Newest papers

Forward-Only Learning for Shared-Kernel Wavelet Neural Networks without Backpropagation

Yongfei Guo, Xuan Sun, Ang Li, Zhao Zhang, Yuanhao Gong · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.29697 · cited by 0 works (OpenAlex) · open-access PDF

French

Bien que l'algorithme Avance-Avance évitera la propagation conventionnelle, il demeure principalement local ou stratégique, motivant l'apprentissage global de l'ensemble du réseau en utilisant uniquement des évaluations avancées. Pour les réseaux neuronaux de wavelet de noyau partagé, nous proposons donc un cadre d'apprentissage global basé sur le CMA-ES. Nous introduisons en outre la décomposition des paramètres conscients de la structure et l'attribution de crédit spécifique aux canaux, facturant la recherche en trois sous-espaces. Sur BSDS500, notre méthode de test réduit le MSE de 4.1783 à 4.0130 et améliore le PSNR de 42.621 à 42.787 dB~B sur Adam/B

Machine translation (Meta M2M100); the English original follows.

Abstract

Backpropagation remains the dominant training paradigm, using backward gradients and local updates that can suffer from vanishing or exploding gradients in deep networks. Although the Forward-Forward algorithm avoids conventional backpropagation, it remains mainly local or layer-wise, motivating global whole-network learning using forward evaluations only. For shared-kernel wavelet neural networks, we therefore propose a CMA-ES-based global forward-only learning framework. We further introduce structure-aware parameter decomposition and channel-specific credit assignment, factorizing the search into three subspaces. On BSDS500, our method reduces test MSE from 4.1783 to 4.0130 and improves PSNR from 42.621 to 42.787~dB over Adam/BP without backward propagation. Beyond these gains, training requires only 6 training-set coverages, compared with 1500 epochs for BP. Ablations show that structural decomposition and aligned credit assignment are more important than increasing population size, demonstrating the feasibility of whole-network forward-only learning for compact reconstruction networks. These results suggest a potential route toward forward-only optimization of compact reconstruction models for resource-constrained imaging and edge or near-sensor systems.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Backpropagation, Artificial intelligence, Artificial neural network, Computer science, Wavelet, Pattern recognition (psychology)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange