AI is going to make our robots a lot smarter. With large language models, it is possible to perform iterative self-improvement for them. For example, it is possible to make a table tennis robot smarter. SAS Prompt, which stands for summarize, analyze, and synthetize, is an approach that enables robots to exhibit new, unseen behavior. As the researchers explain, it is:
a single prompt that enables iterative learning and adaptation of robot behavior by combining the LLM’s ability to retrieve, reason and optimize over previous robot traces in order to synthesize new, unseen behavior.
As the above video shows, this approach was tested in a robot table tennis robot.
[HT]