Fruit Fly-Inspired Algorithm Learns Fast Without Forgetting Old Data

Researchers have developed a new algorithm modeled on the olfactory system of fruit flies that can learn new information rapidly without losing previously acquired knowledge. The approach uses a technique called 'sparse coding,' which mimics how insects process scent signals in their brains. A key challenge in machine learning known as 'catastrophic forgetting' — where systems lose old data when learning new data — is addressed by this method. By drawing on biological neural mechanisms, the algorithm achieves efficient, continual learning that standard AI models typically struggle with.