Unverified50% confidenceFactExact time
The Forward-Forward algorithm replaces one forward pass plus one backward pass with two forward passes: the first runs on positive samples to increase the goodness of each layer, and the second runs on negative samples to decrease goodness
1
Sources
50%
Confidence
Long-term
Relevance
10/5/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedSparse rewards combined with long time spans cause gradient-based optimization signals to decay significantly during backpropagation.68% similarUnverifiedFor candidates below the truncation threshold, the algorithm focuses on improving their reward; for candidates above the threshold, it focuses on maintaining diversity among them.67% similarUnverifiedReverse mode automatic differentiation can obtain gradients of the loss with respect to all parameters in a single backward pass, with complexity only a constant multiple of the forward computation (typically 2–4×).67% similarUnverifiedIn practice, it was found that even when only one gradient update is performed on the target per step—yielding a noisy estimate of the optimal transport map—the algorithm is still able to converge.66% similarUnverifiedMany algorithms that perform well in reversible environments may suffer a cliff-like performance drop once they approach the irreversibility threshold.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/978864API
curl https://kongchang.com/api/v1/knowledge/claims/978864MCP
get_claim(id=978864)