UV - Unmanned Vehicles

SRI to develop continuously learning AI

21st June 2018 - 12:30 GMT | by The Shephard News Team


Save this for later

SRI International has been selected by DARPA to develop a next-generation artificial intelligence (AI) system that will be able to learn continuously and apply that learning to become better and more reliable at performing new tasks, the company announced on 16 June.

Currently available AI systems are not able to handle new scenarios that they are not trained on and repeatedly make the same mistakes. Even with retraining, today’s systems are prone to 'catastrophic forgetting' when a new item disrupts previously learned knowledge.

As part of the DARPA’s Lifelong Learning Machines (L2M) programme, the SRI team will develop AI algorithms based on biological mechanisms in memory consolidation and replay.  With this AI technology, future cognitive systems such as autonomous robots may be able to continually learn after initial deployment, improving execution performance and overall safety.

Biological memory transfer is a complex sequence of dynamic processes, with local and global synchronisation patterns. These processes support memories with flexibility in expression for future thinking, foresight,

planning and creativity. 

Sek Chai, technical director in the Center for Vision Technologies, SRI International, said: ‘Our goal is to address these limitations by enabling AI systems to know what to learn and when. Since memory is a key element of cognitive function, our research is focused on understanding and applying biological memory transfers to new AI algorithms that can fundamentally improve their performance throughout their fielded lifetime experience.’
Back to News

Share to