{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,15]],"date-time":"2026-10-15T17:52:43Z","timestamp":1792086763912,"version":"4.4.0"},"reference-count":71,"publisher":"National Academy of Sciences","issue":"42","license":[{"start":{"date-parts":[[2026,10,15]],"date-time":"2026-10-15T00:00:00Z","timestamp":1792022400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/100027426","name":"Schmidt Sciences","doi-asserted-by":"publisher","award":["AI2050"],"award-info":[{"award-number":["AI2050"]}],"id":[{"id":"10.13039\/100027426","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["www.pnas.org"],"crossmark-restriction":true},"short-container-title":["Proc. Natl. Acad. Sci. U.S.A."],"published-print":{"date-parts":[[2026,10,20]]},"abstract":"<jats:p>Tipping points are abrupt, drastic, and often irreversible changes in the evolution of nonstationary and chaotic dynamical systems. For instance, increased greenhouse gas concentrations are predicted to lead to drastic decreases in low cloud cover, referred to as a climatological tipping point. In this paper, we learn the evolution of such nonstationary dynamical systems using a recurrent neural operator (RNO), which learns mappings between function spaces. After training RNO on only the pre-tipping dynamics, we employ it to detect future tipping points using an uncertainty-based approach. In particular, we propose a conformal prediction framework to forecast tipping points by monitoring deviations from physics constraints (such as conserved quantities and partial differential equations), enabling forecasting of these abrupt changes along with a rigorous measure of uncertainty. We illustrate our proposed methodology on nonstationary ordinary and partial differential equations, such as the Lorenz-63 and Kuramoto\u2013Sivashinsky equations. We also apply our methods to forecast a climate tipping point in stratocumulus cloud cover and airfoil wake and stall transitions using only limited knowledge of the governing equations. For the latter, we show that our proposed method zero-shot generalizes to forecasting multiple future tipping points under varying Reynolds numbers. In our experiments, we demonstrate that even partial or approximate physics constraints can be used to accurately forecast future tipping points.<\/jats:p>","DOI":"10.1073\/pnas.2600331123","type":"journal-article","created":{"date-parts":[[2026,10,15]],"date-time":"2026-10-15T17:37:33Z","timestamp":1792085853000},"update-policy":"https:\/\/doi.org\/10.1073\/pnas.cm10313","source":"Crossref","is-referenced-by-count":0,"title":["Tipping point forecasting in nonstationary dynamics on function spaces"],"prefix":"10.1073","volume":"123","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-9685-8383","authenticated-orcid":false,"given":"Miguel","family":"Liu-Schiaffini","sequence":"first","affiliation":[{"id":[{"id":"https:\/\/ror.org\/05dxps055","id-type":"ROR","asserted-by":"publisher"}],"name":"Division of Engineering and Applied Science, California Institute of Technology","place":["Pasadena"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1708-0997","authenticated-orcid":false,"given":"Clare E.","family":"Singer","sequence":"additional","affiliation":[{"id":[{"id":"https:\/\/ror.org\/02ttsq026","id-type":"ROR","asserted-by":"publisher"}],"name":"University of Colorado Boulder","place":["Boulder"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3650-2972","authenticated-orcid":false,"given":"Nikola","family":"Kovachki","sequence":"additional","affiliation":[{"id":[{"id":"https:\/\/ror.org\/03jdj4y14","id-type":"ROR","asserted-by":"publisher"}],"name":"Nvidia Corporation","place":["Santa Clara"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0006-7356-1046","authenticated-orcid":false,"given":"Sze Chai","family":"Leung","sequence":"additional","affiliation":[{"id":[{"id":"https:\/\/ror.org\/05dxps055","id-type":"ROR","asserted-by":"publisher"}],"name":"Division of Engineering and Applied Science, California Institute of Technology","place":["Pasadena"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6789-6209","authenticated-orcid":false,"given":"Hyunji Jane","family":"Bae","sequence":"additional","affiliation":[{"id":[{"id":"https:\/\/ror.org\/05dxps055","id-type":"ROR","asserted-by":"publisher"}],"name":"Division of Engineering and Applied Science, California Institute of Technology","place":["Pasadena"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8507-1868","authenticated-orcid":false,"given":"Kamyar","family":"Azizzadenesheli","sequence":"additional","affiliation":[{"id":[{"id":"https:\/\/ror.org\/03jdj4y14","id-type":"ROR","asserted-by":"publisher"}],"name":"Nvidia Corporation","place":["Santa Clara"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6974-6797","authenticated-orcid":false,"given":"Anima","family":"Anandkumar","sequence":"additional","affiliation":[{"id":[{"id":"https:\/\/ror.org\/05dxps055","id-type":"ROR","asserted-by":"publisher"}],"name":"Division of Engineering and Applied Science, California Institute of Technology","place":["Pasadena"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"341","published-online":{"date-parts":[[2026,10,15]]},"reference":[{"key":"e_1_3_4_1_2","doi-asserted-by":"publisher","DOI":"10.1257\/pol.6.1.137"},{"key":"e_1_3_4_2_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41561-019-0310-1"},{"key":"e_1_3_4_3_2","first-page":"1","article-title":"Stratocumulus-cumulus transition explained by bulk boundary layer theory","volume":"2023","author":"Singer C. E.","year":"2023","unstructured":"C. E. Singer, T. Schneider, Stratocumulus-cumulus transition explained by bulk boundary layer theory. ESS Open Arch. 2023, 1\u201330 (2023).","journal-title":"ESS Open Arch."},{"key":"e_1_3_4_4_2","doi-asserted-by":"crossref","unstructured":"C. E. Singer T. Schneider Co 2 -driven stratocumulus cloud breakup in a bulk boundary layer model. ESS Open Arch. 10.22541\/essoar.168167204.45220772\/v1 (2023).","DOI":"10.22541\/essoar.168167204.45220772\/v1"},{"key":"e_1_3_4_5_2","doi-asserted-by":"publisher","DOI":"10.1146\/annurev.fl.16.010184.000531"},{"key":"e_1_3_4_6_2","doi-asserted-by":"crossref","unstructured":"D. E. Rumelhart G. E. Hinton R. J. Williams \u201cLearning internal representations by error propagation\u201d in Parallel distributed processing: Explorations in the microstructures of cognition Vol. I D. E. Rumelhart J. L. McClelland Eds. (Cambridge MIT Press MA 1988) pp. 318\u2013362.","DOI":"10.7551\/mitpress\/4943.003.0128"},{"key":"e_1_3_4_7_2","doi-asserted-by":"crossref","unstructured":"D. Patel E. Ott Using machine learning to anticipate tipping points and extrapolate to post-tipping dynamics of non-stationary dynamical systems. Chaos 33 023143 (2023).","DOI":"10.1063\/5.0131787"},{"key":"e_1_3_4_8_2","unstructured":"N. Kovachki et al. Neural operator: Learning maps between function spaces with applications to pdes. J. Mach. Learn. Res. 24 1\u201397 (2023)."},{"key":"e_1_3_4_9_2","doi-asserted-by":"publisher","DOI":"10.1038\/s42254-024-00712-5"},{"key":"e_1_3_4_10_2","doi-asserted-by":"crossref","unstructured":"J. Berner et al. Principled approaches for extending neural architectures to function spaces for operator learning. Nat Mach Intell 8 1173\u20131181 (2026).","DOI":"10.1038\/s42256-026-01267-z"},{"key":"e_1_3_4_11_2","doi-asserted-by":"publisher","DOI":"10.52202\/068431-1220"},{"key":"e_1_3_4_12_2","doi-asserted-by":"publisher","DOI":"10.1098\/rsta.2011.0306"},{"key":"e_1_3_4_13_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.0705414105"},{"key":"e_1_3_4_14_2","doi-asserted-by":"publisher","DOI":"10.1126\/science.abn7950"},{"key":"e_1_3_4_15_2","doi-asserted-by":"publisher","DOI":"10.1029\/2021RG000757"},{"key":"e_1_3_4_16_2","first-page":"1","article-title":"Thermohaline ocean circulation","volume":"5","author":"Rahmstorf S.","year":"2006","unstructured":"S. Rahmstorf, Thermohaline ocean circulation. Encycl. Quat. Sci. 5, 1\u201310 (2006).","journal-title":"Encycl. Quat. Sci."},{"key":"e_1_3_4_17_2","doi-asserted-by":"publisher","DOI":"10.1175\/JCLI-D-21-0912.1"},{"key":"e_1_3_4_18_2","doi-asserted-by":"publisher","DOI":"10.1029\/2017JD027149"},{"key":"e_1_3_4_19_2","doi-asserted-by":"publisher","DOI":"10.1038\/nclimate2226"},{"key":"e_1_3_4_20_2","doi-asserted-by":"publisher","DOI":"10.1038\/nclimate2094"},{"key":"e_1_3_4_21_2","doi-asserted-by":"publisher","DOI":"10.5194\/tc-8-1699-2014"},{"key":"e_1_3_4_22_2","doi-asserted-by":"publisher","DOI":"10.1126\/sciadv.abi5738"},{"key":"e_1_3_4_23_2","doi-asserted-by":"publisher","DOI":"10.1098\/rsta.2014.0423"},{"key":"e_1_3_4_24_2","doi-asserted-by":"publisher","DOI":"10.1038\/s43017-021-00230-3"},{"key":"e_1_3_4_25_2","doi-asserted-by":"publisher","DOI":"10.1017\/jfm.2022.958"},{"key":"e_1_3_4_26_2","unstructured":"W. J. McCroskey \u201cThe phenomenon of dynamic stall\u201d (NASA Technical Memorandum NASA-TM-81264 NASA Ames Research Center Moffett Field California 1981)."},{"key":"e_1_3_4_27_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2106140118"},{"key":"e_1_3_4_28_2","doi-asserted-by":"publisher","DOI":"10.1063\/5.0042598"},{"key":"e_1_3_4_29_2","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevResearch.3.013090"},{"key":"e_1_3_4_30_2","doi-asserted-by":"publisher","DOI":"10.1063\/5.0023764"},{"key":"e_1_3_4_31_2","doi-asserted-by":"publisher","DOI":"10.34133\/research.0174"},{"key":"e_1_3_4_32_2","doi-asserted-by":"publisher","DOI":"10.1098\/rsos.211475"},{"key":"e_1_3_4_33_2","unstructured":"J. Sleeman et al. A generative adversarial network for climate tipping point discovery (tip-gan). arXiv [Preprint] (2023). https:\/\/arxiv.org\/abs\/2302.10274."},{"key":"e_1_3_4_34_2","doi-asserted-by":"crossref","unstructured":"Y. Huang S. Bathiany P. Ashwin N. Boers Deep learning for predicting rate-induced tipping. Nat. Mach. Intell. 6 1\u201310 (2024).","DOI":"10.1038\/s42256-024-00937-0"},{"key":"e_1_3_4_35_2","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevResearch.6.043194"},{"key":"e_1_3_4_36_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-024-48024-7"},{"key":"e_1_3_4_37_2","doi-asserted-by":"publisher","DOI":"10.1103\/2v76-dmg6"},{"key":"e_1_3_4_38_2","doi-asserted-by":"publisher","DOI":"10.1002\/qj.3803"},{"key":"e_1_3_4_39_2","first-page":"1","article-title":"A tutorial on conformal prediction","volume":"9","author":"Shafer G.","year":"2008","unstructured":"G. Shafer, V. Vovk, A tutorial on conformal prediction. J. Mach. Learn. Res. 9, 1\u201351 (2008).","journal-title":"J. Mach. Learn. Res."},{"key":"e_1_3_4_40_2","doi-asserted-by":"publisher","DOI":"10.1214\/aoms\/1177728174"},{"key":"e_1_3_4_41_2","doi-asserted-by":"publisher","DOI":"10.1214\/aop\/1176990746"},{"key":"e_1_3_4_42_2","doi-asserted-by":"publisher","DOI":"10.1140\/epjs\/s11734-021-00114-z"},{"key":"e_1_3_4_43_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-019-44863-3"},{"key":"e_1_3_4_44_2","doi-asserted-by":"publisher","DOI":"10.1063\/5.0226420"},{"key":"e_1_3_4_45_2","doi-asserted-by":"publisher","DOI":"10.1038\/nature08227"},{"key":"e_1_3_4_46_2","doi-asserted-by":"publisher","DOI":"10.1038\/nclimate1143"},{"key":"e_1_3_4_47_2","doi-asserted-by":"crossref","unstructured":"M. Scheffer et al. Anticipating critical transitions. Science 338 344\u2013348 (2012).","DOI":"10.1126\/science.1225244"},{"key":"e_1_3_4_48_2","doi-asserted-by":"publisher","DOI":"10.1029\/2010GL044486"},{"key":"e_1_3_4_49_2","doi-asserted-by":"publisher","DOI":"10.5194\/esd-15-1117-2024"},{"key":"e_1_3_4_50_2","doi-asserted-by":"publisher","DOI":"10.1038\/s42254-023-00650-8"},{"key":"e_1_3_4_51_2","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevLett.133.244201"},{"key":"e_1_3_4_52_2","doi-asserted-by":"publisher","DOI":"10.1007\/s10955-017-1938-0"},{"key":"e_1_3_4_53_2","unstructured":"V. Gopakumar S. Pamela L. Zanisi Fourier-rnns for modelling noisy physics data. arXiv [Preprint] (2023). https:\/\/arxiv.org\/abs\/2302.06534."},{"key":"e_1_3_4_54_2","article-title":"Latent ordinary differential equations for irregularly-sampled time series","volume":"32","author":"Rubanova Y.","year":"2019","unstructured":"Y. Rubanova, R. T. Chen, D. K. Duvenaud, Latent ordinary differential equations for irregularly-sampled time series. Adv. Neural Inf. Process. Syst. 32 (2019).","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"e_1_3_4_55_2","doi-asserted-by":"crossref","unstructured":"M. Habiba B. A. Pearlmutter \u201cNeural ordinary differential equation based recurrent neural network model\u201d in 2020 31st Irish signals and systems conference (ISSC) (IEEE 2020) pp. 1\u20136.","DOI":"10.1109\/ISSC49989.2020.9180182"},{"key":"e_1_3_4_56_2","doi-asserted-by":"crossref","unstructured":"B. Liu M. Trautner A. M. Stuart K. Bhattacharya Learning macroscopic internal variables and history dependence from microscopic models. J. Mech. Phys. Solids. 178 105329 (2023).","DOI":"10.1016\/j.jmps.2023.105329"},{"key":"e_1_3_4_57_2","unstructured":"Z. Li et al. \u201cFourier neural operator for parametric partial differential equations\u201d in International Conference on Learning Representations (2021). https:\/\/openreview.net\/forum?id=c8P9NQVtmnO. Accessed 1 August 2023."},{"key":"e_1_3_4_58_2","unstructured":"M. A. Rahman Z. E. Ross K. Azizzadenesheli U-no: U-shaped neural operators. arXiv [Preprint] (2022). https:\/\/arxiv.org\/abs\/2204.11127."},{"key":"e_1_3_4_59_2","unstructured":"M. Liu-Schiaffini et al. Tipping point forecasting in non-stationary dynamics on function spaces. arXiv [Preprint] (2023). https:\/\/arxiv.org\/abs\/2308.08794."},{"key":"e_1_3_4_60_2","unstructured":"R. Buitrago T. Marwah A. Gu A. Risteski \u201cOn the benefits of memory for modeling time-dependent PDEs\u201d in The Thirteenth International Conference on Learning Representations (2025). https:\/\/openreview.net\/forum?id=o9kqa5K3tB. Accessed 31 December 2025."},{"key":"e_1_3_4_61_2","unstructured":"B. Guo Z. Lin Q. He History-aware neural operator: Robust data-driven constitutive modeling of path-dependent materials. arXiv [Preprint] (2025). https:\/\/arxiv.org\/abs\/2506.10352 (Accessed 31 December 2025)."},{"key":"e_1_3_4_62_2","unstructured":"Z. Ye C. S. Zhang W. Wang Recurrent neural operators: Stable long-term pde prediction. arXiv [Preprint] (2025). https:\/\/arxiv.org\/abs\/2505.20721."},{"key":"e_1_3_4_63_2","unstructured":"N. F. Koren S. Lanthaler Merging memory and space: A spatiotemporal state space neural operator. arXiv [Preprint] (2025). https:\/\/arxiv.org\/abs\/2507.23428 (Accessed 31 December 2025)."},{"key":"e_1_3_4_64_2","unstructured":"K. Cho et al. Learning phrase representations using rnn encoder-decoder for statistical machine translation. arXiv [Preprint] (2014). https:\/\/arxiv.org\/abs\/1406.1078 (Accessed 1 August 2025)."},{"key":"e_1_3_4_65_2","first-page":"1231","article-title":"Nonparametric quantile estimation","volume":"7","author":"Takeuchi I.","year":"2006","unstructured":"I. Takeuchi , Nonparametric quantile estimation. J. Mach. Learn. Res. 7, 1231\u20131264 (2006).","journal-title":"J. Mach. Learn. Res."},{"key":"e_1_3_4_66_2","first-page":"23714","article-title":"Off-policy risk assessment in contextual bandits","volume":"34","author":"Huang A.","year":"2021","unstructured":"A. Huang, L. Leqi, Z. Lipton, K. Azizzadenesheli, Off-policy risk assessment in contextual bandits. Adv. Neural Inf. Process. Syst. 34, 23714\u201323726 (2021).","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"e_1_3_4_67_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2003730117"},{"key":"e_1_3_4_68_2","unstructured":"S. C. Leung D. Zhou H. J. Bae Smart sensor placement: A correlation-aware attribution framework (caaf). for real-world data modeling. arXiv [Preprint] (2025). https:\/\/arxiv.org\/abs\/2510.22517."},{"key":"e_1_3_4_69_2","doi-asserted-by":"publisher","DOI":"10.1260\/1756-8293.7.3.301"},{"key":"e_1_3_4_70_2","unstructured":"M. Liu-Schiaffini tipping-point-forecast [Source code]. GitHub. https:\/\/github.com\/neuraloperator\/tipping-point-forecast\/. Deposited 1 January 2026"},{"key":"e_1_3_4_71_2","unstructured":"M. Liu-Schiaffini et al. Tipping point forecasting in non-stationary dynamics on function spaces (airfoil dataset) (Version 1.0.0) [Data set]. Zenodo. https:\/\/doi.org\/10.5281\/zenodo.18098807. Deposited 30 December 2025."}],"container-title":["Proceedings of the National Academy of Sciences"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/pnas.org\/doi\/pdf\/10.1073\/pnas.2600331123","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,15]],"date-time":"2026-10-15T17:37:37Z","timestamp":1792085857000},"score":1,"resource":{"primary":{"URL":"https:\/\/pnas.org\/doi\/10.1073\/pnas.2600331123"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,10,15]]},"references-count":71,"journal-issue":{"issue":"42","published-print":{"date-parts":[[2026,10,20]]}},"alternative-id":["10.1073\/pnas.2600331123"],"URL":"https:\/\/doi.org\/10.1073\/pnas.2600331123","relation":{},"ISSN":["0027-8424","1091-6490"],"issn-type":[{"value":"0027-8424","type":"print"},{"value":"1091-6490","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10,15]]},"assertion":[{"value":"2026-01-06","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-09-07","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-10-15","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e2600331123"}}