相关论文: DISC: Decoupling Instruction from State-Conditione…

Reward hacking arises when a model improves a proxy reward by exploiting shortcuts rather than solving the intended task. We study this failure mode through the geometry of reinforcement learning updates in language models and argue that…

Learning visuomotor policy for multi-task robotic manipulation has been a long-standing challenge for the robotics community. The difficulty lies in the diversity of action space: typically, a goal can be accomplished in multiple ways,…

Even though most interfaces in the real world are discrete, no efficient way exists to train neural networks to make use of them, yet. We enhance an Interaction Network (a Reinforcement Learning architecture) with discrete interfaces and…

Text-to-image diffusion models have advanced towards more controllable generation via supporting various additional conditions (e.g.,depth map, bounding box) beyond text. However, these models are learned based on the premise of perfect…

Well-designed dense reward functions in robot manipulation not only indicate whether a task is completed but also encode progress along the way. Generally, designing dense rewards is challenging and usually requires access to privileged…

Data is often impractical to share for a range of well considered reasons, such as concerns over privacy, intellectual property, and legal constraints. This not only fragments the statistical power of predictive models, but creates an…

While solving complex manipulation tasks, manipulation policies often need to learn a set of diverse skills to accomplish these tasks. The set of skills is often quite multimodal - each one may have a quite distinct distribution of actions…

We introduce DIP, a novel unsupervised post-training method designed to enhance dense image representations in large-scale pretrained vision encoders for in-context scene understanding. Unlike prior approaches that rely on complex…

Synthesizing natural human motion that adapts to complex environments while allowing creative control remains a fundamental challenge in motion synthesis. Existing models often fall short, either by assuming flat terrain or lacking the…

Conditional diffusion models have made impressive progress in the field of image processing, but the characteristics of constructing data distribution pathways make it difficult to exploit the intrinsic correlation between tasks in…

Discrete diffusion models are a powerful, emerging paradigm for code generation. They construct programs through iterative refinement of partially corrupted token sequences and enable parallel token refinement. Importantly, this paradigm…

Noninterference offers a rigorous end-to-end guarantee for secure propagation of information. However, real-world systems almost always involve security requirements that change during program execution, making noninterference inapplicable.…

Deep neural networks often rely on spurious correlations to make predictions, which hinders generalization beyond training environments. For instance, models that associate cats with bed backgrounds can fail to predict the existence of cats…

Training general robotic policies from heterogeneous data for different tasks is a significant challenge. Existing robotic datasets vary in different modalities such as color, depth, tactile, and proprioceptive information, and collected in…

Can we use reinforcement learning to learn general-purpose policies that can perform a wide range of different tasks, resulting in flexible and reusable skills? Contextual policies provide this capability in principle, but the…

Real-world control systems require policies that are not only high-performing but also interpretable and robust. A promising direction toward this goal is model-based control, which learns system dynamics and cost functions from historical…

Diffusion models have seen rapid adoption in robotic imitation learning, enabling autonomous execution of complex dexterous tasks. However, action synthesis is often slow, requiring many steps of iterative denoising, limiting the extent to…

Masked diffusion models (MDMs) have recently emerged as a novel framework for language modeling. MDMs generate sentences by iteratively denoising masked sequences, filling in [MASK] tokens step by step. Although MDMs support any-order…

Learning generalizable policies for robotic manipulation increasingly relies on large-scale models that map language instructions to actions (L2A). However, this one-way paradigm often produces policies that execute tasks without deeper…

Unlike quasi-static robotic manipulation tasks like pick-and-place, dynamic tasks such as non-prehensile manipulation pose greater challenges, especially for vision-based control. Successful control requires the extraction of features…