Forward-Forward Learning

Ask about this topic

Overview

Forward-Forward (FF) learning trains each layer with a local goodness score, encouraging compatible or positive examples to score above mismatched or negative examples instead of sending error gradients backward through the network. Methods include contrastive and hierarchical objectives, prototype-based classification, covariance-aware goodness, and multi-scale aggregation; researchers test them on vision benchmarks such as CIFAR and ImageNet, as well as regression tasks. The approach may reduce activation memory and could support training in hardware that cannot perform ordinary backpropagation, but removing backward passes does not automatically reduce energy or computation. Current work reports progress on larger tasks, while matched comparisons and theoretical analyses point to open challenges in accuracy, convergence, representation diversity, and scaling relative to backpropagation.

Latest papers

CardsList

17 papers2 papers in the last four weeks, against 2 the four weeks before. 0.0% of all new papers.

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

    Oct 21, 2026Yongfei Guo, Xuan Sun, Ang Li +2Neural Network OptimizationForward-Forward Learning

    Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences · University of Chinese Academy of Sciences, Beijing, China · Chinese Academy of Sciences, Beijing, China

  2. Rescuing Points Trapped Inside Incorrect Decision Regions: A target for Evaluating and Improving Forward-Forward Learning

    Oct 12, 2026Alireza Rahmansetayesh, Amirhossein Yousefiramandi, Maxwell Spiridonov +3Decision Boundary AnalysisForward-Forward Learning

    Department of Computer Science, University of California, Irvine, Irvine, CA, USA · Independent Researcher

  3. The Price of Locality: Why Forward-Forward Underperforms Backpropagation?

    Sep 27, 2026Zhaoxian Wu, Haichuan Liu, Tianyi ChenForward-Forward LearningNeural Network Optimization

    Cornell Tech and Cornell University

  4. Minimizing the Effect of Sleep Deprivation in the Forward-Forward Algorithm

    Sep 5, 2026Joy Datta, Puja Saha, Rawhatur Rabbi +4Forward-Forward Learning

    Brac University, Dhaka 1212, Bangladesh · Rajshahi University of Engineering and Technology, Rajshahi 6204, Bangladesh · Ohio University, Athens, OH 45701, United States

  5. What Does Goodness Measure? A Likelihood-Ratio Account of Forward-Forward Learning

    Jul 14, 2026Paolo GiannitrapaniMahalanobis DistanceLikelihood Ratio Testing

    Department of Information Engineering, Electronics and Telecommunications (DIET) Sapienza University of Rome Rome, Italy

  6. Forward-Only Convolutional Neural Networks with Learnable Channel-Class Assignment

    Jun 7, 2026Mohammadnavid Ghader, Saeed Reza Kheradpisheh, Bahar Farahani +1Adaptive Loss WeightingConvolutional Neural Networks

    Department of Computer and Data Science, Faculty of Mathematical Sciences, Shahid Beheshti University, Tehran, Iran · Cyberspace Research Institute, Shahid Beheshti University, Tehran, Iran · Cybersecurity and Computing Systems Research Group, University of Hertfordshire, Hatfield, United Kingdom

  7. Synthetic Benchmarks Overstate Forward-Forward Scaling: Real-Data Limits of Layer-Local Training

    Jun 4, 2026Yucheng ChenBackpropagationLocal Learning

    Amplimit Toronto, ON, Canada

  8. FFR: Forward-Forward Learning for Regression

    Jun 2, 2026Xinyang Liu, Xuanyu Liang, Shiqi Ding +4Neural Network OptimizationFeedforward Neural Networks

    University of Bristol · University College London · University of Cambridge

  9. HCL-FF: Hierarchical and Contrastive Learning for Forward-Forward Algorithm

    May 24, 2026Jie-En Yao, Hong-En Chen, C. -C. Jay KuoHierarchical Representation LearningForward-Forward Learning

    University of Southern California

  10. Adaptive Multi-Scale Goodness Aggregation for Forward-Forward Learning

    May 11, 2026Salar Beigzad, Vansh VermaNeural Network OptimizationLocal Learning

    Computer Engineering University of St. Thomas Minnesota, USA

  11. Cumulative-Goodness Free-Riding in Forward-Forward Networks: Real, Repairable, but Not Accuracy-Dominant

    May 7, 2026Amirhossein YousefiramandiNeural Network Training DynamicsLocal Learning

  12. Covariance-Aware Goodness for Scalable Forward-Forward Learning

    May 5, 2026Xiaoyi Jiang, Bashir M. Al-Hashimi, Kai XuConvolutional Neural NetworksLocal Learning

    King’s College London, United Kingdom · University of Nottingham

  13. Hyperspherical Forward-Forward with Prototypical Representations

    Apr 30, 2026Shalini Sarode, Brian Moser, Joachim Folz +4Prototype-Based LearningEfficient Neural Network Inference

    German Research Center for Artificial Intelligence (DFKI) · RPTU Kaiserslautern-Landau

  14. Selectivity and Shape in the Design of Forward-Forward Goodness Functions

    Mar 28, 2026Talha Ruzgar Akkus, Suayp Talha Kocabay, Kamer Ali Yuksel +1Forward-Forward Learning

    aiXplain, Inc., San Jose, CA

  15. Local learning for stable backpropagation-free neural network training towards physical learning

    Mar 25, 2026Yaqi Guo, Fabian Braun, Bastiaan Ketelaar +3Physical Neural NetworksNeural Network Optimization

    Department of Materials Science and Engineering, Delft University of Technology, Mekelweg 2, Delft, 2628CD, The Netherlands · Department of Precision and Microsystems Engineering, Delft University of Technology, Mekelweg 2, Delft, 2628CD, The Netherlands · Department of Intelligent Systems, Delft University of Technology, Van Mourik Broekmanweg 6, Delft 2628XE, The Netherlands

  16. When Does Margin Clamping Affect Training Variance? Dataset-Dependent Effects in Contrastive Forward-Forward Learning

    Mar 1, 2026Joshua SteierNeural Network Training DynamicsSupervised Contrastive Learning

    Independent Researcher

  17. Energy-Efficient Deep Learning Without Backpropagation: A Rigorous Hardware-Validated Benchmarking Study of Forward-Only Algorithms

    Nov 2, 2025Przemysław Spyra, Witold DzwinelEnergy-Efficient MLForward-Forward Learning

    AGH University of Krakow, Faculty of Computer Science