{"_id":"QL3Q9LrZLhYkzTkxc","bibbaseid":"elliott-baugh-lacey-efficientexplorationandcalibrationofasemianalyticalmodelofgalaxyformationwithdeeplearning-2021","author_short":["Elliott, E. J.","Baugh, C. M.","Lacey, C. G."],"bibdata":{"bibtype":"article","type":"article","title":"Efficient exploration and calibration of a semi-analytical model of galaxy formation with deep learning","volume":"2103","url":"http://adsabs.harvard.edu/abs/2021arXiv210301072E","abstract":"We implement a sample-efficient method for rapid and accurate emulation of semi-analytical galaxy formation models over a wide range of model outputs. We use ensembled deep learning algorithms to produce a fast emulator of an updated version of the GALFORM model from a small number of training examples. We use the emulator to explore the model's parameter space, and apply sensitivity analysis techniques to better understand the relative importance of the model parameters. We uncover key tensions between observational datasets by applying a heuristic weighting scheme in a Markov chain Monte Carlo framework and exploring the effects of requiring improved fits to certain datasets relative to others. Furthermore, we demonstrate that this method can be used to successfully calibrate the model parameters to a comprehensive list of observational constraints. In doing so, we re-discover previous GALFORM fits in an automatic and transparent way, and discover an improved fit by applying a heavier weighting to the fit to the metallicities of early-type galaxies. The deep learning emulator requires a fraction of the model evaluations needed in similar emulation approaches, achieving an out-of-sample mean absolute error at the knee of the K-band luminosity function of 0.06 dex with less than 1000 model evaluations. We demonstrate that this is an extremely efficient, inexpensive and transparent way to explore multi-dimensional parameter spaces, and can be applied more widely beyond semi-analytical galaxy formation models.","urldate":"2021-05-12","journal":"arXiv e-prints","author":[{"propositions":[],"lastnames":["Elliott"],"firstnames":["Edward","J."],"suffixes":[]},{"propositions":[],"lastnames":["Baugh"],"firstnames":["Carlton","M."],"suffixes":[]},{"propositions":[],"lastnames":["Lacey"],"firstnames":["Cedric","G."],"suffixes":[]}],"month":"March","year":"2021","keywords":"Astrophysics - Astrophysics of Galaxies","pages":"arXiv:2103.01072","bibtex":"@article{elliott_efficient_2021,\n\ttitle = {Efficient exploration and calibration of a semi-analytical model of galaxy formation with deep learning},\n\tvolume = {2103},\n\turl = {http://adsabs.harvard.edu/abs/2021arXiv210301072E},\n\tabstract = {We implement a sample-efficient method for rapid and accurate emulation \nof semi-analytical galaxy formation models over a wide range of model\noutputs. We use ensembled deep learning algorithms to produce a fast\nemulator of an updated version of the GALFORM model from a small number\nof training examples. We use the emulator to explore the model's\nparameter space, and apply sensitivity analysis techniques to better\nunderstand the relative importance of the model parameters. We uncover\nkey tensions between observational datasets by applying a heuristic\nweighting scheme in a Markov chain Monte Carlo framework and exploring\nthe effects of requiring improved fits to certain datasets relative to\nothers. Furthermore, we demonstrate that this method can be used to\nsuccessfully calibrate the model parameters to a comprehensive list of\nobservational constraints. In doing so, we re-discover previous GALFORM\nfits in an automatic and transparent way, and discover an improved fit\nby applying a heavier weighting to the fit to the metallicities of\nearly-type galaxies. The deep learning emulator requires a fraction of\nthe model evaluations needed in similar emulation approaches, achieving\nan out-of-sample mean absolute error at the knee of the K-band\nluminosity function of 0.06 dex with less than 1000 model evaluations.\nWe demonstrate that this is an extremely efficient, inexpensive and\ntransparent way to explore multi-dimensional parameter spaces, and can\nbe applied more widely beyond semi-analytical galaxy formation models.},\n\turldate = {2021-05-12},\n\tjournal = {arXiv e-prints},\n\tauthor = {Elliott, Edward J. and Baugh, Carlton M. and Lacey, Cedric G.},\n\tmonth = mar,\n\tyear = {2021},\n\tkeywords = {Astrophysics - Astrophysics of Galaxies},\n\tpages = {arXiv:2103.01072},\n}\n\n","author_short":["Elliott, E. J.","Baugh, C. M.","Lacey, C. G."],"key":"elliott_efficient_2021","id":"elliott_efficient_2021","bibbaseid":"elliott-baugh-lacey-efficientexplorationandcalibrationofasemianalyticalmodelofgalaxyformationwithdeeplearning-2021","role":"author","urls":{"Paper":"http://adsabs.harvard.edu/abs/2021arXiv210301072E"},"keyword":["Astrophysics - Astrophysics of Galaxies"],"metadata":{"authorlinks":{}}},"bibtype":"article","biburl":"https://bibbase.org/zotero/polyphant","dataSources":["7gvjSdWrEu7z5vjjj"],"keywords":["astrophysics - astrophysics of galaxies"],"search_terms":["efficient","exploration","calibration","semi","analytical","model","galaxy","formation","deep","learning","elliott","baugh","lacey"],"title":"Efficient exploration and calibration of a semi-analytical model of galaxy formation with deep learning","year":2021}