Yihan Zeng

Zehao Wang, Guanglei Yang, Yihan Zeng et al.

Federated fine-tuning of foundation models with Low-Rank Adaptation (LoRA) provides an efficient solution for reducing communication and computation costs while preserving data locality. However, the direct combination of FedAvg and LoRA suffers from three key issues: limited update space, which restricts the model's effective learning capacity; inter-round state mismatch, which disrupts cross-round local optimization continuity; and a client-agnostic starting state, which slows local convergence on clients. Although recent methods mitigate the limited update space issue by merging LoRA updates into the backbone across communication rounds, inter-round state mismatch and the client-agnostic starting state remain insufficiently addressed. To address these issues, we propose FedSmoothLoRA, a federated LoRA tuning framework that preserves the enlarged update space, improves cross-round local optimization continuity, and provides a client-aware starting state for local training. At each communication round, FedSmoothLoRA constructs the local LoRA initialization using two matrices: a Round-Matching matrix that preserves cross-round local state continuity, and a Gradient-Aligned matrix that provides client-specific optimization guidance from gradient signals estimated on local data. Together, these designs enable smoother and faster convergence. Extensive experiments on image classification and natural language generation tasks demonstrate that FedSmoothLoRA consistently outperforms existing federated LoRA tuning methods. Code: https://github.com/wangzehao0704/FedSmoothLoRA

Kai Chen, Yunhao Gou, Runhui Huang et al.

GPT-4o, an omni-modal model that enables vocal conversations with diverse emotions and tones, marks a milestone for omni-modal foundation models. However, empowering Large Language Models to perceive and generate images, texts, and speeches end-to-end with publicly available data remains challenging for the open-source community. Existing vision-language models rely on external tools for speech processing, while speech-language models still suffer from limited or totally without vision-understanding capabilities. To address this gap, we propose the EMOVA (EMotionally Omni-present Voice Assistant), to enable Large Language Models with end-to-end speech abilities while maintaining the leading vision-language performance. With a semantic-acoustic disentangled speech tokenizer, we surprisingly notice that omni-modal alignment can further enhance vision-language and speech abilities compared with the bi-modal aligned counterparts. Moreover, a lightweight style module is introduced for the flexible speech style controls including emotions and pitches. For the first time, EMOVA achieves state-of-the-art performance on both the vision-language and speech benchmarks, and meanwhile, supporting omni-modal spoken dialogue with vivid emotions.

Cuican Yu, Guansong Lu, Yihan Zeng et al. · tsinghua

Generating 3D faces from textual descriptions has a multitude of applications, such as gaming, movie, and robotics. Recent progresses have demonstrated the success of unconditional 3D face generation and text-to-3D shape generation. However, due to the limited text-3D face data pairs, text-driven 3D face generation remains an open problem. In this paper, we propose a text-guided 3D faces generation method, refer as TG-3DFace, for generating realistic 3D faces using text guidance. Specifically, we adopt an unconditional 3D face generation framework and equip it with text conditions, which learns the text-guided 3D face generation with only text-2D face data. On top of that, we propose two text-to-face cross-modal alignment techniques, including the global contrastive learning and the fine-grained alignment module, to facilitate high semantic consistency between generated 3D faces and input texts. Besides, we present directional classifier guidance during the inference process, which encourages creativity for out-of-domain generations. Compared to the existing methods, TG-3DFace creates more realistic and aesthetically pleasing 3D faces, boosting 9% multi-view consistency (MVIC) over Latent3D. The rendered face images generated by TG-3DFace achieve higher FID and CLIP score than text-to-2D face/image generation models, demonstrating our superiority in generating realistic and semantic-consistent textures.

Yihan Zeng, Chenhan Jiang, Jiageng Mao et al.

Contrastive Language-Image Pre-training, benefiting from large-scale unlabeled text-image pairs, has demonstrated great performance in open-world vision understanding tasks. However, due to the limited Text-3D data pairs, adapting the success of 2D Vision-Language Models (VLM) to the 3D space remains an open problem. Existing works that leverage VLM for 3D understanding generally resort to constructing intermediate 2D representations for the 3D data, but at the cost of losing 3D geometry information. To take a step toward open-world 3D vision understanding, we propose Contrastive Language-Image-Point Cloud Pretraining (CLIP$^2$) to directly learn the transferable 3D point cloud representation in realistic scenarios with a novel proxy alignment mechanism. Specifically, we exploit naturally-existed correspondences in 2D and 3D scenarios, and build well-aligned and instance-based text-image-point proxies from those complex scenarios. On top of that, we propose a cross-modal contrastive objective to learn semantic and instance-level aligned point cloud representation. Experimental results on both indoor and outdoor scenarios show that our learned 3D representation has great transfer ability in downstream tasks, including zero-shot and few-shot 3D recognition, which boosts the state-of-the-art methods by large margins. Furthermore, we provide analyses of the capability of different representations in real scenarios and present the optional ensemble scheme.

Tianyu Huang, Yihan Zeng, Bowen Dong et al.

Recent works learn 3D representation explicitly under text-3D guidance. However, limited text-3D data restricts the vocabulary scale and text control of generations. Generators may easily fall into a stereotype concept for certain text prompts, thus losing open-vocabulary generation ability. To tackle this issue, we introduce a conditional 3D generative model, namely TextField3D. Specifically, rather than using the text prompts as input directly, we suggest to inject dynamic noise into the latent space of given text prompts, i.e., Noisy Text Fields (NTFs). In this way, limited 3D data can be mapped to the appropriate range of textual latent space that is expanded by NTFs. To this end, an NTFGen module is proposed to model general text latent code in noisy fields. Meanwhile, an NTFBind module is proposed to align view-invariant image latent code to noisy fields, further supporting image-conditional 3D generation. To guide the conditional generation in both geometry and texture, multi-modal discrimination is constructed with a text-3D discriminator and a text-2.5D discriminator. Compared to previous methods, TextField3D includes three merits: 1) large vocabulary, 2) text consistency, and 3) low latency. Extensive experiments demonstrate that our method achieves a potential open-vocabulary 3D generation capability.

Chenhan Jiang, Yihan Zeng, Tianyang Hu et al.

Score Distillation Sampling (SDS) by well-trained 2D diffusion models has shown great promise in text-to-3D generation. However, this paradigm distills view-agnostic 2D image distributions into the rendering distribution of 3D representation for each view independently, overlooking the coherence across views and yielding 3D inconsistency in generations. In this work, we propose \textbf{J}oint \textbf{S}core \textbf{D}istillation (JSD), a new paradigm that ensures coherent 3D generations. Specifically, we model the joint image distribution, which introduces an energy function to capture the coherence among denoised images from the diffusion model. We then derive the joint score distillation on multiple rendered views of the 3D representation, as opposed to a single view in SDS. In addition, we instantiate three universal view-aware models as energy functions, demonstrating compatibility with JSD. Empirically, JSD significantly mitigates the 3D inconsistency problem in SDS, while maintaining text congruence. Moreover, we introduce the Geometry Fading scheme and Classifier-Free Guidance (CFG) Switching strategy to enhance generative details. Our framework, JointDreamer, establishes a new benchmark in text-to-3D generation, achieving outstanding results with an 88.5\% CLIP R-Precision and 27.7\% CLIP Score. These metrics demonstrate exceptional text congruence, as well as remarkable geometric consistency and texture fidelity.

Runhui Huang, Jianhua Han, Guansong Lu et al.

Recently, large-scale diffusion models, e.g., Stable diffusion and DallE2, have shown remarkable results on image synthesis. On the other hand, large-scale cross-modal pre-trained models (e.g., CLIP, ALIGN, and FILIP) are competent for various downstream tasks by learning to align vision and language embeddings. In this paper, we explore the possibility of jointly modeling generation and discrimination. Specifically, we propose DiffDis to unify the cross-modal generative and discriminative pretraining into one single framework under the diffusion process. DiffDis first formulates the image-text discriminative problem as a generative diffusion process of the text embedding from the text encoder conditioned on the image. Then, we propose a novel dual-stream network architecture, which fuses the noisy text embedding with the knowledge of latent images from different scales for image-text discriminative learning. Moreover, the generative and discriminative tasks can efficiently share the image-branch network structure in the multi-modality model. Benefiting from diffusion-based unified training, DiffDis achieves both better generation ability and cross-modal semantic alignment in one architecture. Experimental results show that DiffDis outperforms single-task models on both the image generation and the image-text discriminative tasks, e.g., 1.65% improvement on average accuracy of zero-shot classification over 12 datasets and 2.42 improvement on FID of zero-shot image synthesis.

Yang Cao, Yihan Zeng, Hang Xu et al.

Open-vocabulary 3D Object Detection (OV-3DDet) aims to detect objects from an arbitrary list of categories within a 3D scene, which remains seldom explored in the literature. There are primarily two fundamental problems in OV-3DDet, i.e., localizing and classifying novel objects. This paper aims at addressing the two problems simultaneously via a unified framework, under the condition of limited base categories. To localize novel 3D objects, we propose an effective 3D Novel Object Discovery strategy, which utilizes both the 3D box geometry priors and 2D semantic open-vocabulary priors to generate pseudo box labels of the novel objects. To classify novel object boxes, we further develop a cross-modal alignment module based on discovered novel boxes, to align feature spaces between 3D point cloud and image/text modalities. Specifically, the alignment process contains a class-agnostic and a class-discriminative alignment, incorporating not only the base objects with annotations but also the increasingly discovered novel objects, resulting in an iteratively enhanced alignment. The novel box discovery and crossmodal alignment are jointly learned to collaboratively benefit each other. The novel object discovery can directly impact the cross-modal alignment, while a better feature alignment can, in turn, boost the localization capability, leading to a unified OV-3DDet framework, named CoDA, for simultaneous novel object localization and classification. Extensive experiments on two challenging datasets (i.e., SUN-RGBD and ScanNet) demonstrate the effectiveness of our method and also show a significant mAP improvement upon the best-performing alternative method by 80%. Codes and pre-trained models are released on the project page.

Zehao Wang, Yihan Zeng, Zidong Gong et al.

Post-training via Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) is crucial for enhancing reasoning in Multimodal Large Language Models (MLLMs), yet existing paradigms often reach a performance bottleneck due to the limitations of static data. While current methods leverage self-reflection or self-evolution to push these boundaries, they still suffer from cognitive drift and hallucinated reasoning paths caused by low-quality synthetic data. To address these challenges, we propose Anchor Evolution (AnE), a new paradigm that integrates truth-anchored data curation and model evolution, achieving faithful and steady performance gains at the reasoning frontier. Specifically, we propose Truth Anchor Expansion, which pinpoints the model failing frontier via trajectory rollouts and leverages ground-truth databases to retrieve high-fidelity anchors for faithful data curation. Subsequently, we introduce the Scaffold-Stripping Mechanism to internalize reasoning capabilities. This mechanism first anchors reasoning paths via scaffold-augmented supervision to mitigate the learning complexity and distribution drift of direct SFT on raw data, then leverages RL to strip the scaffold template, thereby effectively transitioning the reasoning paths into intrinsic model capabilities. Experimental results on multimodal reasoning benchmarks show that our method substantially advances the model performance frontier, improving the base model by 10.3\% across eight multimodal benchmarks and achieving state-of-the-art results. The code will be made publicly available.

Zhenquan Yao, Zitong Huang, Yihan Zeng et al.

Graphical User Interface (GUI) Agents, benefiting from recent advances in multimodal large language models (MLLM), have achieved significant development. However, due to the frequent updates of GUI applications, adapting to new tasks without forgetting old tasks in GUI continual learning remains an open problem. In this work, we reveal that while Supervised Fine-Tuning (SFT) facilitates fast adaptation, it often triggers knowledge overwriting, whereas Reinforcement Learning (RL) demonstrates an inherent resilience that shields prior interaction logic from erasure. Based on this insight, we propose a \textbf{C}ontinual \textbf{G}UI \textbf{L}earning (CGL) framework that dynamically balances adaptation efficiency and skill retention by enhancing the synergy between SFT and RL. Specifically, we introduce an SFT proportion adjustment mechanism guided by policy entropy to dynamically control the weight allocation between the SFT and RL training phases. To resolve explicit gradient interference, we further develop a specialized gradient surgery strategy. By projecting exploratory SFT gradients onto GRPO-based anchor gradients, our method explicitly clips the components of SFT gradients that conflict with GRPO. On top of that, we establish an AndroidControl-CL benchmark, which divides GUI applications into distinct task groups to effectively simulate and evaluate the performance of continual GUI learning. Experimental results demonstrate the effectiveness of our proposed CGL framework across continual learning scenarios. The benchmark, code, and model will be made publicly available.

Tianyu Huang, Yihan Zeng, Zhilu Zhang et al.

3D generation has raised great attention in recent years. With the success of text-to-image diffusion models, the 2D-lifting technique becomes a promising route to controllable 3D generation. However, these methods tend to present inconsistent geometry, which is also known as the Janus problem. We observe that the problem is caused mainly by two aspects, i.e., viewpoint bias in 2D diffusion models and overfitting of the optimization objective. To address it, we propose a two-stage 2D-lifting framework, namely DreamControl, which optimizes coarse NeRF scenes as 3D self-prior and then generates fine-grained objects with control-based score distillation. Specifically, adaptive viewpoint sampling and boundary integrity metric are proposed to ensure the consistency of generated priors. The priors are then regarded as input conditions to maintain reasonable geometries, in which conditional LoRA and weighted score are further proposed to optimize detailed textures. DreamControl can generate high-quality 3D content in terms of both geometry consistency and texture fidelity. Moreover, our control-based optimization guidance is applicable to more downstream tasks, including user-guided generation and 3D animation. The project page is available at https://github.com/tyhuang0428/DreamControl.

Feng Zhu, Shuyang Xie, Yihan Zeng et al.

Real-world image restoration is challenging due to complex and interacting mixed degradations. Recent agent-based approaches address this problem by composing multiple task-specific restoration tools. However, empirical analysis reveals that their performance is fundamentally limited by implicitly constrained planning spaces and the lack of coordination among independently pretrained tools. To address these issues, we propose OPERA (Optimized Planning-Execution Restoration Agent), a framework that jointly optimizes restoration planning and tool execution in an end-to-end manner. On the planning side, OPERA uses reinforcement learning to directly optimize tool composition over a combinatorial plan space, with the final restoration quality as the reward. On the execution side, OPERA introduces agent-guided co-training of restoration tools, enabling them to learn cooperative behaviors under sequential composition. Extensive experiments on multi-degradation benchmarks and real-world datasets demonstrate that OPERA consistently outperforms both all-in-one restoration models and existing agent-based methods across diverse and complex degradation scenarios.

Zheyuan Zhou, Jiachen Lu, Yihan Zeng et al.

3D object detection from LiDAR point cloud is of critical importance for autonomous driving and robotics. While sequential point cloud has the potential to enhance 3D perception through temporal information, utilizing these temporal features effectively and efficiently remains a challenging problem. Based on the observation that the foreground information is sparsely distributed in LiDAR scenes, we believe sufficient knowledge can be provided by sparse format rather than dense maps. To this end, we propose to learn Significance-gUided Information for 3D Temporal detection (SUIT), which simplifies temporal information as sparse features for information fusion across frames. Specifically, we first introduce a significant sampling mechanism that extracts information-rich yet sparse features based on predicted object centroids. On top of that, we present an explicit geometric transformation learning technique, which learns the object-centric transformations among sparse features across frames. We evaluate our method on large-scale nuScenes and Waymo dataset, where our SUIT not only significantly reduces the memory and computation cost of temporal fusion, but also performs well over the state-of-the-art baselines.

Kun Xiang, Zhili Liu, Zihao Jiang et al.

In this paper, we address the challenging task of multimodal mathematical reasoning by incorporating the ability of "slow thinking" into multimodal large language models (MLLMs). Our core idea is that different levels of reasoning abilities can be combined dynamically to tackle questions with different complexity. To this end, we propose a paradigm of Self-structured Chain of Thought (SCoT), which is composed of minimal semantic atomic steps. Different from existing methods that rely on structured templates or free-form paradigms, our method can not only generate cognitive CoT structures for various complex tasks but also mitigates the phenomenon of overthinking. To introduce structured reasoning capabilities into visual understanding models, we further design a novel AtomThink framework with four key modules, including (i) a data engine to generate high-quality multimodal reasoning paths; (ii) a supervised fine-tuning process with serialized inference data; (iii) a policy-guided multi-turn inference method; and (iv) an atomic capability metric to evaluate the single step utilization rate. We conduct extensive experiments to show that the proposed AtomThink significantly improves the performance of baseline MLLMs, achieving more than 10\% average accuracy gains on MathVista and MathVerse. Compared to state-of-the-art structured CoT approaches, our method not only achieves higher accuracy but also improves data utilization by 5 times and boosts inference efficiency by 85.3\%. Our code is now public available in https://github.com/Quinn777/AtomThink.

Yabo Zhang, Xinpeng Zhou, Yihan Zeng et al.

Interactive image editing allows users to modify images through visual interaction operations such as drawing, clicking, and dragging. Existing methods construct such supervision signals from videos, as they capture how objects change with various physical interactions. However, these models are usually built upon text-to-image diffusion models, so necessitate (i) massive training samples and (ii) an additional reference encoder to learn real-world dynamics and visual consistency. In this paper, we reformulate this task as an image-to-video generation problem, so that inherit powerful video diffusion priors to reduce training costs and ensure temporal consistency. Specifically, we introduce FramePainter as an efficient instantiation of this formulation. Initialized with Stable Video Diffusion, it only uses a lightweight sparse control encoder to inject editing signals. Considering the limitations of temporal attention in handling large motion between two frames, we further propose matching attention to enlarge the receptive field while encouraging dense correspondence between edited and source image tokens. We highlight the effectiveness and efficiency of FramePainter across various of editing signals: it domainantly outperforms previous state-of-the-art methods with far less training data, achieving highly seamless and coherent editing of images, \eg, automatically adjust the reflection of the cup. Moreover, FramePainter also exhibits exceptional generalization in scenarios not present in real-world videos, \eg, transform the clownfish into shark-like shape. Our code will be available at https://github.com/YBYBZhang/FramePainter.

Xiuwei Chen, Wentao Hu, Hanhui Li et al.

Recent advances in multimodal large language models (MLLMs) have shown impressive reasoning capabilities. However, further enhancing existing MLLMs necessitates high-quality vision-language datasets with carefully curated task complexities, which are both costly and challenging to scale. Although recent self-improving models that iteratively refine themselves offer a feasible solution, they still suffer from two core challenges: (i) most existing methods augment visual or textual data separately, resulting in discrepancies in data complexity (e.g., over-simplified diagrams paired with redundant textual descriptions); and (ii) the evolution of data and models is also separated, leading to scenarios where models are exposed to tasks with mismatched difficulty levels. To address these issues, we propose C2-Evo, an automatic, closed-loop self-improving framework that jointly evolves both training data and model capabilities. Specifically, given a base dataset and a base model, C2-Evo enhances them by a cross-modal data evolution loop and a data-model evolution loop. The former loop expands the base dataset by generating complex multimodal problems that combine structured textual sub-problems with iteratively specified geometric diagrams, while the latter loop adaptively selects the generated problems based on the performance of the base model, to conduct supervised fine-tuning and reinforcement learning alternately. Consequently, our method continuously refines its model and training data, and consistently obtains considerable performance gains across multiple mathematical reasoning benchmarks. Our code, models, and datasets will be released.

Kun Xiang, Zhili Liu, Terry Jingchen Zhang et al.

In this paper, we address the challenging task of multimodal mathematical reasoning by incorporating the notion of ``slow thinking'' into multimodal large language models (MLLMs). Our core idea is that models can learn to adaptively use different levels of reasoning to tackle questions of different complexity. We propose a novel paradigm of Self-structured Chain of Thought (SCoT), which comprises of minimal semantic atomic steps. Different from existing methods that rely on structured templates or free-form paradigms, our method can not only generate cognitive CoT structures for various complex tasks but also mitigates the phenomena of overthinking for easier tasks. To introduce structured reasoning into visual cognition, we further design a novel AtomThink framework with four key modules, including (i) a data engine to generate high-quality multimodal reasoning paths; (ii) a supervised fine-tuning (SFT) process with serialized inference data; (iii) a policy-guided multi-turn inference method; and (iv) an atomic capability metric to evaluate the single step utilization rate. We conduct extensive experiments to show that the proposed AtomThink significantly improves the performance of baseline MLLMs, achieving more than 10\% average accuracy gains on MathVista and MathVerse. Compared to state-of-the-art structured CoT approaches, our method not only achieves higher accuracy but also improves data utilization by 5 times and boosts inference efficiency by 85.3\%. Our code is now public available in https://github.com/Quinn777/AtomThink.

Yabo Zhang, Yihan Zeng, Qingyun Li et al.

Large language models (LLMs) have demonstrated strong capabilities in language understanding and reasoning, yet they remain limited when tackling real-world tasks that require up-to-date knowledge, precise operations, or specialized tool use. To address this, we propose Tool-R1, a reinforcement learning framework that enables LLMs to perform general, compositional, and multi-step tool use by generating executable Python code. Tool-R1 supports integration of user-defined tools and standard libraries, with variable sharing across steps to construct coherent workflows. An outcome-based reward function, combining LLM-based answer judgment and code execution success, guides policy optimization. To improve training efficiency, we maintain a dynamic sample queue to cache and reuse high-quality trajectories, reducing the overhead of costly online sampling. Experiments on the GAIA benchmark show that Tool-R1 substantially improves both accuracy and robustness, achieving about 10\% gain over strong baselines, with larger improvements on complex multi-step tasks. These results highlight the potential of Tool-R1 for enabling reliable and efficient tool-augmented reasoning in real-world applications. Our code will be available at https://github.com/YBYBZhang/Tool-R1.

Wan Xu, Feng Zhu, Yihan Zeng et al.

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.

Guansong Lu, Yuanfan Guo, Jianhua Han et al.

Current large-scale diffusion models represent a giant leap forward in conditional image synthesis, capable of interpreting diverse cues like text, human poses, and edges. However, their reliance on substantial computational resources and extensive data collection remains a bottleneck. On the other hand, the integration of existing diffusion models, each specialized for different controls and operating in unique latent spaces, poses a challenge due to incompatible image resolutions and latent space embedding structures, hindering their joint use. Addressing these constraints, we present "PanGu-Draw", a novel latent diffusion model designed for resource-efficient text-to-image synthesis that adeptly accommodates multiple control signals. We first propose a resource-efficient Time-Decoupling Training Strategy, which splits the monolithic text-to-image model into structure and texture generators. Each generator is trained using a regimen that maximizes data utilization and computational efficiency, cutting data preparation by 48% and reducing training resources by 51%. Secondly, we introduce "Coop-Diffusion", an algorithm that enables the cooperative use of various pre-trained diffusion models with different latent spaces and predefined resolutions within a unified denoising process. This allows for multi-control image synthesis at arbitrary resolutions without the necessity for additional data or retraining. Empirical validations of Pangu-Draw show its exceptional prowess in text-to-image and multi-control image generation, suggesting a promising direction for future model training efficiencies and generation versatility. The largest 5B T2I PanGu-Draw model is released on the Ascend platform. Project page: $\href{https://pangu-draw.github.io}{this~https~URL}$

Haoyuan Li, Yanpeng Zhou, Tao Tang et al.

Recent advancements in multi-modal 3D pre-training methods have shown promising efficacy in learning joint representations of text, images, and point clouds. However, adopting point clouds as 3D representation fails to fully capture the intricacies of the 3D world and exhibits a noticeable gap between the discrete points and the dense 2D pixels of images. To tackle this issue, we propose UniGS, integrating 3D Gaussian Splatting (3DGS) into multi-modal pre-training to enhance the 3D representation. We first rely on the 3DGS representation to model the 3D world as a collection of 3D Gaussians with color and opacity, incorporating all the information of the 3D scene while establishing a strong connection with 2D images. Then, to achieve Language-Image-3D pertaining, UniGS starts with a pre-trained vision-language model to establish a shared visual and textual space through extensive real-world image-text pairs. Subsequently, UniGS employs a 3D encoder to align the optimized 3DGS with the Language-Image representations to learn unified multi-modal representations. To facilitate the extraction of global explicit 3D features by the 3D encoder and achieve better cross-modal alignment, we additionally introduce a novel Gaussian-Aware Guidance module that guides the learning of fine-grained representations of the 3D domain. Through extensive experiments across the Objaverse, ABO, MVImgNet and SUN RGBD datasets with zero-shot classification, text-driven retrieval and open-world understanding tasks, we demonstrate the effectiveness of UniGS in learning a more general and stronger aligned multi-modal representation. Specifically, UniGS achieves leading results across different 3D tasks with remarkable improvements over previous SOTA, Uni3D, including on zero-shot classification (+9.36%), text-driven retrieval (+4.3%) and open-world understanding (+7.92%).

Haochen Jiang, Yueming Xu, Yihan Zeng et al.

3D reconstruction has been widely used in autonomous navigation fields of mobile robotics. However, the former research can only provide the basic geometry structure without the capability of open-world scene understanding, limiting advanced tasks like human interaction and visual navigation. Moreover, traditional 3D scene understanding approaches rely on expensive labeled 3D datasets to train a model for a single task with supervision. Thus, geometric reconstruction with zero-shot scene understanding i.e. Open vocabulary 3D Understanding and Reconstruction, is crucial for the future development of mobile robots. In this paper, we propose OpenOcc, a novel framework unifying the 3D scene reconstruction and open vocabulary understanding with neural radiance fields. We model the geometric structure of the scene with occupancy representation and distill the pre-trained open vocabulary model into a 3D language field via volume rendering for zero-shot inference. Furthermore, a novel semantic-aware confidence propagation (SCP) method has been proposed to relieve the issue of language field representation degeneracy caused by inconsistent measurements in distilled features. Experimental results show that our approach achieves competitive performance in 3D scene understanding tasks, especially for small and long-tail objects.

Yu Yang, Zhilu Zhang, Xiang Zhang et al.

Interactive world models that simulate object dynamics are crucial for robotics, VR, and AR. However, it remains a significant challenge to learn physics-consistent dynamics models from limited real-world video data, especially for deformable objects with spatially-varying physical properties. To overcome the challenge of data scarcity, we propose PhysWorld, a novel framework that utilizes a simulator to synthesize physically plausible and diverse demonstrations to learn efficient world models. Specifically, we first construct a physics-consistent digital twin within MPM simulator via constitutive model selection and global-to-local optimization of physical properties. Subsequently, we apply part-aware perturbations to the physical properties and generate various motion patterns for the digital twin, synthesizing extensive and diverse demonstrations. Finally, using these demonstrations, we train a lightweight GNN-based world model that is embedded with physical properties. The real video can be used to further refine the physical properties. PhysWorld achieves accurate and fast future predictions for various deformable objects, and also generalizes well to novel interactions. Experiments show that PhysWorld has competitive performance while enabling inference speeds 47 times faster than the recent state-of-the-art method, i.e., PhysTwin.

Haoyuan Li, Yanpeng Zhou, Yihan Zeng et al.

3D Shape represented as point cloud has achieve advancements in multimodal pre-training to align image and language descriptions, which is curial to object identification, classification, and retrieval. However, the discrete representations of point cloud lost the object's surface shape information and creates a gap between rendering results and 2D correspondences. To address this problem, we propose GS-CLIP for the first attempt to introduce 3DGS (3D Gaussian Splatting) into multimodal pre-training to enhance 3D representation. GS-CLIP leverages a pre-trained vision-language model for a learned common visual and textual space on massive real world image-text pairs and then learns a 3D Encoder for aligning 3DGS optimized per object. Additionally, a novel Gaussian-Aware Fusion is proposed to extract and fuse global explicit feature. As a general framework for language-image-3D pre-training, GS-CLIP is agnostic to 3D backbone networks. Experiments on challenging shows that GS-CLIP significantly improves the state-of-the-art, outperforming the previously best results.

Chenhan Jiang, Yihan Zeng, Dit-Yan Yeung

Score Distillation Sampling (SDS) has achieved remarkable success in text-to-3D content generation. However, SDS-based methods struggle to maintain semantic fidelity for user prompts, particularly when involving multiple objects with intricate interactions. While existing approaches often address 3D consistency through multiview diffusion model fine-tuning on 3D datasets, this strategy inadvertently exacerbates text-3D alignment degradation. The limitation stems from SDS's inherent accumulation of view-independent biases during optimization, which progressively diverges from the ideal text alignment direction. To alleviate this limitation, we propose a novel SDS objective, dubbed as Textual Coherent Score Distillation (TCSD), which integrates alignment feedback from multimodal large language models (MLLMs). Our TCSD leverages cross-modal understanding capabilities of MLLMs to assess and guide the text-3D correspondence during the optimization. We further develop 3DLLaVA-CRITIC - a fine-tuned MLLM specialized for evaluating multiview text alignment in 3D generations. Additionally, we introduce an LLM-layout initialization that significantly accelerates optimization convergence through semantic-aware spatial configuration. Our framework, CoherenDream, achieves consistent improvement across multiple metrics on TIFA subset.As the first study to incorporate MLLMs into SDS optimization, we also conduct extensive ablation studies to explore optimal MLLM adaptations for 3D generation tasks.

Yunhao Gou, Hansi Yang, Zhili Liu et al.

Visual Instruction Tuning (VIT) aims to enhance Multimodal Large Language Models (MLLMs), yet its effectiveness is often compromised by corrupted datasets with issues such as hallucinated content, incorrect responses, and poor OCR quality. Previous approaches to address these challenges have focused on refining datasets through high-quality data collection or rule-based filtering that can be costly or limited in scope. In this paper, we conduct a systematic investigation into the impact of corrupted data on MLLMs and discover that, although corrupted data degrade model performance, such adverse effects are largely reversible, and MLLMs are {\bf corrupted but not broken}. Specifically, we find that disabling a small subset of parameters can almost fully restore performance. Moreover, corrupted MLLMs inherently possess the capability to differentiate between clean and corrupted samples, facilitating dataset cleaning without external intervention. Building on these insights, we introduce a corruption-robust training paradigm that significantly surpasses existing strategies for mitigating the effects of corrupted data.

Tianyu Huang, Haoze Zhang, Yihan Zeng et al.

Dynamic 3D interaction has been attracting a lot of attention recently. However, creating such 4D content remains challenging. One solution is to animate 3D scenes with physics-based simulation, which requires manually assigning precise physical properties to the object or the simulated results would become unnatural. Another solution is to learn the deformation of 3D objects with the distillation of video generative models, which, however, tends to produce 3D videos with small and discontinuous motions due to the inappropriate extraction and application of physics priors. In this work, to combine the strengths and complementing shortcomings of the above two solutions, we propose to learn the physical properties of a material field with video diffusion priors, and then utilize a physics-based Material-Point-Method (MPM) simulator to generate 4D content with realistic motions. In particular, we propose motion distillation sampling to emphasize video motion information during distillation. In addition, to facilitate the optimization, we further propose a KAN-based material field with frame boosting. Experimental results demonstrate that our method enjoys more realistic motions than state-of-the-arts do.

Yang Cao, Yihan Zeng, Hang Xu et al.

Open-vocabulary 3D Object Detection (OV-3DDet) addresses the detection of objects from an arbitrary list of novel categories in 3D scenes, which remains a very challenging problem. In this work, we propose CoDAv2, a unified framework designed to innovatively tackle both the localization and classification of novel 3D objects, under the condition of limited base categories. For localization, the proposed 3D Novel Object Discovery (3D-NOD) strategy utilizes 3D geometries and 2D open-vocabulary semantic priors to discover pseudo labels for novel objects during training. 3D-NOD is further extended with an Enrichment strategy that significantly enriches the novel object distribution in the training scenes, and then enhances the model's ability to localize more novel objects. The 3D-NOD with Enrichment is termed 3D-NODE. For classification, the Discovery-driven Cross-modal Alignment (DCMA) module aligns features from 3D point clouds and 2D/textual modalities, employing both class-agnostic and class-specific alignments that are iteratively refined to handle the expanding vocabulary of objects. Besides, 2D box guidance boosts the classification accuracy against complex background noises, which is coined as Box-DCMA. Extensive evaluation demonstrates the superiority of CoDAv2. CoDAv2 outperforms the best-performing method by a large margin (AP_Novel of 9.17 vs. 3.61 on SUN-RGBD and 9.12 vs. 3.74 on ScanNetv2). Source code and pre-trained models are available at the GitHub project page.