Jiawei Ren

Tong Wu, Jiarui Zhang, Xiao Fu et al.

Recent advances in modeling 3D objects mostly rely on synthetic datasets due to the lack of large-scale realscanned 3D databases. To facilitate the development of 3D perception, reconstruction, and generation in the real world, we propose OmniObject3D, a large vocabulary 3D object dataset with massive high-quality real-scanned 3D objects. OmniObject3D has several appealing properties: 1) Large Vocabulary: It comprises 6,000 scanned objects in 190 daily categories, sharing common classes with popular 2D datasets (e.g., ImageNet and LVIS), benefiting the pursuit of generalizable 3D representations. 2) Rich Annotations: Each 3D object is captured with both 2D and 3D sensors, providing textured meshes, point clouds, multiview rendered images, and multiple real-captured videos. 3) Realistic Scans: The professional scanners support highquality object scans with precise shapes and realistic appearances. With the vast exploration space offered by OmniObject3D, we carefully set up four evaluation tracks: a) robust 3D perception, b) novel-view synthesis, c) neural surface reconstruction, and d) 3D object generation. Extensive studies are performed on these four benchmarks, revealing new observations, challenges, and opportunities for future research in realistic 3D vision.

Jiaxiang Tang, Jiawei Ren, Hang Zhou et al.

Recent advances in 3D content creation mostly leverage optimization-based 3D generation via score distillation sampling (SDS). Though promising results have been exhibited, these methods often suffer from slow per-sample optimization, limiting their practical usage. In this paper, we propose DreamGaussian, a novel 3D content generation framework that achieves both efficiency and quality simultaneously. Our key insight is to design a generative 3D Gaussian Splatting model with companioned mesh extraction and texture refinement in UV space. In contrast to the occupancy pruning used in Neural Radiance Fields, we demonstrate that the progressive densification of 3D Gaussians converges significantly faster for 3D generative tasks. To further enhance the texture quality and facilitate downstream applications, we introduce an efficient algorithm to convert 3D Gaussians into textured meshes and apply a fine-tuning stage to refine the details. Extensive experiments demonstrate the superior efficiency and competitive generation quality of our proposed approach. Notably, DreamGaussian produces high-quality textured meshes in just 2 minutes from a single-view image, achieving approximately 10 times acceleration compared to existing methods.

Tianchang Shen, Sherwin Bahmani, Kai He et al. · nvidia, utoronto

Recent advances in video generation enable a new paradigm for 3D scene creation: generating camera-controlled videos that simulate scene walkthroughs, then lifting them to 3D via feed-forward reconstruction techniques. This generative reconstruction approach combines the visual fidelity and creative capacity of video models with 3D outputs ready for real-time rendering and simulation. Scaling to large, complex environments requires 3D-consistent video generation over long camera trajectories with large viewpoint changes and location revisits, a setting where current video models degrade quickly. Existing methods for long-horizon generation are fundamentally limited by two forms of degradation: spatial forgetting and temporal drifting. As exploration proceeds, previously observed regions fall outside the model's temporal context, forcing the model to hallucinate structures when revisited. Meanwhile, autoregressive generation accumulates small synthesis errors over time, gradually distorting scene appearance and geometry. We present Lyra 2.0, a framework for generating persistent, explorable 3D worlds at scale. To address spatial forgetting, we maintain per-frame 3D geometry and use it solely for information routing -- retrieving relevant past frames and establishing dense correspondences with the target viewpoints -- while relying on the generative prior for appearance synthesis. To address temporal drifting, we train with self-augmented histories that expose the model to its own degraded outputs, teaching it to correct drift rather than propagate it. Together, these enable substantially longer and 3D-consistent video trajectories, which we leverage to fine-tune feed-forward reconstruction models that reliably recover high-quality 3D scenes.

Lingdong Kong, Youquan Liu, Xin Li et al.

The robustness of 3D perception systems under natural corruptions from environments and sensors is pivotal for safety-critical applications. Existing large-scale 3D perception datasets often contain data that are meticulously cleaned. Such configurations, however, cannot reflect the reliability of perception models during the deployment stage. In this work, we present Robo3D, the first comprehensive benchmark heading toward probing the robustness of 3D detectors and segmentors under out-of-distribution scenarios against natural corruptions that occur in real-world environments. Specifically, we consider eight corruption types stemming from severe weather conditions, external disturbances, and internal sensor failure. We uncover that, although promising results have been progressively achieved on standard benchmarks, state-of-the-art 3D perception models are at risk of being vulnerable to corruptions. We draw key observations on the use of data representations, augmentation schemes, and training strategies, that could severely affect the model's performance. To pursue better robustness, we propose a density-insensitive training framework along with a simple flexible voxelization strategy to enhance the model resiliency. We hope our benchmark and approach could inspire future research in designing more robust and reliable 3D perception models. Our robustness benchmark suite is publicly available.

Lingdong Kong, Jiawei Ren, Liang Pan et al.

Densely annotating LiDAR point clouds is costly, which restrains the scalability of fully-supervised learning methods. In this work, we study the underexplored semi-supervised learning (SSL) in LiDAR segmentation. Our core idea is to leverage the strong spatial cues of LiDAR point clouds to better exploit unlabeled data. We propose LaserMix to mix laser beams from different LiDAR scans, and then encourage the model to make consistent and confident predictions before and after mixing. Our framework has three appealing properties: 1) Generic: LaserMix is agnostic to LiDAR representations (e.g., range view and voxel), and hence our SSL framework can be universally applied. 2) Statistically grounded: We provide a detailed analysis to theoretically explain the applicability of the proposed framework. 3) Effective: Comprehensive experimental analysis on popular LiDAR segmentation datasets (nuScenes, SemanticKITTI, and ScribbleKITTI) demonstrates our effectiveness and superiority. Notably, we achieve competitive results over fully-supervised counterparts with 2x to 5x fewer labels and improve the supervised-only baseline significantly by 10.8% on average. We hope this concise yet high-performing framework could facilitate future research in semi-supervised LiDAR segmentation. Code is publicly available.

Shaoyuan Xie, Lingdong Kong, Wenwei Zhang et al.

The recent advances in camera-based bird's eye view (BEV) representation exhibit great potential for in-vehicle 3D perception. Despite the substantial progress achieved on standard benchmarks, the robustness of BEV algorithms has not been thoroughly examined, which is critical for safe operations. To bridge this gap, we introduce RoboBEV, a comprehensive benchmark suite that encompasses eight distinct corruptions, including Bright, Dark, Fog, Snow, Motion Blur, Color Quant, Camera Crash, and Frame Lost. Based on it, we undertake extensive evaluations across a wide range of BEV-based models to understand their resilience and reliability. Our findings indicate a strong correlation between absolute performance on in-distribution and out-of-distribution datasets. Nonetheless, there are considerable variations in relative performance across different approaches. Our experiments further demonstrate that pre-training and depth-free BEV transformation has the potential to enhance out-of-distribution robustness. Additionally, utilizing long and rich temporal information largely helps with robustness. Our findings provide valuable insights for designing future BEV models that can achieve both accuracy and robustness in real-world deployments.

Bo Li, Yifei Shen, Jingkang Yang et al.

Human visual perception can easily generalize to out-of-distributed visual data, which is far beyond the capability of modern machine learning models. Domain generalization (DG) aims to close this gap, with existing DG methods mainly focusing on the loss function design. In this paper, we propose to explore an orthogonal direction, i.e., the design of the backbone architecture. It is motivated by an empirical finding that transformer-based models trained with empirical risk minimization (ERM) outperform CNN-based models employing state-of-the-art (SOTA) DG algorithms on multiple DG datasets. We develop a formal framework to characterize a network's robustness to distribution shifts by studying its architecture's alignment with the correlations in the dataset. This analysis guides us to propose a novel DG model built upon vision transformers, namely Generalizable Mixture-of-Experts (GMoE). Extensive experiments on DomainBed demonstrate that GMoE trained with ERM outperforms SOTA DG baselines by a large margin. Moreover, GMoE is complementary to existing DG methods and its performance is substantially improved when trained with DG algorithms.

Jiawei Ren, Cunjun Yu, Siwei Chen et al.

Motion mimicking is a foundational task in physics-based character animation. However, most existing motion mimicking methods are built upon reinforcement learning (RL) and suffer from heavy reward engineering, high variance, and slow convergence with hard explorations. Specifically, they usually take tens of hours or even days of training to mimic a simple motion sequence, resulting in poor scalability. In this work, we leverage differentiable physics simulators (DPS) and propose an efficient motion mimicking method dubbed DiffMimic. Our key insight is that DPS casts a complex policy learning task to a much simpler state matching problem. In particular, DPS learns a stable policy by analytical gradients with ground-truth physical priors hence leading to significantly faster and stabler convergence than RL-based methods. Moreover, to escape from local optima, we utilize a Demonstration Replay mechanism to enable stable gradient backpropagation in a long horizon. Extensive experiments on standard benchmarks show that DiffMimic has a better sample efficiency and time efficiency than existing methods (e.g., DeepMimic). Notably, DiffMimic allows a physically simulated character to learn Backflip after 10 minutes of training and be able to cycle it after 3 hours of training, while the existing approach may require about a day of training to cycle Backflip. More importantly, we hope DiffMimic can benefit more differentiable animation systems with techniques like differentiable clothes simulation in future research.

Yuren Cong, Mengmeng Xu, Christian Simon et al.

Text-to-video editing aims to edit the visual appearance of a source video conditional on textual prompts. A major challenge in this task is to ensure that all frames in the edited video are visually consistent. Most recent works apply advanced text-to-image diffusion models to this task by inflating 2D spatial attention in the U-Net into spatio-temporal attention. Although temporal context can be added through spatio-temporal attention, it may introduce some irrelevant information for each patch and therefore cause inconsistency in the edited video. In this paper, for the first time, we introduce optical flow into the attention module in the diffusion model's U-Net to address the inconsistency issue for text-to-video editing. Our method, FLATTEN, enforces the patches on the same flow path across different frames to attend to each other in the attention module, thus improving the visual consistency in the edited videos. Additionally, our method is training-free and can be seamlessly integrated into any diffusion-based text-to-video editing methods and improve their visual consistency. Experiment results on existing text-to-video editing benchmarks show that our proposed method achieves the new state-of-the-art performance. In particular, our method excels in maintaining the visual consistency in the edited videos.

Alessandro Burzio, Tobias Fischer, Sven Elflein et al.

Recent feed-forward 3D reconstruction transformers have scaled to over a billion parameters, following the broader trend of increasing model capacity in computer vision. Yet emerging evidence suggests that contiguous transformer layers often behave like repeated applications of similar operations, and multi-view reconstruction transformers refine their predictions progressively across decoder depth. We posit that model depth partially buys iteration, paid for inefficiently in unique parameters, and instead make that iteration explicit in architecture. Our model, DéjàView, applies a single looped transformer block recurrently to per-view features for K refinement steps. Trained once, it exposes K as an inference-time compute knob, matching or outperforming substantially larger feed-forward baselines across five reconstruction benchmarks spanning indoor, outdoor, object-centric, and driving scenes, while using a fraction of their parameters and comparable or lower compute. Importantly, the same looped block formulation outperforms an otherwise identical variant with independent per-step parameters under matched training data and compute, suggesting that explicit iteration is not merely a compute-efficient substitute for capacity but a stronger inductive bias for multi-view 3D reconstruction.

Jiawei Ren, Michal Jan Tyszkiewicz, Jiahui Huang et al.

In this work, we revisit several key design choices of modern Transformer-based approaches for feed-forward 3D Gaussian Splatting (3DGS) prediction. We argue that the common practice of regressing Gaussian means as depths along camera rays is suboptimal, and instead propose to directly regress 3D mean coordinates using only a self-supervised rendering loss. This formulation allows us to move from the standard encoder-only design to an encoder-decoder architecture with learnable Gaussian tokens, thereby unbinding the number of predicted primitives from input image resolution and number of views. Our resulting method, TokenGS, demonstrates improved robustness to pose noise and multiview inconsistencies, while naturally supporting efficient test-time optimization in token space without degrading learned priors. TokenGS achieves state-of-the-art feed-forward reconstruction performance on both static and dynamic scenes, producing more regularized geometry and more balanced 3DGS distribution, while seamlessly recovering emergent scene attributes such as static-dynamic decomposition and scene flow.

Tianshi Cao, Jiawei Ren, Yuxuan Zhang et al.

Closed-loop simulation is a core component of autonomous vehicle (AV) development, enabling scalable testing, training, and safety validation before real-world deployment. Neural scene reconstruction converts driving logs into interactive 3D environments for simulation, but it does not produce complete 3D object assets required for agent manipulation and large-viewpoint novel-view synthesis. To address this challenge, we present Asset Harvester, an image-to-3D model and end-to-end pipeline that converts sparse, in-the-wild object observations from real driving logs into complete, simulation-ready assets. Rather than relying on a single model component, we developed a system-level design for real-world AV data that combines large-scale curation of object-centric training tuples, geometry-aware preprocessing across heterogeneous sensors, and a robust training recipe that couples sparse-view-conditioned multiview generation with 3D Gaussian lifting. Within this system, SparseViewDiT is explicitly designed to address limited-angle views and other real-world data challenges. Together with hybrid data curation, augmentation, and self-distillation, this system enables scalable conversion of sparse AV object observations into reusable 3D assets.

Jiawei Ren, Yan Zhuang, Xiaokang Ye et al.

While LLM/VLM-powered AI agents have advanced rapidly in math, coding, and computer use, their applications in complex physical and social environments remain challenging. Building agents that can survive and thrive in the real world (for example, by autonomously earning income or running a business) requires massive-scale interaction, reasoning, training, and evaluation across diverse embodied scenarios. However, existing world simulators for such development fall short: they often rely on limited hand-crafted environments, simulate simplified game-like physics and social rules, and lack native support for LLM/VLM agents. We introduce SimWorld, a new simulator built on Unreal Engine 5, designed for developing and evaluating LLM/VLM agents in rich, real-world-like settings. SimWorld offers three core capabilities: (1) realistic, open-ended world simulation, including accurate physical and social dynamics and language-driven procedural environment generation; (2) a rich interface for LLM/VLM agents, with multimodal world inputs and open-vocabulary actions at varying levels of abstraction; and (3) diverse and extensible physical and social reasoning scenarios that are easily customizable by users. We demonstrate SimWorld by deploying frontier LLM agents (e.g., GPT-4o, Gemini-2.5-Flash, Claude-3.5, and DeepSeek-Prover-V2) on long-horizon multi-agent delivery tasks involving strategic cooperation and competition. The results reveal distinct reasoning patterns and limitations across models. We open-source SimWorld and hope it becomes a foundational platform for advancing real-world agent intelligence across disciplines: https://simworld.org.

Jiahui Huang, Qunjie Zhou, Hesam Rabeti et al. · nvidia, utoronto

Accurate 3D geometric perception is an important prerequisite for a wide range of spatial AI systems. While state-of-the-art methods depend on large-scale training data, acquiring consistent and precise 3D annotations from in-the-wild videos remains a key challenge. In this work, we introduce ViPE, a handy and versatile video processing engine designed to bridge this gap. ViPE efficiently estimates camera intrinsics, camera motion, and dense, near-metric depth maps from unconstrained raw videos. It is robust to diverse scenarios, including dynamic selfie videos, cinematic shots, or dashcams, and supports various camera models such as pinhole, wide-angle, and 360° panoramas. We have benchmarked ViPE on multiple benchmarks. Notably, it outperforms existing uncalibrated pose estimation baselines by 18%/50% on TUM/KITTI sequences, and runs at 3-5FPS on a single GPU for standard input resolutions. We use ViPE to annotate a large-scale collection of videos. This collection includes around 100K real-world internet videos, 1M high-quality AI-generated videos, and 2K panoramic videos, totaling approximately 96M frames -- all annotated with accurate camera poses and dense depth maps. We open-source ViPE and the annotated dataset with the hope of accelerating the development of spatial AI systems.

Jiawei Ren, Liang Pan, Jiaxiang Tang et al.

4D content generation has achieved remarkable progress recently. However, existing methods suffer from long optimization times, a lack of motion controllability, and a low quality of details. In this paper, we introduce DreamGaussian4D (DG4D), an efficient 4D generation framework that builds on Gaussian Splatting (GS). Our key insight is that combining explicit modeling of spatial transformations with static GS makes an efficient and powerful representation for 4D generation. Moreover, video generation methods have the potential to offer valuable spatial-temporal priors, enhancing the high-quality 4D generation. Specifically, we propose an integral framework with two major modules: 1) Image-to-4D GS - we initially generate static GS with DreamGaussianHD, followed by HexPlane-based dynamic generation with Gaussian deformation; and 2) Video-to-Video Texture Refinement - we refine the generated UV-space texture maps and meanwhile enhance their temporal consistency by utilizing a pre-trained image-to-video diffusion model. Notably, DG4D reduces the optimization time from several hours to just a few minutes, allows the generated 3D motion to be visually controlled, and produces animated meshes that can be realistically rendered in 3D engines.

Yan Zhuang, Jiawei Ren, Xiaokang Ye et al.

Recent advances in foundation models have shown promising results in developing generalist robotics that can perform diverse tasks in open-ended scenarios given multimodal inputs. However, current work has been mainly focused on indoor, household scenarios. In this work, we present SimWorld-Robotics~(SWR), a simulation platform for embodied AI in large-scale, photorealistic urban environments. Built on Unreal Engine 5, SWR procedurally generates unlimited photorealistic urban scenes populated with dynamic elements such as pedestrians and traffic systems, surpassing prior urban simulations in realism, complexity, and scalability. It also supports multi-robot control and communication. With these key features, we build two challenging robot benchmarks: (1) a multimodal instruction-following task, where a robot must follow vision-language navigation instructions to reach a destination in the presence of pedestrians and traffic; and (2) a multi-agent search task, where two robots must communicate to cooperatively locate and meet each other. Unlike existing benchmarks, these two new benchmarks comprehensively evaluate a wide range of critical robot capacities in realistic scenarios, including (1) multimodal instructions grounding, (2) 3D spatial reasoning in large environments, (3) safe, long-range navigation with people and traffic, (4) multi-robot collaboration, and (5) grounded communication. Our experimental results demonstrate that state-of-the-art models, including vision-language models (VLMs), struggle with our tasks, lacking robust perception, reasoning, and planning abilities necessary for urban environments.

Mingyuan Zhang, Huirong Li, Zhongang Cai et al.

Text-driven motion generation has achieved substantial progress with the emergence of diffusion models. However, existing methods still struggle to generate complex motion sequences that correspond to fine-grained descriptions, depicting detailed and accurate spatio-temporal actions. This lack of fine controllability limits the usage of motion generation to a larger audience. To tackle these challenges, we present FineMoGen, a diffusion-based motion generation and editing framework that can synthesize fine-grained motions, with spatial-temporal composition to the user instructions. Specifically, FineMoGen builds upon diffusion model with a novel transformer architecture dubbed Spatio-Temporal Mixture Attention (SAMI). SAMI optimizes the generation of the global attention template from two perspectives: 1) explicitly modeling the constraints of spatio-temporal composition; and 2) utilizing sparsely-activated mixture-of-experts to adaptively extract fine-grained features. To facilitate a large-scale study on this new fine-grained motion generation task, we contribute the HuMMan-MoGen dataset, which consists of 2,968 videos and 102,336 fine-grained spatio-temporal descriptions. Extensive experiments validate that FineMoGen exhibits superior motion generation quality over state-of-the-art methods. Notably, FineMoGen further enables zero-shot motion editing capabilities with the aid of modern large language models (LLM), which faithfully manipulates motion sequences with fine-grained instructions. Project Page: https://mingyuan-zhang.github.io/projects/FineMoGen.html

Shoufa Chen, Mengmeng Xu, Jiawei Ren et al.

In this study, we explore Transformer-based diffusion models for image and video generation. Despite the dominance of Transformer architectures in various fields due to their flexibility and scalability, the visual generative domain primarily utilizes CNN-based U-Net architectures, particularly in diffusion-based models. We introduce GenTron, a family of Generative models employing Transformer-based diffusion, to address this gap. Our initial step was to adapt Diffusion Transformers (DiTs) from class to text conditioning, a process involving thorough empirical exploration of the conditioning mechanism. We then scale GenTron from approximately 900M to over 3B parameters, observing significant improvements in visual quality. Furthermore, we extend GenTron to text-to-video generation, incorporating novel motion-free guidance to enhance video quality. In human evaluations against SDXL, GenTron achieves a 51.1% win rate in visual quality (with a 19.8% draw rate), and a 42.3% win rate in text alignment (with a 42.9% draw rate). GenTron also excels in the T2I-CompBench, underscoring its strengths in compositional generation. We believe this work will provide meaningful insights and serve as a valuable reference for future research.

Lingdong Kong, Xiang Xu, Jiawei Ren et al.

Efficient data utilization is crucial for advancing 3D scene understanding in autonomous driving, where reliance on heavily human-annotated LiDAR point clouds challenges fully supervised methods. Addressing this, our study extends into semi-supervised learning for LiDAR semantic segmentation, leveraging the intrinsic spatial priors of driving scenes and multi-sensor complements to augment the efficacy of unlabeled datasets. We introduce LaserMix++, an evolved framework that integrates laser beam manipulations from disparate LiDAR scans and incorporates LiDAR-camera correspondences to further assist data-efficient learning. Our framework is tailored to enhance 3D scene consistency regularization by incorporating multi-modality, including 1) multi-modal LaserMix operation for fine-grained cross-sensor interactions; 2) camera-to-LiDAR feature distillation that enhances LiDAR feature learning; and 3) language-driven knowledge guidance generating auxiliary supervisions using open-vocabulary models. The versatility of LaserMix++ enables applications across LiDAR representations, establishing it as a universally applicable solution. Our framework is rigorously validated through theoretical analysis and extensive experiments on popular driving perception datasets. Results demonstrate that LaserMix++ markedly outperforms fully supervised alternatives, achieving comparable accuracy with five times fewer annotations and significantly improving the supervised-only baselines. This substantial advancement underscores the potential of semi-supervised approaches in reducing the reliance on extensive labeled data in LiDAR-based 3D scene understanding systems.

Hanxue Liang, Jiawei Ren, Ashkan Mirzaei et al.

Recent advancements in static feed-forward scene reconstruction have demonstrated significant progress in high-quality novel view synthesis. However, these models often struggle with generalizability across diverse environments and fail to effectively handle dynamic content. We present BTimer (short for BulletTimer), the first motion-aware feed-forward model for real-time reconstruction and novel view synthesis of dynamic scenes. Our approach reconstructs the full scene in a 3D Gaussian Splatting representation at a given target ('bullet') timestamp by aggregating information from all the context frames. Such a formulation allows BTimer to gain scalability and generalization by leveraging both static and dynamic scene datasets. Given a casual monocular dynamic video, BTimer reconstructs a bullet-time scene within 150ms while reaching state-of-the-art performance on both static and dynamic scene datasets, even compared with optimization-based approaches.

Jiawei Ren, Mingyuan Zhang, Cunjun Yu et al.

Generating animation of physics-based characters with intuitive control has long been a desirable task with numerous applications. However, generating physically simulated animations that reflect high-level human instructions remains a difficult problem due to the complexity of physical environments and the richness of human language. In this paper, we present InsActor, a principled generative framework that leverages recent advancements in diffusion-based human motion models to produce instruction-driven animations of physics-based characters. Our framework empowers InsActor to capture complex relationships between high-level human instructions and character motions by employing diffusion policies for flexibly conditioned motion planning. To overcome invalid states and infeasible state transitions in planned motions, InsActor discovers low-level skills and maps plans to latent skill sequences in a compact latent space. Extensive experiments demonstrate that InsActor achieves state-of-the-art results on various tasks, including instruction-driven motion generation and instruction-driven waypoint heading. Notably, the ability of InsActor to generate physically simulated animations using high-level human instructions makes it a valuable tool, particularly in executing long-horizon tasks with a rich set of instructions.

Lingjun Mao, Jiawei Ren, Kun Zhou et al.

LLMs and VLMs are increasingly deployed as embodied agents, yet existing benchmarks largely revolve around simple short-term tasks and struggle to capture rich realistic constraints that shape real-world decision making. To close this gap, we propose DeliveryBench, a city-scale embodied benchmark grounded in the real-world profession of food delivery. Food couriers naturally operate under long-horizon objectives (maximizing net profit over hours) while managing diverse constraints, e.g., delivery deadline, transportation expense, vehicle battery, and necessary interactions with other couriers and customers. DeliveryBench instantiates this setting in procedurally generated 3D cities with diverse road networks, buildings, functional locations, transportation modes, and realistic resource dynamics, enabling systematic evaluation of constraint-aware, long-horizon planning. We benchmark a range of VLM-based agents across nine cities and compare them with human players. Our results reveal a substantial performance gap to humans, and find that these agents are short-sighted and frequently break basic commonsense constraints. Additionally, we observe distinct personalities across models (e.g., adventurous GPT-5 vs. conservative Claude), highlighting both the brittleness and the diversity of current VLM-based embodied agents in realistic, constraint-dense environments. Our code, data, and benchmark are available at https://deliverybench.github.io.

Jiawei Ren, Mengmeng Xu, Jui-Chieh Wu et al.

Diffusion models generate images with an unprecedented level of quality, but how can we freely rearrange image layouts? Recent works generate controllable scenes via learning spatially disentangled latent codes, but these methods do not apply to diffusion models due to their fixed forward process. In this work, we propose SceneDiffusion to optimize a layered scene representation during the diffusion sampling process. Our key insight is that spatial disentanglement can be obtained by jointly denoising scene renderings at different spatial layouts. Our generated scenes support a wide range of spatial editing operations, including moving, resizing, cloning, and layer-wise appearance editing operations, including object restyling and replacing. Moreover, a scene can be generated conditioned on a reference image, thus enabling object moving for in-the-wild images. Notably, this approach is training-free, compatible with general text-to-image diffusion models, and responsive in less than a second.

Yuan Xie, Jiawei Ren, Junfeng Li et al.

Underwater acoustic target recognition has emerged as a prominent research area within the field of underwater acoustics. However, the current availability of authentic underwater acoustic signal recordings remains limited, which hinders data-driven acoustic recognition models from learning robust patterns of targets from a limited set of intricate underwater signals, thereby compromising their stability in practical applications. To overcome these limitations, this study proposes a recognition framework called M3 (Multi-task, Multi-gate, Multi-expert) to enhance the model's ability to capture robust patterns by making it aware of the inherent properties of targets. In this framework, an auxiliary task that focuses on target properties, such as estimating target size, is designed. The auxiliary task then shares parameters with the recognition task to realize multi-task learning. This paradigm allows the model to concentrate on shared information across tasks and identify robust patterns of targets in a regularized manner, thereby enhancing the model's generalization ability. Moreover, M3 incorporates multi-expert and multi-gate mechanisms, allowing for the allocation of distinct parameter spaces to various underwater signals. This enables the model to process intricate signal patterns in a fine-grained and differentiated manner. To evaluate the effectiveness of M3, extensive experiments were implemented on the ShipsEar underwater ship-radiated noise dataset. The results substantiate that M3 has the ability to outperform the most advanced single-task recognition models, thereby achieving the state-of-the-art performance.

Yuan Xie, Ji Xu, Jiawei Ren et al.

Underwater acoustic target recognition based on passive sonar faces numerous challenges in practical maritime applications. One of the main challenges lies in the susceptibility of signal characteristics to diverse environmental conditions and data acquisition configurations, which can lead to instability in recognition systems. While significant efforts have been dedicated to addressing these influential factors in other domains of underwater acoustics, they are often neglected in the field of underwater acoustic target recognition. To overcome this limitation, this study designs auxiliary tasks that model influential factors (e.g., source range, water column depth, or wind speed) based on available annotations and adopts a multi-task framework to connect these factors to the recognition task. Furthermore, we integrate an adversarial learning mechanism into the multi-task framework to prompt the model to extract representations that are robust against influential factors. Through extensive experiments and analyses on the ShipsEar dataset, our proposed adversarial multi-task model demonstrates its capacity to effectively model the influential factors and achieve state-of-the-art performance on the 12-class recognition task.

Sherwin Bahmani, Tianchang Shen, Jiawei Ren et al. · nvidia, utoronto

The ability to generate virtual environments is crucial for applications ranging from gaming to physical AI domains such as robotics, autonomous driving, and industrial AI. Current learning-based 3D reconstruction methods rely on the availability of captured real-world multi-view data, which is not always readily available. Recent advancements in video diffusion models have shown remarkable imagination capabilities, yet their 2D nature limits the applications to simulation where a robot needs to navigate and interact with the environment. In this paper, we propose a self-distillation framework that aims to distill the implicit 3D knowledge in the video diffusion models into an explicit 3D Gaussian Splatting (3DGS) representation, eliminating the need for multi-view training data. Specifically, we augment the typical RGB decoder with a 3DGS decoder, which is supervised by the output of the RGB decoder. In this approach, the 3DGS decoder can be purely trained with synthetic data generated by video diffusion models. At inference time, our model can synthesize 3D scenes from either a text prompt or a single image for real-time rendering. Our framework further extends to dynamic 3D scene generation from a monocular input video. Experimental results show that our framework achieves state-of-the-art performance in static and dynamic 3D scene generation.

Zhaoxi Chen, Tianqi Liu, Long Zhuo et al.

We present 4DNeX, the first feed-forward framework for generating 4D (i.e., dynamic 3D) scene representations from a single image. In contrast to existing methods that rely on computationally intensive optimization or require multi-frame video inputs, 4DNeX enables efficient, end-to-end image-to-4D generation by fine-tuning a pretrained video diffusion model. Specifically, 1) to alleviate the scarcity of 4D data, we construct 4DNeX-10M, a large-scale dataset with high-quality 4D annotations generated using advanced reconstruction approaches. 2) we introduce a unified 6D video representation that jointly models RGB and XYZ sequences, facilitating structured learning of both appearance and geometry. 3) we propose a set of simple yet effective adaptation strategies to repurpose pretrained video diffusion models for 4D modeling. 4DNeX produces high-quality dynamic point clouds that enable novel-view video synthesis. Extensive experiments demonstrate that 4DNeX outperforms existing 4D generation methods in efficiency and generalizability, offering a scalable solution for image-to-4D modeling and laying the foundation for generative 4D world models that simulate dynamic scene evolution.

Yuan Xie, Xiaowei Zhang, Jiawei Ren et al.

Building a robust underwater acoustic recognition system in real-world scenarios is challenging due to the complex underwater environment and the dynamic motion states of targets. A promising optimization approach is to leverage the intrinsic physical characteristics of targets, which remain invariable regardless of environmental conditions, to provide robust insights. However, our study reveals that while physical characteristics exhibit robust properties, they may lack class-specific discriminative patterns. Consequently, directly incorporating physical characteristics into model training can potentially introduce unintended inductive biases, leading to performance degradation. To utilize the benefits of physical characteristics while mitigating possible detrimental effects, we propose DEMONet in this study, which utilizes the detection of envelope modulation on noise (DEMON) to provide robust insights into the shaft frequency or blade counts of targets. DEMONet is a multi-expert network that allocates various underwater signals to their best-matched expert layer based on DEMON spectra for fine-grained signal processing. Thereinto, DEMON spectra are solely responsible for providing implicit physical characteristics without establishing a mapping relationship with the target category. Furthermore, to mitigate noise and spurious modulation spectra in DEMON features, we introduce a cross-temporal alignment strategy and employ a variational autoencoder (VAE) to reconstruct noise-resistant DEMON spectra to replace the raw DEMON features. The effectiveness of the proposed DEMONet with cross-temporal VAE was primarily evaluated on the DeepShip dataset and our proprietary datasets. Experimental results demonstrated that our approach could achieve state-of-the-art performance on both datasets.

Tianyang Zhang, Zhuoxuan Jiang, Shengguang Bai et al.

With the ever-increasing demands on Question Answering (QA) systems for IT operations and maintenance, an efficient and supervised fine-tunable framework is necessary to ensure the data security, private deployment and continuous upgrading. Although Large Language Models (LLMs) have notably improved the open-domain QA's performance, how to efficiently handle enterprise-exclusive corpora and build domain-specific QA systems are still less-studied for industrial applications. In this paper, we propose a general and comprehensive framework based on Retrieval Augmented Generation (RAG) and facilitate the whole business process of establishing QA systems for IT operations and maintenance. In accordance with the prevailing RAG method, our proposed framework, named with RAG4ITOps, composes of two major stages: (1) Models Fine-tuning \& Data Vectorization, and (2) Online QA System Process. At the Stage 1, we leverage a contrastive learning method with two negative sampling strategies to fine-tune the embedding model, and design the instruction templates to fine-tune the LLM with a Retrieval Augmented Fine-Tuning method. At the Stage 2, an efficient process of QA system is built for serving. We collect enterprise-exclusive corpora from the domain of cloud computing, and the extensive experiments show that our method achieves superior results than counterparts on two kinds of QA tasks. Our experiment also provide a case for applying the RAG4ITOps to real-world enterprise-level applications.

Akihisa Watanabe, Jiawei Ren, Li Siyao et al.

Generating physically plausible human motion is crucial for applications such as character animation and virtual reality. Existing approaches often incorporate a simulator-based motion projection layer to the diffusion process to enforce physical plausibility. However, such methods are computationally expensive due to the sequential nature of the simulator, which prevents parallelization. We show that simulator-based motion projection can be interpreted as a form of guidance, either classifier-based or classifier-free, within the diffusion process. Building on this insight, we propose SimDiff, a Simulator-constrained Diffusion Model that integrates environment parameters (e.g., gravity, wind) directly into the denoising process. By conditioning on these parameters, SimDiff generates physically plausible motions efficiently, without repeated simulator calls at inference, and also provides fine-grained control over different physical coefficients. Moreover, SimDiff successfully generalizes to unseen combinations of environmental parameters, demonstrating compositional generalization.

Jiawei Ren, Kevin Xie, Ashkan Mirzaei et al.

We present L4GM, the first 4D Large Reconstruction Model that produces animated objects from a single-view video input -- in a single feed-forward pass that takes only a second. Key to our success is a novel dataset of multiview videos containing curated, rendered animated objects from Objaverse. This dataset depicts 44K diverse objects with 110K animations rendered in 48 viewpoints, resulting in 12M videos with a total of 300M frames. We keep our L4GM simple for scalability and build directly on top of LGM, a pretrained 3D Large Reconstruction Model that outputs 3D Gaussian ellipsoids from multiview image input. L4GM outputs a per-frame 3D Gaussian Splatting representation from video frames sampled at a low fps and then upsamples the representation to a higher fps to achieve temporal smoothness. We add temporal self-attention layers to the base LGM to help it learn consistency across time, and utilize a per-timestep multiview rendering loss to train the model. The representation is upsampled to a higher framerate by training an interpolation model which produces intermediate 3D Gaussian representations. We showcase that L4GM that is only trained on synthetic data generalizes extremely well on in-the-wild videos, producing high quality animated 3D assets.

Yuan Xie, Jiawei Ren, Ji Xu

Underwater acoustic target recognition is an intractable task due to the complex acoustic source characteristics and sound propagation patterns. Limited by insufficient data and narrow information perspective, recognition models based on deep learning seem far from satisfactory in practical underwater scenarios. Although underwater acoustic signals are severely influenced by distance, channel depth, or other factors, annotations of relevant information are often non-uniform, incomplete, and hard to use. In our work, we propose to implement Underwater Acoustic Recognition based on Templates made up of rich relevant information (hereinafter called "UART"). We design templates to integrate relevant information from different perspectives into descriptive natural language. UART adopts an audio-spectrogram-text tri-modal contrastive learning framework, which endows UART with the ability to guide the learning of acoustic representations by descriptive natural language. Our experiments reveal that UART has better recognition capability and generalization performance than traditional paradigms. Furthermore, the pre-trained UART model could provide superior prior knowledge for the recognition model in the scenario without any auxiliary annotation.

Yuan Xie, Jiawei Ren, Ji Xu

Analyzing the ocean acoustic environment is a tricky task. Background noise and variable channel transmission environment make it complicated to implement accurate ship-radiated noise recognition. Existing recognition systems are weak in addressing the variable underwater environment, thus leading to disappointing performance in practical application. In order to keep the recognition system robust in various underwater environments, this work proposes an adaptive generalized recognition system - AGNet (Adaptive Generalized Network). By converting fixed wavelet parameters into fine-grained learnable parameters, AGNet learns the characteristics of underwater sound at different frequencies. Its flexible and fine-grained design is conducive to capturing more background acoustic information (e.g., background noise, underwater transmission channel). To utilize the implicit information in wavelet spectrograms, AGNet adopts the convolutional neural network with parallel convolution attention modules as the classifier. Experiments reveal that our AGNet outperforms all baseline methods on several underwater acoustic datasets, and AGNet could benefit more from transfer learning. Moreover, AGNet shows robust performance against various interference factors.

Jiawei Ren, Mingyuan Zhang, Cunjun Yu et al.

Data imbalance exists ubiquitously in real-world visual regressions, e.g., age estimation and pose estimation, hurting the model's generalizability and fairness. Thus, imbalanced regression gains increasing research attention recently. Compared to imbalanced classification, imbalanced regression focuses on continuous labels, which can be boundless and high-dimensional and hence more challenging. In this work, we identify that the widely used Mean Square Error (MSE) loss function can be ineffective in imbalanced regression. We revisit MSE from a statistical view and propose a novel loss function, Balanced MSE, to accommodate the imbalanced training label distribution. We further design multiple implementations of Balanced MSE to tackle different real-world scenarios, particularly including the one that requires no prior knowledge about the training label distribution. Moreover, to the best of our knowledge, Balanced MSE is the first general solution to high-dimensional imbalanced regression. Extensive experiments on both synthetic and three real-world benchmarks demonstrate the effectiveness of Balanced MSE.

Jiawei Ren, Liang Pan, Ziwei Liu

3D perception, especially point cloud classification, has achieved substantial progress. However, in real-world deployment, point cloud corruptions are inevitable due to the scene complexity, sensor inaccuracy, and processing imprecision. In this work, we aim to rigorously benchmark and analyze point cloud classification under corruptions. To conduct a systematic investigation, we first provide a taxonomy of common 3D corruptions and identify the atomic corruptions. Then, we perform a comprehensive evaluation on a wide range of representative point cloud models to understand their robustness and generalizability. Our benchmark results show that although point cloud classification performance improves over time, the state-of-the-art methods are on the verge of being less robust. Based on the obtained observations, we propose several effective techniques to enhance point cloud classifier robustness. We hope our comprehensive benchmark, in-depth analysis, and proposed techniques could spark future research in robust 3D perception.

Zhongang Cai, Mingyuan Zhang, Jiawei Ren et al.

Image- and video-based 3D human recovery (i.e., pose and shape estimation) have achieved substantial progress. However, due to the prohibitive cost of motion capture, existing datasets are often limited in scale and diversity. In this work, we obtain massive human sequences by playing the video game with automatically annotated 3D ground truths. Specifically, we contribute GTA-Human, a large-scale 3D human dataset generated with the GTA-V game engine, featuring a highly diverse set of subjects, actions, and scenarios. More importantly, we study the use of game-playing data and obtain five major insights. First, game-playing data is surprisingly effective. A simple frame-based baseline trained on GTA-Human outperforms more sophisticated methods by a large margin. For video-based methods, GTA-Human is even on par with the in-domain training set. Second, we discover that synthetic data provides critical complements to the real data that is typically collected indoor. Our investigation into domain gap provides explanations for our data mixture strategies that are simple yet useful. Third, the scale of the dataset matters. The performance boost is closely related to the additional data available. A systematic study reveals the model sensitivity to data density from multiple key aspects. Fourth, the effectiveness of GTA-Human is also attributed to the rich collection of strong supervision labels (SMPL parameters), which are otherwise expensive to acquire in real datasets. Fifth, the benefits of synthetic data extend to larger models such as deeper convolutional neural networks (CNNs) and Transformers, for which a significant impact is also observed. We hope our work could pave the way for scaling up 3D human recovery to the real world. Homepage: https://caizhongang.github.io/projects/GTA-Human/

Jiawei Ren, Xiao Ma, Chen Xu et al.

Person Re-Identification (Re-ID) is of great importance to the many video surveillance systems. Learning discriminative features for Re-ID remains a challenge due to the large variations in the image space, e.g., continuously changing human poses, illuminations and point of views. In this paper, we propose HAVANA, a novel extensible, light-weight HierArchical and VAriation-Normalized Autoencoder that learns features robust to intra-class variations. In contrast to existing generative approaches that prune the variations with heavy extra supervised signals, HAVANA suppresses the intra-class variations with a Variation-Normalized Autoencoder trained with no additional supervision. We also introduce a novel Jensen-Shannon triplet loss for contrastive distribution learning in Re-ID. In addition, we present Hierarchical Variation Distiller, a hierarchical VAE to factorize the latent representation and explicitly model the variations. To the best of our knowledge, HAVANA is the first VAE-based framework for person ReID.

Jiawei Ren, Cunjun Yu, Zhongang Cai et al.

Deep classifiers have achieved great success in visual recognition. However, real-world data is long-tailed by nature, leading to the mismatch between training and testing distributions. In this report, we introduce Balanced Activation (Balanced Softmax and Balanced Sigmoid), an elegant unbiased, and simple extension of Sigmoid and Softmax activation function, to accommodate the label distribution shift between training and testing in object detection. We derive the generalization bound for multiclass Softmax regression and show our loss minimizes the bound. In our experiments, we demonstrate that Balanced Activation generally provides ~3% gain in terms of mAP on LVIS-1.0 and outperforms the current state-of-the-art methods without introducing any extra parameters.

Zhongang Cai, Cunjun Yu, Junzhe Zhang et al.

We present McAssoc, a deep learning approach to the as-sociation of detection bounding boxes in different views ofa multi-camera system. The vast majority of the academiahas been developing single-camera computer vision algo-rithms, however, little research attention has been directedto incorporating them into a multi-camera system. In thispaper, we designed a 3-branch architecture that leveragesdirect association and additional cross localization infor-mation. A new metric, image-pair association accuracy(IPAA) is designed specifically for performance evaluationof cross-camera detection association. We show in the ex-periments that localization information is critical to suc-cessful cross-camera association, especially when similar-looking objects are present. This paper is an experimentalwork prior to MessyTable, which is a large-scale bench-mark for instance association in mutliple cameras.

Jiawei Ren, Cunjun Yu, Shunan Sheng et al.

Deep classifiers have achieved great success in visual recognition. However, real-world data is long-tailed by nature, leading to the mismatch between training and testing distributions. In this paper, we show that the Softmax function, though used in most classification tasks, gives a biased gradient estimation under the long-tailed setup. This paper presents Balanced Softmax, an elegant unbiased extension of Softmax, to accommodate the label distribution shift between training and testing. Theoretically, we derive the generalization bound for multiclass Softmax regression and show our loss minimizes the bound. In addition, we introduce Balanced Meta-Softmax, applying a complementary Meta Sampler to estimate the optimal class sample rate and further improve long-tailed learning. In our experiments, we demonstrate that Balanced Meta-Softmax outperforms state-of-the-art long-tailed classification solutions on both visual recognition and instance segmentation tasks.

Cunjun Yu, Xiao Ma, Jiawei Ren et al.

Understanding crowd motion dynamics is critical to real-world applications, e.g., surveillance systems and autonomous driving. This is challenging because it requires effectively modeling the socially aware crowd spatial interaction and complex temporal dependencies. We believe attention is the most important factor for trajectory prediction. In this paper, we present STAR, a Spatio-Temporal grAph tRansformer framework, which tackles trajectory prediction by only attention mechanisms. STAR models intra-graph crowd interaction by TGConv, a novel Transformer-based graph convolution mechanism. The inter-graph temporal dependencies are modeled by separate temporal Transformers. STAR captures complex spatio-temporal interactions by interleaving between spatial and temporal Transformers. To calibrate the temporal prediction for the long-lasting effect of disappeared pedestrians, we introduce a read-writable external memory module, consistently being updated by the temporal Transformer. We show that with only attention mechanism, STAR achieves state-of-the-art performance on 5 commonly used real-world pedestrian prediction datasets.