Learning Distributed Representations

Hinton's research emphasized systems that learn internal representations rather than relying entirely on hand-designed features. Distributed representations encode concepts across patterns of activity, allowing learned features to support generalization.