Achieving Occam's Razor: Deep Learning for Optimal Model Reduction. Antal, B. B., Chesebro, A. G., Strey, H. H., Mujica-Parodi, L. R., & Weistuch, C. September, 2023. arXiv:2303.13746![Achieving Occam's Razor: Deep Learning for Optimal Model Reduction [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper doi abstract bibtex 4 downloads All fields of science depend on mathematical models. Occam's razor refers to the principle that good models should exclude parameters beyond those minimally required to describe the systems they represent. This is because redundancy can lead to incorrect estimates of model parameters from data, and thus inaccurate or ambiguous conclusions. Here, we show how deep learning can be powerfully leveraged to address Occam's razor. FixFit, our new method, uses a feedforward deep neural network with a bottleneck layer to characterize and predict the behavior of a given model from its input parameters. FixFit has three major benefits. First, it provides a metric to quantify the original model's degree of complexity. Second, it allows for the unique fitting of data. Third, it provides an unbiased way to discriminate between experimental hypotheses that add value versus those that do not. In two use cases, we demonstrate the broad applicability of this method across scientific domains. To validate the method using a known system, we apply FixFit to recover known composite parameters for the Kepler orbit model. To illustrate how the method can be applied to less well-established fields, we use it to identify parameters for a multi-scale brain model and reduce the search space for viable candidate mechanisms.
@misc{antal2023,
title = {Achieving {Occam}'s {Razor}: {Deep} {Learning} for {Optimal} {Model} {Reduction}},
shorttitle = {Achieving {Occam}'s {Razor}},
url = {http://arxiv.org/abs/2303.13746},
doi = {10.48550/arXiv.2303.13746},
abstract = {All fields of science depend on mathematical models. Occam's razor refers to the principle that good models should exclude parameters beyond those minimally required to describe the systems they represent. This is because redundancy can lead to incorrect estimates of model parameters from data, and thus inaccurate or ambiguous conclusions. Here, we show how deep learning can be powerfully leveraged to address Occam's razor. FixFit, our new method, uses a feedforward deep neural network with a bottleneck layer to characterize and predict the behavior of a given model from its input parameters. FixFit has three major benefits. First, it provides a metric to quantify the original model's degree of complexity. Second, it allows for the unique fitting of data. Third, it provides an unbiased way to discriminate between experimental hypotheses that add value versus those that do not. In two use cases, we demonstrate the broad applicability of this method across scientific domains. To validate the method using a known system, we apply FixFit to recover known composite parameters for the Kepler orbit model. To illustrate how the method can be applied to less well-established fields, we use it to identify parameters for a multi-scale brain model and reduce the search space for viable candidate mechanisms.},
urldate = {2024-04-04},
publisher = {arXiv},
author = {Antal, Botond B. and Chesebro, Anthony G. and Strey, Helmut H. and Mujica-Parodi, Lilianne R. and Weistuch, Corey},
month = sep,
year = {2023},
note = {arXiv:2303.13746},
keywords = {Computer Science - Machine Learning, Mathematics - Optimization and Control},
file = {arXiv Fulltext PDF:/Users/lcneuro/Zotero/storage/VGJNWH4N/Antal et al. - 2023 - Achieving Occam's Razor Deep Learning for Optimal.pdf:application/pdf;arXiv.org Snapshot:/Users/lcneuro/Zotero/storage/SCTLXL82/2303.html:text/html},
}
Downloads: 4
{"_id":"GZP9RRMJQw8sYYL97","bibbaseid":"antal-chesebro-strey-mujicaparodi-weistuch-achievingoccamsrazordeeplearningforoptimalmodelreduction-2023","author_short":["Antal, B. B.","Chesebro, A. G.","Strey, H. H.","Mujica-Parodi, L. R.","Weistuch, C."],"bibdata":{"bibtype":"misc","type":"misc","title":"Achieving Occam's Razor: Deep Learning for Optimal Model Reduction","shorttitle":"Achieving Occam's Razor","url":"http://arxiv.org/abs/2303.13746","doi":"10.48550/arXiv.2303.13746","abstract":"All fields of science depend on mathematical models. Occam's razor refers to the principle that good models should exclude parameters beyond those minimally required to describe the systems they represent. This is because redundancy can lead to incorrect estimates of model parameters from data, and thus inaccurate or ambiguous conclusions. Here, we show how deep learning can be powerfully leveraged to address Occam's razor. FixFit, our new method, uses a feedforward deep neural network with a bottleneck layer to characterize and predict the behavior of a given model from its input parameters. FixFit has three major benefits. First, it provides a metric to quantify the original model's degree of complexity. Second, it allows for the unique fitting of data. Third, it provides an unbiased way to discriminate between experimental hypotheses that add value versus those that do not. In two use cases, we demonstrate the broad applicability of this method across scientific domains. To validate the method using a known system, we apply FixFit to recover known composite parameters for the Kepler orbit model. To illustrate how the method can be applied to less well-established fields, we use it to identify parameters for a multi-scale brain model and reduce the search space for viable candidate mechanisms.","urldate":"2024-04-04","publisher":"arXiv","author":[{"propositions":[],"lastnames":["Antal"],"firstnames":["Botond","B."],"suffixes":[]},{"propositions":[],"lastnames":["Chesebro"],"firstnames":["Anthony","G."],"suffixes":[]},{"propositions":[],"lastnames":["Strey"],"firstnames":["Helmut","H."],"suffixes":[]},{"propositions":[],"lastnames":["Mujica-Parodi"],"firstnames":["Lilianne","R."],"suffixes":[]},{"propositions":[],"lastnames":["Weistuch"],"firstnames":["Corey"],"suffixes":[]}],"month":"September","year":"2023","note":"arXiv:2303.13746","keywords":"Computer Science - Machine Learning, Mathematics - Optimization and Control","file":"arXiv Fulltext PDF:/Users/lcneuro/Zotero/storage/VGJNWH4N/Antal et al. - 2023 - Achieving Occam's Razor Deep Learning for Optimal.pdf:application/pdf;arXiv.org Snapshot:/Users/lcneuro/Zotero/storage/SCTLXL82/2303.html:text/html","bibtex":"@misc{antal2023,\n\ttitle = {Achieving {Occam}'s {Razor}: {Deep} {Learning} for {Optimal} {Model} {Reduction}},\n\tshorttitle = {Achieving {Occam}'s {Razor}},\n\turl = {http://arxiv.org/abs/2303.13746},\n\tdoi = {10.48550/arXiv.2303.13746},\n\tabstract = {All fields of science depend on mathematical models. Occam's razor refers to the principle that good models should exclude parameters beyond those minimally required to describe the systems they represent. This is because redundancy can lead to incorrect estimates of model parameters from data, and thus inaccurate or ambiguous conclusions. Here, we show how deep learning can be powerfully leveraged to address Occam's razor. FixFit, our new method, uses a feedforward deep neural network with a bottleneck layer to characterize and predict the behavior of a given model from its input parameters. FixFit has three major benefits. First, it provides a metric to quantify the original model's degree of complexity. Second, it allows for the unique fitting of data. Third, it provides an unbiased way to discriminate between experimental hypotheses that add value versus those that do not. In two use cases, we demonstrate the broad applicability of this method across scientific domains. To validate the method using a known system, we apply FixFit to recover known composite parameters for the Kepler orbit model. To illustrate how the method can be applied to less well-established fields, we use it to identify parameters for a multi-scale brain model and reduce the search space for viable candidate mechanisms.},\n\turldate = {2024-04-04},\n\tpublisher = {arXiv},\n\tauthor = {Antal, Botond B. and Chesebro, Anthony G. and Strey, Helmut H. and Mujica-Parodi, Lilianne R. and Weistuch, Corey},\n\tmonth = sep,\n\tyear = {2023},\n\tnote = {arXiv:2303.13746},\n\tkeywords = {Computer Science - Machine Learning, Mathematics - Optimization and Control},\n\tfile = {arXiv Fulltext PDF:/Users/lcneuro/Zotero/storage/VGJNWH4N/Antal et al. - 2023 - Achieving Occam's Razor Deep Learning for Optimal.pdf:application/pdf;arXiv.org Snapshot:/Users/lcneuro/Zotero/storage/SCTLXL82/2303.html:text/html},\n}\n\n","author_short":["Antal, B. B.","Chesebro, A. G.","Strey, H. H.","Mujica-Parodi, L. R.","Weistuch, C."],"bibbaseid":"antal-chesebro-strey-mujicaparodi-weistuch-achievingoccamsrazordeeplearningforoptimalmodelreduction-2023","role":"author","urls":{"Paper":"http://arxiv.org/abs/2303.13746"},"keyword":["Computer Science - Machine Learning","Mathematics - Optimization and Control"],"metadata":{"authorlinks":{}},"downloads":4},"bibtype":"misc","biburl":"https://bibbase.org/f/8yBxW5neHauDryu9w/LCNeuro Site.bib","dataSources":["2ReyafjWJ3M2Gumqx","7f5wSqszKdrNetFrD","gEffMWRsF2gZ6Khv9","6TgWJK4SnMk85y7XM","Yz54tqD9cq9C8KFW5","hsZz2NvuN5pfuTomQ","dmYjc2WKm2QfHdZmE","4cjmrYTW8kkpYDyAZ","ijSyzcSEvf3Y9grZL","ZHHpjMFTPYP8NqJif"],"keywords":["computer science - machine learning","mathematics - optimization and control"],"search_terms":["achieving","occam","razor","deep","learning","optimal","model","reduction","antal","chesebro","strey","mujica-parodi","weistuch"],"title":"Achieving Occam's Razor: Deep Learning for Optimal Model Reduction","year":2023,"downloads":4}