Inspired by sea slugs and octopuses, researchers have identified a new approach which uses simple associative learning rules and enhanced episodic memory.Scientists have discovered an AI that can help with navigating new environments, look out for rewards, and map landmarks. It’s also equipped to overcome obstacles. The team achieved this by exposing artificial intelligence to simple associative learning rules.
The researchers believe that this method is far more efficient and generates vast amounts of data compared to the others. Simply adding a memory module can help the AI retain information about the past. This approach could also expand simple spatial learning to complex learning. Asimov was programmed to monitor its internal state and seek validation and rewards. It was trained to live in the present. It could also learn from experience but had difficulties integrating information from the past.