The Forward-Forward algorithm (FF) is comparable in speed to backpropagation but has the advantage that it can be used when the precise details of the forward computation are unknown.

In early December, dozens of alternatives to traditional backpropagation were proposed during a workshop at the NeurIPS 2020 conference, which took place virtually.

Beijing, April 23, 2025 (GLOBE NEWSWIRE) -- WiMi Developed a Quantum Computing-Based Feedforward Neural Network (QFNN) Algorithm BEIJING, Apr. 23, 2025––WiMi Hologram Cloud Inc. (NASDAQ: WIMI ...

Backpropagation is not limited to function derivatives. Any algorithm that effectively takes the loss function and applies gradual, positive changes back through the network is valid. Conclusion ...

The researchers showed the gradient could be found experimentally using a 3D-printed model of a mechanical neural network. Credit S. Li and X. Mao, Nat. Commun. 2024, DOI: 10.1038/s41467-024-54849-z) ...