{"title":"Neurons as Autoencoders.","authors":"Larry Bull","doi":"10.1162/artl_c_00461","DOIUrl":null,"url":null,"abstract":"<p><p>This letter presents the idea that neural backpropagation is exploiting dendritic processing to enable individual neurons to perform autoencoding. Using a very simple connection weight search heuristic and artificial neural network model, the effects of interleaving autoencoding for each neuron in a hidden layer of a feedforward network are explored. This is contrasted with the equivalent standard layered approach to autoencoding. It is shown that such individualized processing is not detrimental and can improve network learning.</p>","PeriodicalId":55574,"journal":{"name":"Artificial Life","volume":" ","pages":"1-6"},"PeriodicalIF":1.6000,"publicationDate":"2024-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Artificial Life","FirstCategoryId":"94","ListUrlMain":"https://doi.org/10.1162/artl_c_00461","RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE","Score":null,"Total":0}
引用次数: 0
Abstract
This letter presents the idea that neural backpropagation is exploiting dendritic processing to enable individual neurons to perform autoencoding. Using a very simple connection weight search heuristic and artificial neural network model, the effects of interleaving autoencoding for each neuron in a hidden layer of a feedforward network are explored. This is contrasted with the equivalent standard layered approach to autoencoding. It is shown that such individualized processing is not detrimental and can improve network learning.
期刊介绍:
Artificial Life, launched in the fall of 1993, has become the unifying forum for the exchange of scientific information on the study of artificial systems that exhibit the behavioral characteristics of natural living systems, through the synthesis or simulation using computational (software), robotic (hardware), and/or physicochemical (wetware) means. Each issue features cutting-edge research on artificial life that advances the state-of-the-art of our knowledge about various aspects of living systems such as:
Artificial chemistry and the origins of life
Self-assembly, growth, and development
Self-replication and self-repair
Systems and synthetic biology
Perception, cognition, and behavior
Embodiment and enactivism
Collective behaviors of swarms
Evolutionary and ecological dynamics
Open-endedness and creativity
Social organization and cultural evolution
Societal and technological implications
Philosophy and aesthetics
Applications to biology, medicine, business, education, or entertainment.