# Hebbian Deep Learning Without Feedback Source: [Hebbian Deep Learning Without Feedback]() Collection index: [llms.txt]() Data updated: 2025-10-25T19:13:45.935Z Authors: Adrien Journ\\'e, Hector Garcia Rodriguez, Qinghai Guo, Timoleon Moraitis Published: 2022-09-23T23:12:59.000Z ## Abstract Recent approximations to backpropagation (BP) have mitigated many of BP's computational inefficiencies and incompatibilities with biology, but important limitations still remain. Moreover, the approximations significantly decrease accuracy in benchmarks, suggesting that an entirely different approach may be more fruitful. Here, grounded on recent theory for Hebbian learning in soft winner-take-all networks, we present multilayer SoftHebb, i.e. an algorithm that trains deep neural networks, without any feedback, target, or error signals. As a result, it achieves efficiency by avoiding weight transport, non-local plasticity, time-locking of layer updates, iterative equilibria, and (self-) supervisory or other feedback signals -- which were necessary in other approaches. Its increased efficiency and biological compatibility do not trade off accuracy compared to state-of-the-art bio-plausible learning, but rather improve it. With up to five hidden layers and an added linear classifier, accuracies on MNIST, CIFAR-10, STL-10, and ImageNet, respectively reach 99.4%, 80.3%, 76.2%, and 27.3%. In conclusion, SoftHebb shows with a radically different approach from BP that Deep Learning over few layers may be plausible in the brain and increases the accuracy of bio-plausible machine learning. Code is available at https://github.com/NeuromorphicComputing/SoftHebb. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Image Classification on CIFAR-10 (test) (Accuracy and Backprop Difference)](): Image Classification; CIFAR-10 (test); Accuracy: 80.36. - [Image Classification on MNIST (test) (Accuracy Only)](): Image Classification; MNIST (test); Accuracy: 99.35. - [Image Classification on CIFAR-100 (ACC, NLL, ECE, ESS)](): Image Classification; CIFAR-100; Accuracy: 56. - [Image Classification on ImageNet ILSVRC-2012 (val)](): Image Classification; ImageNet ILSVRC-2012 (val); Top-1 Accuracy: 26.08. - [Image Classification on CIFAR-10 (Accuracy)](): Image Classification; CIFAR-10; Accuracy: 80.31. - [Image Classification on ImageNet-1K (test) (Standard Accuracy)](): Image Classification; ImageNet 1k (test); Top-1 Accuracy: 27.83. ## Original sources and implementation [Original paper]() [PDF]()