The researchers at DeepMind have developed an algorithm to solve the problem that even though the AI could beat the games, it needed to be retrained to beat each new game, and quickly forgot how to play a game after it had moved on to the next one.
DeepMind developed EWC in order to overcome a phenomenon known as “catastrophic forgetting,” where new tasks and adaptations overwrite previously acquired knowledge and memories.The researchers explain in the paper that this usually occurs because the deep neural networks used for machine learning are typically only capable of learning multiple tasks when data is presented all at once.
While the researchers state that today’s “computer programs cannot learn from data adaptively and in real time,” they “hope that this research represents a step towards programs that can learn in a more flexible and efficient way.”
To read more http://www.theverge.com/2017/3/15/14935936/google-deepmind-ewc-memory-algorithm-atari-games
