Hi! I'm Yuan Yin

PhD from the MLIA Team at Sorbonne Université, ISIR.


Younger me at the Rotterdam Centraal train station back in 2018.

My research focuses on machine learning and deep learning for spatiotemporal sequence modeling, prediction, and analysis of complex behaviors, with a special emphasis on physical dynamical systems.

I have explored the following directions in the past years:

  • improving neural network-based dynamics modeling by leveraging physical models through hybrid modeling,
  • extending the generalization power of dynamics models by learning commonalities from data of different dynamics to extrapolate to unseen systems, and
  • handling free-form data and continuously predicting phenomena in time and space through continuous modeling.

I successfully defended my PhD thesis in June 2023 at Sorbonne Université, supervised by Prof. Patrick Gallinari and Assoc. Prof. Nicolas Baskiotis.

Prior to that, I earned my BSc in Computer Science from Beihang University back in 2016, followed by an MSc in Computer Science from Université Paris Cité and Sorbonne Université in 2019.

News More >>>

Jan 31, 2024 I gave a talk at the Workshop “Mathematical Foundations of AI”, at DATAIA-SCAI
Jan 25, 2024 I gave a talk at the seminar at Valeo.ai.
Dec 13, 2023 I have been recognized as a Top Reviewer at NeurIPS 2023!
Nov 30, 2023 I gave a talk at the seminar of UMR MIA Paris-Saclay, at AgroParisTech.
Nov 20, 2023 I gave a talk on APHYNITY at the seminar of Team MCS of Laboratoire Analyse, Géométrie et Applications, at Université Sorbonne Paris Nord (Paris-13).
Sep 22, 2023 Patrick and I co-presented at ECML-PKDD 2023 the tutorial on Physics-Aware Deep Learning (Site & slides)
Sep 21, 2023 CORAL, a paper I co-ordinated, has been accepted at NeurIPS 2023!
Jun 28, 2023 I successfully defended my PhD thesis.
Jun 19, 2023 INFINITY, a paper I co-ordinated, has been accepted at SynS & ML Workshop @ ICML 2023!
May 4, 2023 I presented CoDA and DINo at the Seminar of Signal Processing Laboratory (LTS4) at EPFL
May 3, 2023 I’m sorry to announce that I won’t be able to attend our ICLR spotlight presentation of DINo in Kigali, Rwanda. Please check our paper, poster and slides here.
Apr 17, 2023 Matthieu and I gave a talk on DINo at AI for Science Talks.
Mar 27, 2023 I gave a talk on APHINITY at the seminar of SIG LearnSFluid.
Mar 20, 2023 I was invited to give a talk at AP-HP on applying machine learning to medical biology.
Jan 21, 2023 Our latest work, DINo, has been accepted at ICLR 2023 as a Spotlight paper (officially called Notable-Top-25%).
Nov 28, 2022 Jean-Yves and I gave a talk on DINo at the seminar of Criteo AI Lab.
Oct 18, 2022 I presented APHYNITY at the joint seminar of Sorbonne Université, ISAE, and Cerfacs.
Jul 17, 2022 Matthieu presented our work CoDA in person at ICML 2022 in New Orleans.
May 14, 2022 Our work, CoDA, has been accepted at ICML 2022 as Spotlight presentation.
Feb 8, 2022 I gave a talk on my work on hybrid modeling and generalization at Extrality, a French startup working on machine learning accelerated CFD models.
Dec 8, 2021 I presented our work LEADS at NeurIPS 2021 and NeurIPS@Paris 2021.
Sep 27, 2021 Our work, LEADS, has been accepted at NeurIPS 2021.
May 6, 2021 Ibrahim presented APHYNITY at ICLR 2021.
Mar 22, 2021 I presented the first version of LEADS at AAAI 2021 Spring Symposium on Combining Artificial Intelligence and Machine Learning with Physics Sciences (MLPS).
Jan 7, 2021 Our latest work, APHYNITY, has been accepted at ICLR 2021 as Oral presentation.
Oct 4, 2019 I presented “Unsupervised Inpainting for Occluded Sea Surface Temperature Sequences” at Climate Informatics 2019.