A major breakthrough for artificial intelligence has been discovered, thus eliminating a serious flaw that researchers have been hard at work for. Forgetfulness serves as AI's leading weakness at the moment but Google's DeepMind researchers were finally able to resolve this issue.
Deriving their inspiration from neuroscience-based theories, researchers were able to generate the first AI that learns like a human being.
— DeepMind (@DeepMindAI) March 14, 2017
“When a new task is introduced, new adaptations overwrite the knowledge that the neural network had previously acquired,” the DeepMind team explains. “This phenomenon is known in cognitive science as ‘catastrophic forgetting’, and is considered one of the fundamental limitations of neural networks.”
The AI's new algorithm was tested through Atari video games by assimilating the networks to recall old skills particularly moderating the imperative purpose of those tasks.
“We show that the learning rule can be modified so that a program can remember old tasks when learning a new one. This is an important step towards more intelligent programs that are able to learn progressively and adaptively. “Today, computer programs cannot learn from data adaptively and in real time. However, we have shown that catastrophic forgetting is not an insurmountable challenge for neural networks. We hope that this research represents a step towards programs that can learn in a more flexible and efficient way.”
Is it time to embrace artificial intelligence?
Subscribe today and get future blog posts your email.