{"title":"论学习网络的权重动态","authors":"Nahal Sharafi, Christoph Martin, Sarah Hallerberg","doi":"arxiv-2405.00743","DOIUrl":null,"url":null,"abstract":"Neural networks have become a widely adopted tool for tackling a variety of\nproblems in machine learning and artificial intelligence. In this contribution\nwe use the mathematical framework of local stability analysis to gain a deeper\nunderstanding of the learning dynamics of feed forward neural networks.\nTherefore, we derive equations for the tangent operator of the learning\ndynamics of three-layer networks learning regression tasks. The results are\nvalid for an arbitrary numbers of nodes and arbitrary choices of activation\nfunctions. Applying the results to a network learning a regression task, we\ninvestigate numerically, how stability indicators relate to the final\ntraining-loss. Although the specific results vary with different choices of\ninitial conditions and activation functions, we demonstrate that it is possible\nto predict the final training loss, by monitoring finite-time Lyapunov\nexponents or covariant Lyapunov vectors during the training process.","PeriodicalId":501167,"journal":{"name":"arXiv - PHYS - Chaotic Dynamics","volume":"60 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-04-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"On the weight dynamics of learning networks\",\"authors\":\"Nahal Sharafi, Christoph Martin, Sarah Hallerberg\",\"doi\":\"arxiv-2405.00743\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Neural networks have become a widely adopted tool for tackling a variety of\\nproblems in machine learning and artificial intelligence. In this contribution\\nwe use the mathematical framework of local stability analysis to gain a deeper\\nunderstanding of the learning dynamics of feed forward neural networks.\\nTherefore, we derive equations for the tangent operator of the learning\\ndynamics of three-layer networks learning regression tasks. The results are\\nvalid for an arbitrary numbers of nodes and arbitrary choices of activation\\nfunctions. Applying the results to a network learning a regression task, we\\ninvestigate numerically, how stability indicators relate to the final\\ntraining-loss. Although the specific results vary with different choices of\\ninitial conditions and activation functions, we demonstrate that it is possible\\nto predict the final training loss, by monitoring finite-time Lyapunov\\nexponents or covariant Lyapunov vectors during the training process.\",\"PeriodicalId\":501167,\"journal\":{\"name\":\"arXiv - PHYS - Chaotic Dynamics\",\"volume\":\"60 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-04-30\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - PHYS - Chaotic Dynamics\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2405.00743\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - PHYS - Chaotic Dynamics","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2405.00743","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Neural networks have become a widely adopted tool for tackling a variety of
problems in machine learning and artificial intelligence. In this contribution
we use the mathematical framework of local stability analysis to gain a deeper
understanding of the learning dynamics of feed forward neural networks.
Therefore, we derive equations for the tangent operator of the learning
dynamics of three-layer networks learning regression tasks. The results are
valid for an arbitrary numbers of nodes and arbitrary choices of activation
functions. Applying the results to a network learning a regression task, we
investigate numerically, how stability indicators relate to the final
training-loss. Although the specific results vary with different choices of
initial conditions and activation functions, we demonstrate that it is possible
to predict the final training loss, by monitoring finite-time Lyapunov
exponents or covariant Lyapunov vectors during the training process.