Alexander Dack, Benjamin Qureshi, Thomas E. Ouldridge, Tomislav Plesa
{"title":"近似任意动力学的递归神经化学反应网络","authors":"Alexander Dack, Benjamin Qureshi, Thomas E. Ouldridge, Tomislav Plesa","doi":"arxiv-2406.03456","DOIUrl":null,"url":null,"abstract":"Many important phenomena in chemistry and biology are realized via dynamical\nfeatures such as multi-stability, oscillations, and chaos. Construction of\nnovel chemical systems with such finely-tuned dynamics is a challenging problem\ncentral to the growing field of synthetic biology. In this paper, we address\nthis problem by putting forward a molecular version of a recurrent artificial\nneural network, which we call a recurrent neural chemical reaction network\n(RNCRN). We prove that the RNCRN, with sufficiently many auxiliary chemical\nspecies and suitable fast reactions, can be systematically trained to achieve\nany dynamics. This approximation ability is shown to hold independent of the\ninitial conditions for the auxiliary species, making the RNCRN more\nexperimentally feasible. To demonstrate the results, we present a number of\nrelatively simple RNCRNs trained to display a variety of biologically-important\ndynamical features.","PeriodicalId":501325,"journal":{"name":"arXiv - QuanBio - Molecular Networks","volume":"25 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Recurrent neural chemical reaction networks that approximate arbitrary dynamics\",\"authors\":\"Alexander Dack, Benjamin Qureshi, Thomas E. Ouldridge, Tomislav Plesa\",\"doi\":\"arxiv-2406.03456\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Many important phenomena in chemistry and biology are realized via dynamical\\nfeatures such as multi-stability, oscillations, and chaos. Construction of\\nnovel chemical systems with such finely-tuned dynamics is a challenging problem\\ncentral to the growing field of synthetic biology. In this paper, we address\\nthis problem by putting forward a molecular version of a recurrent artificial\\nneural network, which we call a recurrent neural chemical reaction network\\n(RNCRN). We prove that the RNCRN, with sufficiently many auxiliary chemical\\nspecies and suitable fast reactions, can be systematically trained to achieve\\nany dynamics. This approximation ability is shown to hold independent of the\\ninitial conditions for the auxiliary species, making the RNCRN more\\nexperimentally feasible. To demonstrate the results, we present a number of\\nrelatively simple RNCRNs trained to display a variety of biologically-important\\ndynamical features.\",\"PeriodicalId\":501325,\"journal\":{\"name\":\"arXiv - QuanBio - Molecular Networks\",\"volume\":\"25 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-06-05\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - QuanBio - Molecular Networks\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2406.03456\",\"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 - QuanBio - Molecular Networks","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2406.03456","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Recurrent neural chemical reaction networks that approximate arbitrary dynamics
Many important phenomena in chemistry and biology are realized via dynamical
features such as multi-stability, oscillations, and chaos. Construction of
novel chemical systems with such finely-tuned dynamics is a challenging problem
central to the growing field of synthetic biology. In this paper, we address
this problem by putting forward a molecular version of a recurrent artificial
neural network, which we call a recurrent neural chemical reaction network
(RNCRN). We prove that the RNCRN, with sufficiently many auxiliary chemical
species and suitable fast reactions, can be systematically trained to achieve
any dynamics. This approximation ability is shown to hold independent of the
initial conditions for the auxiliary species, making the RNCRN more
experimentally feasible. To demonstrate the results, we present a number of
relatively simple RNCRNs trained to display a variety of biologically-important
dynamical features.