Adaptive Learning with Binary Neurons

2009/04/29 by Juan‐Manuel Torres‐Moreno, Juan-Manuel Torres-Moreno, Torres-Moreno, Juan-Manuel +2
Computer Science · #Artificial Intelligence (cs.AI) #FOS: Computer and information sciences #Fuzzy Logic and Control Systems #Machine Learning and ELM #Neural Networks and Applications #Neural and Evolutionary Computing (cs.NE) #cs.AI #cs.NE

paper · pdf · doi:10.48550/arxiv.0904.4587

29 pages, 7 figures

arxiv created 2009/04/29 · openalex publication_date 2009/04/29 · arxiv updated 2009/12/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

A efficient incremental learning algorithm for classification tasks, called NetLines, well adapted for both binary and real-valued input patterns is presented. It generates small compact feedforward neural networks with one hidden layer of binary units and binary output units. A convergence theorem ensures that solutions with a finite number of hidden units exist for both binary and real-valued input patterns. An implementation for problems with more than two classes, valid for any binary classifier, is proposed. The generalization error and the size of the resulting networks are compared to the best published results on well-known classification benchmarks. Early stopping is shown to decrease overfitting, without improving the generalization performance.

Related

• Training binary neural networks without floating point precision
• Binary output layer of feedforward neural networks for solving multi-class classification problems
• Towards Experience Replay for Class-Incremental Learning in Fully-Binary Networks
• AdaSTE: An Adaptive Straight-Through Estimator to Train Binary Neural Networks
• Automatic Construction of Pattern Classifiers Capable of Continuous Incremental Learning and Unlearning Tasks Based on Compact-Sized Probabilistic Neural Network