10 papers
Affordance-Conditioned Decision Making: Bridging the Semantic-Spatial Gap in Zero-Shot Cross-Floor Vision-and-Language Navigation
Xuekang Yang, Lu Chen, Shuang Luo +4
Vision-and-language navigation increasingly relies on general-purpose semantic planners, yet translating correct high-level intent into reliable physical execution remains difficul…
MotionJEPA: Preventing Temporal Feature Collapse by Capturing Visual Changes in Latent Space
Markus Karmann, Shile Li, Christian Internò +10
Joint Embedding Predictive Architectures (JEPAs) are a promising paradigm for learning task-agnostic latent world models without visual reconstruction. However, standard JEPA train…
A4A: Cross-Embodiment Transfer of Action-Oriented 4D Affordances from Human Demonstrations
Yifan Han, Litao Liu, Yuqi Gu +7
Human demonstrations contain rich manipulation knowledge, but it remains unclear what information can be transferred effectively to robot control. Existing affordance representatio…
HELP: Human-Efficient Large-Scale Robot Post-Training with Rollout Segmentation
Shaopeng Zhai, Qi Zhang, Tianyi Zhang +5
When adapting Vision Language Action (VLA) models to downstream tasks, multiple rounds of post-training are often required to progressively address policy weaknesses. In this repor…
Affordance2Action: Task-Conditioned Scene-level Affordance Grounding for Real-Time Manipulation
Litao Liu, Yifan Han, Pengfei Yi +9
Task-conditioned manipulation requires grounding instructions to task-relevant functional parts rather than object categories. This setting is scene-dependent and often one-to-many…
Preserving Foundational Capabilities in Flow-Matching VLAs through Conservative SFT
Tianyi Zhang, Shaopeng Zhai, Haoran Zhang +2
Unconstrained fine-tuning of flow-matching Vision-Language-Action (VLA) models drives dense parameter overwrites, degrading pre-trained capabilities. We present Conservative Superv…
Neural Video Compression with Domain Transfer
Tiange Zhang, Rongqun Lin, Xiandong Meng +4
Content-adaptive compression has always been a key direction in neural video coding (NVC), aiming to mitigate the domain gap between training and testing data. Such gaps often aris…
A Vision-Language-Action-Critic Model for Robotic Real-World Reinforcement Learning
Shaopeng Zhai, Qi Zhang, Tianyi Zhang +7
Robotic real-world reinforcement learning (RL) with vision-language-action (VLA) models is bottlenecked by sparse, handcrafted rewards and inefficient exploration. We introduce VLA…
CLSP: High-Fidelity Contrastive Language-State Pre-training for Agent State Representation
Fuxian Huang, Qi Zhang, Shaopeng Zhai +6
With the rapid development of artificial intelligence, multimodal learning has become an important research area. For intelligent agents, the state is a crucial modality to convey…
Building Open-Ended Embodied Agent via Language-Policy Bidirectional Adaptation
Shaopeng Zhai, Jie Wang, Tianyi Zhang +6
Building embodied agents on integrating Large Language Models (LLMs) and Reinforcement Learning (RL) have revolutionized human-AI interaction: researchers can now leverage language…