Daalụ maka ileta Nature.com. Ụdị ihe nchọgharị ị na-eji nwere nkwado CSS nwere oke. Maka nsonaazụ kacha mma, anyị na-akwado ka ị jiri ụdị ihe nchọgharị gị kachasị ọhụrụ (ma ọ bụ gbanyụọ ọnọdụ ndakọrịta na Internet Explorer). Ka ọ dị ugbu a, iji hụ na nkwado na-aga n'ihu, anyị na-egosi saịtị na-enweghị styling ma ọ bụ Javascript.
Otu n'ime ngwa kachasị mma nke mmụta igwe na kọmpụta kọmpụta bụ ngwangwa ngwangwa nke nha nhata dị iche iche (PDEs). Ebumnuche bụ isi nke igwe mmụta na-edozi akụkụ dị iche iche nha nhata bụ imepụta azịza ziri ezi ngwa ngwa karịa usoro ọnụọgụgụ iji rụọ ọrụ dịka ntụnyere ntọala. Anyị na-ebu ụzọ mee nyocha nhazi nke igwe na-amụ akwụkwọ na-edozi nha anya akụkụ dị iche iche. N'ime akwụkwọ niile na-akọwa iji ML dozie nha nha nha anya dị iche iche ma na-ekwu na ịdị elu karịa usoro ọnụọgụgụ, anyị chọpụtara 79% (60/76) ma e jiri ya tụnyere ntọala ndị na-adịghị ike. Nke abụọ, anyị chọtara ihe akaebe na-egosi enweghị mkpesa zuru oke, ọkachasị n'ịkọpụta nsonaazụ yana mkparị mbipụta. Anyị kwubiri na nchọnchọ mmụta igwe maka idozi nha nha anya dị iche na-enwe oke nchekwube: data ntinye adịghị ike nwere ike ibute nsonaazụ dị oke mma, na ịkọ akụkọ nwere ike bute ịkọpụta nsonaazụ ọjọọ. N'akụkụ dị ukwuu, nsogbu ndị a na-egosi na ọ bụ ihe ndị yiri ya na ọgba aghara mmeghari n'oge gara aga: ezi uche nke onye nyocha na nhụsianya dị mma. Anyị na-akpọ maka mgbanwe omenala nke dị n'okpuru ala ka ibelata akụkọ na-adịghị mma na mgbanwe nhazi nke elu-ala iji belata ihe mkpali gbagọrọ agbagọ ime ya.
Ndepụta nke ndị edemede na akụkọ ndị a na-eme site na nyochaa usoro nhazi, yana nhazi nke isiokwu ọ bụla n'ime ihe atụ, dị n'ihu ọha na https://doi.org/10.17605/OSF.IO/GQ5B3 (ref. 124).
Enwere ike ịchọta koodu achọrọ iji mepụtaghachi nsonaazụ dị na tebụl 2 na GitHub: https://github.com/nickmcgreivy/WeakBaselinesMLPDE/ (ref. 125) na na Code Ocean: https://codeocean.com/capsule/9605539/Tree/ v1 (njikọ 126) na https://codeocean.com/capsule/0799002/tree/v1 (njikọ 127).
Randall, D., na Welser, K., Nsogbu enweghị ike ime na sayensị ọgbara ọhụrụ: Ihe kpatara, nsonaazụ, na ụzọ maka mgbanwe (National Association of Scientists, 2018).
Ritchie, S. Akụkọ Sayensị Sayensị: Otu Wayo, Agbaghara, ịgbachi nkịtị, na Hype si mebie Ọchụchọ Eziokwu (Vintage, 2020).
Mepee mmekorita sayensị. Ịtụle reproducibility na nkà mmụta uche. Sayensị 349, AAAC4716 (2015).
Prinz, F., Schlange, T., na Asadullah, K. Kwere ya ma ọ bụ ghara: Ego ole ka anyị nwere ike ịdabere na data ebipụta na ebumnuche ọgwụ nwere ike? Mba. Mkpu. “Nchọpụta Ọgwụgwọ.” 10, 712 (2011).
Begley, KG na Ellis, LM Ịkwalite ụkpụrụ na nyocha ọrịa kansa. Ọdịdị 483, 531-533 (2012).
A. Gelman na E. Loken, Ogige nke ụzọ ịhịa aka n'ahụ: Gịnị kpatara ọtụtụ ntụnyere ji bụrụ nsogbu ọbụlagodi na "njem azụ azụ" ma ọ bụ "p-hacks" na Preformed Research Hypotheses, vol. 348, 1–17 (Ngalaba Statistics, 2013).
Karagiorgi, G., Kasecka, G., Kravitz, S., Nachman, B., na Shi, D. Machine mmụta na-achọ ọhụrụ isi physics. Mba. Dọkịta nke Philosophy na Physics. 4, 399–412 (2022).
Dara S, Damercherla S, Jadhav SS, Babu CM, Ahsan MJ. Ọmụmụ igwe na nchọpụta ọgwụ: nyocha. Atif. Intel. Ed. 55, 1947–1999 (2022).
Mather, AS na Coote, ML mmụta miri emi na kemistri. J. Chemistry. gosi. Nlereanya. 59, 2545–2559 (2019).
Rajkomar A., Dean J. na Kohan I. Machine mmụta na nkà mmụta ọgwụ. Akwụkwọ akụkọ Ọgwụ nke New England. 380, 1347–1358 (2019).
Grimmer J, Roberts ME. na Stewart BM Machine mmụta na sayensị mmekọrịta ọha na eze: ụzọ agnostic. Rev. Ann Ball. sayensị. 24, 395–419 (2021).
Jump, J. et al. Jiri alfafold mee amụma nhazi protein ziri ezi nke ukwuu. Ọdịdị 596, 583–589 (2021).
Gundersen, OE, Coakley, K., Kirkpatrick, K., na Gil, Y. Isi mmalite nke enweghị ike imepụtaghachi na mmụta igwe: nyocha. Mbipụta dị na https://arxiv.org/abs/2204.07610 (2022).
Scully, D., Snook, J., Wiltschko, A., na Rahimi, A. Onye mmeri? Na ọsọ, ọganihu na ike nke ihe akaebe pụtara ìhè (ICLR, 2018).
Armstrong, TG, Moffat, A., Webber, W., na Zobel, J. Nkwalite na-abụghị mgbakwunye: nsonaazụ ọchụchọ mbido kemgbe 1998. 18th ACM Conference on Information and Knowledge Management 601-610 (ACM 2009).
Kapoor, S. na Narayanan, A. Leakage na nsogbu ịmụgharị na sayensị dabere na igwe mmụta. Ụdị, 4, 100804 (2023).
Kapoor S. et al. Ndozigharị: ụkpụrụ mkpesa sayensị dabere na mmụta igwe. Mbipụta dị na https://arxiv.org/abs/2308.07832 (2023).
DeMasi, O., Cording, C., na Recht, B. Ntụnyere enweghị isi nwere ike ibute nchekwube ụgha na mmụta igwe ọgwụ. PloS ONE 12, e0184604 (2017).
Roberts, M., et al. Ọnyà ndị a na-ahụkarị na omume kacha mma maka iji mmụta igwe iji chọpụta na ịkọ amụma COVID-19 site na ụzarị obi yana tomography gbakọọ. Mba. Oke. Intel. 3, 199–217 (2021).
Winantz L. et al. Ụdị amụma maka nchoputa na prognosis nke COVID-19: nyocha usoro na nyocha dị oke egwu. BMJ 369, m1328 (2020).
Whalen S., Schreiber J., Noble WS na Pollard KS imeri ọnyà nke iji mmụta igwe na genomics. Mba. Pastor Ginette. 23, 169–181 (2022).
Atris N. et al. Omume kacha mma maka mmụta igwe na kemistri. Mba. Chemical. 13, 505–508 (2021).
Brunton SL na Kutz JN na-ekwe nkwa ntụzịaka maka mmụta igwe nke nha anya dị iche iche. Mba. gbakọọ. sayensị. 4, 483–494 (2024).
Vinuesa, R. na Brunton, SL Na-emeziwanye ihe omimi mgbakọ na mwepụ site na mmụta igwe. Mba. gbakọọ. sayensị. 2, 358–366 (2022).
Comeau, S. et al. Igwe mmụta sayensị na netwọkụ akwara ozi anụ ahụ: Ebe anyị nọ ugbu a na ihe na-esote. J. Sayensị. gbakọọ. 92, 88 (2022).
Duraisamy, K., Yaccarino, G., na Xiao, H. Turbulence modeling na oge data. Mbipụta Ann emegharịgharịrị. 51, 357–377 (2019).
Durran, DR Ụzọ ọnụọgụgụ maka idozi nha nha ebili mmiri na geophysical hydrodynamics, vol. 32 (Springer, 2013).
Mishra, S. Usoro mmụta igwe maka ịgbakọ data na-akwalite ngwa ngwa nke nha nhata dị iche iche. mgbakọ na mwepụ. injinia. https://doi.org/10.3934/Mine.2018.1.118 (2018).
Kochikov D. et al. Ịmụ ihe n'ịmụ igwe - osooso usoro ịgbakọ mmiri mmiri. usoro. National Academy of Sciences. sayensị. US 118, e2101784118 (2021).
Kadapa, K. Machine mmụta maka sayensị kọmputa na injinịa - okwu mmeghe dị nkenke na ụfọdụ isi okwu. Mbipụta dị na https://arxiv.org/abs/2112.12054 (2021).
Ross, A., Li, Z., Perezhogin, P., Fernandez-Granda, C., na Zanna, L. Comparative analysis of machine learning Ocean subgrid parameterization in idealized model. J.Adv. Nlereanya. usoro ụwa. 15. e2022MS003258 (2023).
Lippe, P., Wieling, B., Perdikaris, P., Turner, R., na Brandstetter, J. PDE nnụcha: nweta ezi extrusions ogologo na a neural PDE solver. Nzukọ 37th na Sistemụ Nhazi Ozi Neural (NeurIPS 2023).
Frachas, PR et al. Backpropagation algọridim na reservoir ngụkọta oge na netwọk neural ugboro ugboro maka ịkọ mgbagwoju spatiotemporal ike. neural netwọk. 126, 191–217 (2020).
Raissi, M., Perdikaris, P. na Karniadakis, GE Physics, sayensị kọmputa, netwọkụ akwara: usoro mmụta miri emi maka idozi nsogbu na-aga n'ihu na mgbagha nke metụtara nha anya akụkụ dị iche iche na-adịghị adị. J. Kọmputa. physics. 378, 686–707 (2019).
Grossmann, TG, Komorowska, UJ, Lutz, J., na Schönlieb, K.-B. Netwọk akwara nke dabere na physics ọ nwere ike karịa ụzọ mmewere nwere oke? Ngwa IMA J. mgbakọ na mwepụ. 89, 143–174 (2024).
de la Mata, FF, Gijon, A., Molina-Solana, M., na Gómez-Romero, J. Physics dabeere neural netwọk maka data-chụpụrụ modeling: uru, njedebe, na ohere. physics. A 610, 128415 (2023).
Zhuang, P.-Y. & Barba, LA Akụkọ dị egwu na netwọkụ akwara gbasara physics na nhazi mmiri: ọnyà na ndakpọ olileanya. Mbipụta dị na https://arxiv.org/abs/2205.14249 (2022).
Zhuang, P.-Y. na Barba, LA Mgbochi amụma nke netwọk akwara ozi anụ ahụ na nhazi vortex. Mbipụta dị na https://arxiv.org/abs/2306.00230 (2023).
Wang, S., Yu, H., na Perdikaris, P. Mgbe na ihe kpatara pinn na-ejighị azụ: Anya akwara tangent akwara. J. Kọmputa. physics. 449, 110768 (2022).
Krishnapriyan, A., Gholami, A., Zhe, S., Kirby, R., na Mahoney, MW Njirimara nke ụdị ọdịda enwere ike na netwọkụ akwara ozi. Nzukọ 35 na Sistemụ nhazi ozi akwara Vol. 34, 26548–26560 (NeurIPS 2021).
Basir, S. na Senokak, I. Nnyocha dị oke egwu nke ụdị ọdịda ọdịda na netwọkụ akwara dabere na physics. Na AiAA SCITECH 2022 Forum 2353 (ARK, 2022).
Karnakov P., Litvinov S. na Koumoutsakos P. Na-edozi nsogbu anụ ahụ na-agbanwe agbanwe site n'ịmepụta ihe efu efu: mmụta ngwa ngwa na nke ziri ezi na-enweghị netwọkụ akwara. usoro. National Academy of Sciences. sayensị. Nexus 3, pgae005 (2024).
Gundersen OE Ụkpụrụ bụ isi nke ịmụgharị. Phil.cross. R. Shuker. A 379, 20200210 (2021).
Aromataris E na Pearson A. Nlebanya usoro: nyocha. Ee. J. Nọọsụ 114, 53–58 (2014).
Magiera, J., Ray, D., Hesthaven, JS, na Rohde, K. Constraint-aware neural networks maka nsogbu Riemann. J. Kọmputa. physics. 409, 109345 (2020).
Bezgin DA, Schmidt SJ na Adams NA Ihe data na-eme ka sekit nwere oke anụ ahụ na-ebute maka mbelata voltaji na-abụghị oge gboo. J. Kọmputa. physics. 437, 110324 (2021).
Oge nzipu: Sep-29-2024