing search calls by over 30%. We further analyze key empirical findings to offer actionable insights for advancing research in multimodal search.
Abstract:Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS) have emerged as powerful tools for 3D reconstruction and SLAM tasks. However, their performance depends heavily on accurate camera pose priors. Existing approaches attempt to address this issue by introducing external constraints but fall short of achieving satisfactory accuracy, particularly when camera trajectories are complex. In this paper, we propose a novel method, RA-NeRF, capable of predicting highly accurate camera poses even with complex camera trajectories. Following the incremental pipeline, RA-NeRF reconstructs the scene using NeRF with photometric consistency and incorporates flow-driven pose regulation to enhance robustness during initialization and localization. Additionally, RA-NeRF employs an implicit pose filter to capture the camera movement pattern and eliminate the noise for pose estimation. To validate our method, we conduct extensive experiments on the Tanks\&Temple dataset for standard evaluation, as well as the NeRFBuster dataset, which presents challenging camera pose trajectories. On both datasets, RA-NeRF achieves state-of-the-art results in both camera pose estimation and visual quality, demonstrating its effectiveness and robustness in scene reconstruction under complex pose trajectories.
Abstract:This paper presents Seewo's systems for both tracks of the Multilingual Conversational Speech Language Model Challenge (MLC-SLM), addressing automatic speech recognition (ASR) and speaker diarization with ASR (SD-ASR). We introduce a multi-stage training pipeline that explicitly enhances reasoning and self-correction in speech language models for ASR. Our approach combines curriculum learning for progressive capability acquisition, Chain-of-Thought data augmentation to foster intermediate reflection, and Reinforcement Learning with Verifiable Rewards (RLVR) to further refine self-correction through reward-driven optimization. This approach achieves substantial improvements over the official challenge baselines. On the evaluation set, our best system attains a WER/CER of 11.57% for Track 1 and a tcpWER/tcpCER of 17.67% for Track 2. Comprehensive ablation studies demonstrate the effectiveness of each component under challenge constraints.
Abstract:Large language models (LLMs) often struggle with knowledge-intensive tasks due to a lack of background knowledge and a tendency to hallucinate. To address these limitations, integrating knowledge graphs (KGs) with LLMs has been intensively studied. Existing KG-enhanced LLMs focus on supplementary factual knowledge, but still struggle with solving complex questions. We argue that refining the relationships among facts and organizing them into a logically consistent reasoning path is equally important as factual knowledge itself. Despite their potential, extracting reliable reasoning paths from KGs poses the following challenges: the complexity of graph structures and the existence of multiple generated paths, making it difficult to distinguish between useful and redundant ones. To tackle these challenges, we propose the RRP framework to mine the knowledge graph, which combines the semantic strengths of LLMs with structural information obtained through relation embedding and bidirectional distribution learning. Additionally, we introduce a rethinking module that evaluates and refines reasoning paths according to their significance. Experimental results on two public datasets show that RRP achieves state-of-the-art performance compared to existing baseline methods. Moreover, RRP can be easily integrated into various LLMs to enhance their reasoning abilities in a plug-and-play manner. By generating high-quality reasoning paths tailored to specific questions, RRP distills effective guidance for LLM reasoning.
Abstract:Large Audio-Language Models (LALMs) have significantly advanced intelligent human-computer interaction, yet their reliance on text-based outputs limits their ability to generate natural speech responses directly, hindering seamless audio interactions. To address this, we introduce Step-Audio-AQAA, a fully end-to-end LALM designed for Audio Query-Audio Answer (AQAA) tasks. The model integrates a dual-codebook audio tokenizer for linguistic and semantic feature extraction, a 130-billion-parameter backbone LLM and a neural vocoder for high-fidelity speech synthesis. Our post-training approach employs interleaved token-output of text and audio to enhance semantic coherence and combines Direct Preference Optimization (DPO) with model merge to improve performance. Evaluations on the StepEval-Audio-360 benchmark demonstrate that Step-Audio-AQAA excels especially in speech control, outperforming the state-of-art LALMs in key areas. This work contributes a promising solution for end-to-end LALMs and highlights the critical role of token-based vocoder in enhancing overall performance for AQAA tasks.
Abstract:Recent advances in diffusion models have significantly improved conditional video generation, particularly in the pose-guided human image animation task. Although existing methods are capable of generating high-fidelity and time-consistent animation sequences in regular motions and static scenes, there are still obvious limitations when facing complex human body motions (Hypermotion) that contain highly dynamic, non-standard motions, and the lack of a high-quality benchmark for evaluation of complex human motion animations. To address this challenge, we introduce the \textbf{Open-HyperMotionX Dataset} and \textbf{HyperMotionX Bench}, which provide high-quality human pose annotations and curated video clips for evaluating and improving pose-guided human image animation models under complex human motion conditions. Furthermore, we propose a simple yet powerful DiT-based video generation baseline and design spatial low-frequency enhanced RoPE, a novel module that selectively enhances low-frequency spatial feature modeling by introducing learnable frequency scaling. Our method significantly improves structural stability and appearance consistency in highly dynamic human motion sequences. Extensive experiments demonstrate the effectiveness of our dataset and proposed approach in advancing the generation quality of complex human motion image animations. Code and dataset will be made publicly available.
Abstract:Clinical decision-making is inherently complex and fast-paced, particularly in emergency departments (EDs) where critical, rapid and high-stakes decisions are made. Clinical Decision Rules (CDRs) are standardized evidence-based tools that combine signs, symptoms, and clinical variables into decision trees to make consistent and accurate diagnoses. CDR usage is often hindered by the clinician's cognitive load, limiting their ability to quickly recall and apply the appropriate rules. We introduce CDR-Agent, a novel LLM-based system designed to enhance ED decision-making by autonomously identifying and applying the most appropriate CDRs based on unstructured clinical notes. To validate CDR-Agent, we curated two novel ED datasets: synthetic and CDR-Bench, although CDR-Agent is applicable to non ED clinics. CDR-Agent achieves a 56.3\% (synthetic) and 8.7\% (CDR-Bench) accuracy gain relative to the standalone LLM baseline in CDR selection. Moreover, CDR-Agent significantly reduces computational overhead. Using these datasets, we demonstrated that CDR-Agent not only selects relevant CDRs efficiently, but makes cautious yet effective imaging decisions by minimizing unnecessary interventions while successfully identifying most positively diagnosed cases, outperforming traditional LLM prompting approaches. Code for our work can be found at: https://github.com/zhenxianglance/medagent-cdr-agent
Abstract:Video Virtual Try-On (VVT) aims to simulate the natural appearance of garments across consecutive video frames, capturing their dynamic variations and interactions with human body motion. However, current VVT methods still face challenges in terms of spatiotemporal consistency and garment content preservation. First, they use diffusion models based on the U-Net, which are limited in their expressive capability and struggle to reconstruct complex details. Second, they adopt a separative modeling approach for spatial and temporal attention, which hinders the effective capture of structural relationships and dynamic consistency across frames. Third, their expression of garment details remains insufficient, affecting the realism and stability of the overall synthesized results, especially during human motion. To address the above challenges, we propose MagicTryOn, a video virtual try-on framework built upon the large-scale video diffusion Transformer. We replace the U-Net architecture with a diffusion Transformer and combine full self-attention to jointly model the spatiotemporal consistency of videos. We design a coarse-to-fine garment preservation strategy. The coarse strategy integrates garment tokens during the embedding stage, while the fine strategy incorporates multiple garment-based conditions, such as semantics, textures, and contour lines during the denoising stage. Moreover, we introduce a mask-aware loss to further optimize garment region fidelity. Extensive experiments on both image and video try-on datasets demonstrate that our method outperforms existing SOTA methods in comprehensive evaluations and generalizes to in-the-wild scenarios.
Abstract:Recent advances in diffusion based editing models have enabled realistic camera simulation and image-based bokeh, but video bokeh remains largely unexplored. Existing video editing models cannot explicitly control focus planes or adjust bokeh intensity, limiting their applicability for controllable optical effects. Moreover, naively extending image-based bokeh methods to video often results in temporal flickering and unsatisfactory edge blur transitions due to the lack of temporal modeling and generalization capability. To address these challenges, we propose a novel one-step video bokeh framework that converts arbitrary input videos into temporally coherent, depth-aware bokeh effects. Our method leverages a multi-plane image (MPI) representation constructed through a progressively widening depth sampling function, providing explicit geometric guidance for depth-dependent blur synthesis. By conditioning a single-step video diffusion model on MPI layers and utilizing the strong 3D priors from pre-trained models such as Stable Video Diffusion, our approach achieves realistic and consistent bokeh effects across diverse scenes. Additionally, we introduce a progressive training strategy to enhance temporal consistency, depth robustness, and detail preservation. Extensive experiments demonstrate that our method produces high-quality, controllable bokeh effects and achieves state-of-the-art performance on multiple evaluation benchmarks.
Abstract:Multimodal large language models (MLLMs) remain vulnerable to transferable adversarial examples. While existing methods typically achieve targeted attacks by aligning global features-such as CLIP's [CLS] token-between adversarial and target samples, they often overlook the rich local information encoded in patch tokens. This leads to suboptimal alignment and limited transferability, particularly for closed-source models. To address this limitation, we propose a targeted transferable adversarial attack method based on feature optimal alignment, called FOA-Attack, to improve adversarial transfer capability. Specifically, at the global level, we introduce a global feature loss based on cosine similarity to align the coarse-grained features of adversarial samples with those of target samples. At the local level, given the rich local representations within Transformers, we leverage clustering techniques to extract compact local patterns to alleviate redundant local features. We then formulate local feature alignment between adversarial and target samples as an optimal transport (OT) problem and propose a local clustering optimal transport loss to refine fine-grained feature alignment. Additionally, we propose a dynamic ensemble model weighting strategy to adaptively balance the influence of multiple models during adversarial example generation, thereby further improving transferability. Extensive experiments across various models demonstrate the superiority of the proposed method, outperforming state-of-the-art methods, especially in transferring to closed-source MLLMs. The code is released at https://github.com/jiaxiaojunQAQ/FOA-Attack.