The SKILL tool could have wide-ranging applications, including in the medical field, where different AI systems could specialize in learning about different illnesses and treatments and then consolidate their knowledge to serve as a comprehensive medical assistant. It could also be used to create advanced digital tour guides for smartphone users. The researchers are now looking at deploying AIs to perform more sophisticated tasks beyond simple recognition.
Key takeaways:
- A new study by USC researchers has found that robots can teach other robots how to learn by sharing knowledge.
- The researchers developed a tool called SKILL (Shared Knowledge Lifelong Learning), where AI agents learn 102 distinct tasks and then share their knowledge over a decentralized communication network.
- The SKILL tool significantly speeds up the learning process as the AI agents learn tasks in parallel, reducing the time needed by a factor of 101.5 after accounting for necessary communications and knowledge consolidation among agents.
- The researchers believe that this approach can be scaled up to thousands or millions of tasks in the future, potentially transforming various facets of our lives and bringing us closer to a truly connected, intelligent, and efficient global community.