Yijie Su, Yakun Zhu, Yi Bin et al.cs.CV
Language-guided 3D affordance grounding aims to localize functional regions on 3D objects according to interaction intents. However, existing methods exhibit poor generalization to unseen objects.
Qixin Sun, WenLong Gao, Xiansheng Chen et al.cs.RO
Robotic manipulation often requires remembering interactions that are no longer visible. We introduce Latent Transition Memory (LTM), which stores history as compact latent transitions between consecutive observations.
Jing Xia, Changyan Li, Yuduo Wang et al.cs.RO
Indoor service robots often enter unmapped buildings whose layouts and task constraints are already familiar to users.
Zhiyuan Ma, Jiabao Wei, Tianyu Wang et al.cs.AI
Diffusion models (DMs) have steadily demonstrated outstanding performance in high-fidelity, controllable generation across diverse scenarios, following the "pretraining-finetuning" paradigm established in LLMs.
Zhuowen Liang, Xiaotian Lin, Yuyu Luo et al.cs.AI
Multimodal long-document question answering requires models to seek sparse evidence across pages and modalities, understand heterogeneous document content, and reason over the collected evidence.
Jaeyeong Kim, Jinhyuk Jang, Jongmin Lee et al.cs.CV
We propose Tetris3D, a generative framework for single-image 3D scene reconstruction that recovers objects which are physically and geometrically coherent as a scene.
Yuling Xi, Haokai Zhang, Muzhi Zhu et al.cs.CV
Metric reasoning is a critical and challenging task for Vision Language Models (VLMs), playing a pivotal role in embodied AI tasks such as robotic manipulation and autonomous navigation.
Tianjiao Yu, Zhilin Ou, Yifan Shen et al.cs.RO
Generative simulation can automate robot environment construction, but physically valid environments are not necessarily useful for training. The challenge is to identify environments that address a policy's failures and supervision that helps it overcome them.
Shuanglin Li, Jiaying Lin, Shuquan Yecs.CV
Recent advances in multimodal large language models (MLLMs) and 3D generation have expanded 3D understanding and synthesis. However, beyond visual fidelity, task-specific correctness grounded in 3D evidence remains a major challenge.
Qi Lü, Jiarun Li, Ce Zhang et al.cs.CV
Vision-language-action (VLA) models transfer semantic knowledge acquired through vision-language pretraining to robot control, achieving strong performance across various benchmarks.
Zirui Wu, Chen Guo, Tianjian Jiang et al.cs.CV
We introduce P^3-WAM, a unified world action model for egocentric dexterous manipulation that learns 3D perception, action-conditioned prediction, and goal-directed planning through three complementary modes: Perceive, which reconstructs articulated hand motion and camera trajectories from video; Predict, which ge…
Han Su, Meng Tian, Tianyu Huang et al.cs.CV
Vision-language models still struggle to understand 3D spatial relations across views and update spatial judgments as the observer moves.
Yun Zhang, Zhiyu Huang, Tianhui Cai et al.cs.RO
Long-horizon robotic manipulation fundamentally requires memory: agents must retain task-relevant information from past observations and interactions and use it to guide future actions.
Jiuyi Xu, Chengyuan Qian, Zhengzhong Tu et al.cs.RO
Vision-Language-Action (VLA) models present a promising paradigm for general-purpose robot manipulation, but closed-loop deployment is often limited by inference efficiency: at each timestep, high-resolution, often multi-view images are encoded into hundreds of visual tokens that dominate transformer computation and l…
Tianxing Chen, Wenwei Lin, Chuqiao Lyu et al.cs.RO
Generalist robot policies benefit from large-scale visual pretraining but often struggle with contact-rich dexterous manipulation, where force, compliance, and incipient slip are hard to observe from images.
Siwei Cai, Yuwei Wu, Vijay Kumar et al.cs.RO
Aerial robots operating near people must treat a pedestrian, a parked car, and a wall differently, yet conventional perception--planning stacks reduce all of them to occupied geometry---a \emph{human-aware safety gap} in which category-specific clearance is unavailable to the planner and temporary occlusions remove pe…
Luozheng Qin, Yuhan Zhu, Changhai Zhou et al.cs.CV
Spatial intelligence lets an agent perceive and act in the physical world, and many of its judgments must be made at interaction rates: which spatial relation holds, how far an object is, or where to place it.
Juliette Marrie, Paul Pacaud, Adrien Ramanana Rahary et al.cs.RO
We introduce REKI, a 2.4B vision-language-action model (VLA) pretrained exclusively on public data that sets a new Pareto frontier between performance and model size: on LIBERO-Plus, RoboTwin Clean2Rand and RoboCasa365, it is outperformed only by models about twice its size.
20 papers listed