Liming Chen

Self-training could be misleadingly perceived as beneficial, yet it often degrades model performance on tasks the base model already solves well.

Achieving state-of-the-art grasp synthesis without the need for expensive, object-annotated datasets could revolutionize dexterous manipulation in robotics.

Object-agnostic grasp planning can achieve state-of-the-art performance without the need for object-specific training data, revolutionizing how robots learn to manipulate diverse objects.

Slot-RAE achieves state-of-the-art object-centric learning without the cumbersome reliance on pretrained generative models, streamlining the process significantly.

PriGo enhances robotic manipulation by refining actions in real-time, leading to improved robustness and generalization without retraining.

Achieving zero-shot sim-to-real transfer, the CORE Planner reduces travel distance by up to 48% compared to existing learning-based methods, revolutionizing robot navigation in unknown environments.

LAFM boosts robotic manipulation success rates by over 23% by dynamically adapting to the complexities of action spaces.

Open-source Mixture-of-Experts models now rival proprietary systems in fake news detection, but LiveFact reveals a critical "reasoning gap" where all models struggle with the temporal uncertainty inherent in real-world misinformation.

By decoupling camera and manipulation actions and training them in a coordinated manner, SaPaVe achieves significantly higher success rates in real-world robotic manipulation tasks compared to existing end-to-end vision-language-action models.