person1967-2011

David Rumelhart

Disciplines:

Summary

American cognitive scientist who co-developed the backpropagation algorithm and the parallel distributed processing (PDP) framework — the foundation of modern deep learning.

Core Contribution

Backpropagation (1986, with McClelland and Williams): an algorithm for training multi-layer neural networks by propagating error gradients backward through the network — making deep learning feasible. Parallel Distributed Processing (PDP, 1986, 2 vols): the manifesto for connectionist cognitive science — cognition as the result of interactions among large numbers of simple units.

Key Spikes & Highlights

The PDP volumes were the intellectual foundation of the connectionist turn in cognitive science — they argued that traditional symbolic AI had the wrong architecture, and that the right model was neural networks with distributed representations.
Developed early-onset frontotemporal dementia in the 1990s and died without knowing that his work on backpropagation would become the foundation of the deep learning revolution.

Notable Works & Achievements

Parallel Distributed Processing

Learning Representations by Back-propagating Errors

Historical Influences

Warren McCulloch Frank Rosenblatt

Intellectual Lineage

Cognitive Science Connectionism

Contemporaries & Peers

Geoffrey Hinton James McClelland

Minds Influenced

Geoffrey Hinton Deep Learning Research

Intellectual Posterity

Deep Learning Connectionism Backpropagation

Discussing

David Rumelhart

Ask anything about David Rumelhart — their ideas, influences, contradictions, legacy.

person1967-2011

David Rumelhart

Disciplines:

psychology

cognitive science

artificial intelligence

Summary

American cognitive scientist who co-developed the backpropagation algorithm and the parallel distributed processing (PDP) framework — the foundation of modern deep learning.

Core Contribution

Key Spikes & Highlights

The PDP volumes were the intellectual foundation of the connectionist turn in cognitive science — they argued that traditional symbolic AI had the wrong architecture, and that the right model was neural networks with distributed representations.
Developed early-onset frontotemporal dementia in the 1990s and died without knowing that his work on backpropagation would become the foundation of the deep learning revolution.

Notable Works & Achievements

Parallel Distributed Processing

Learning Representations by Back-propagating Errors

Historical Influences

Warren McCulloch Frank Rosenblatt

Intellectual Lineage

Cognitive Science Connectionism

Contemporaries & Peers

Geoffrey Hinton James McClelland

Minds Influenced

Geoffrey Hinton Deep Learning Research

Intellectual Posterity

Deep Learning Connectionism Backpropagation

Discussing

David Rumelhart

Ask anything about David Rumelhart — their ideas, influences, contradictions, legacy.