Action Perception

Action perception research focuses on understanding how agents, both biological and artificial, integrate sensory information with actions to achieve goals. Current research emphasizes the development of multimodal models, often incorporating large language models and graph neural networks, to improve perception-action cycles in complex scenarios like robotics and autonomous driving. This work is significant for advancing artificial intelligence, particularly in areas requiring robust decision-making in dynamic environments, and for providing insights into human cognition through the lens of perception-action interactions. The development of benchmarks and evaluation protocols is also a key focus, enabling more rigorous comparison and improvement of these models.