1939–, Cognitive scientist, neural-network theorist
Stephen Grossberg is an American cognitive scientist who, since the late 1960s, has produced a long and unified research programme on biologically-grounded neural networks. His Adaptive Resonance Theory (ART), developed with Gail Carpenter from the late 1970s onwards, addresses the stability–plasticity dilemma: how to learn new patterns without catastrophically forgetting old ones. ART networks use a feedback loop between bottom-up recognition and top-down expectation, declaring a "resonance" only when the two match closely enough.
Grossberg's work has been influential in computational neuroscience but somewhat peripheral to mainstream deep learning, which has historically prioritised gradient-based learning over the biologically-motivated learning rules ART employs. The stability–plasticity dilemma he highlighted has, however, returned to centre stage with modern interest in continual and lifelong learning.
Video
Related people: Terry Sejnowski, John Hopfield
Discussed in:
- Chapter 9: Neural Networks, Neural Networks