ML-DE@ECAI 2024 : Machine Learning Meets Differential Equations: From Theory to Applications
ML-DE@ECAI 2024 : Machine Learning Meets Differential Equations: From Theory to Applications

ML-DE@ECAI 2024 : Machine Learning Meets Differential Equations: From Theory to Applications

ECAI, Santiago de Compostela, Spain
Event Date: September 19, 2024 - September 20, 2024
Submission Deadline: May 15, 2024
Notification of Acceptance: July 01, 2024


Call for Papers

The [ML-DE] Workshop on "Machine Learning Meets Differential Equations: From Theory to Applications", co-located with ECAI 2024, is designed to spotlight the dynamic interplay between Machine Learning (ML) and Differential Equations (DE), two fields at the heart of numerous technological and scientific breakthroughs. This workshop aims to delve into how DEs, foundational in modeling complex systems across various domains, can be ingeniously coupled with ML to unlock new potentials, from enhancing prediction accuracies to fostering advancements in explainable AI. It is motivated by the emerging need to transcend traditional boundaries, leveraging the predictive power of ML to tackle DE-driven challenges in novel ways, thereby catalyzing a deeper understanding and innovative solutions to problems that were once considered intractable. By emphasizing energy-efficient algorithms and aiming to reduce the computational footprint of ML, the workshop underscores a commitment to sustainable AI practices. Participants will explore the integration of DEs into ML architectures, the application of ML in solving intricate DE problems, and the potential of these convergences to revolutionize fields as diverse as physics, biology, and beyond. Our purpose is to forge a community that not only shares insights but actively contributes to expanding the frontiers of what's possible at the intersection of ML and DE, setting a new paradigm for research and application in the era of intelligent technologies.

Publication Types:
-Full-Length Papers: Maximum of 8 pages, excluding references and supplementary material.
-Extended Abstracts: Limited to 2 pages, including references, designed for poster sessions and
brief elevator pitches (approximately 5 minutes). This format provides a snapshot of your research, perfect for generating interest and discussion.
-Presentation Only: Authors of papers recently published in top-tier conferences and journals (JMLR, JAIR, MLJ, PAMI, IJCAI, NeurIPS, ICLR, AISTATS, ICML, AAAI) are encouraged to submit a 2-page extended abstract, including references, for presentation. Please indicate the original publication venue in your submission form.
-Reproducibility Track: Contributions that enhance the reproducibility of research findings are crucial. We invite interactive tutorials, demos, libraries, packages or datasets (e.g., Jupyter notebooks) and their respective 2-page extended abstracts. This track emphasizes the practical application and implementation of research, facilitating a deeper understanding and broader use of ML-DE techniques. Demo code (e.g. Jupyter notebooks etc.) will be published jointly at our github together with a link to the paper.

Full-Length Papers will be in a volume by the Proceedings of Machine Learning Research (PMLR)

List of topics:
      - Embedding differential equations into machine learning (Neural ODEs, normalising flows, ...);
      - Solving differential equations using machine learning (PINNs, Neural Operators, ...);
   - Machine Learning-augmented numerical methods for solving differential equations (hybrid solvers, ...);
   - Analysis of numerical methods for incorporating differential equations' solvers into machine learning algorithms (trade-offs, benchmarks, ...);
   - Incorporation of expert-knowledge given by differential equations into machine learning algorithms (physics-inspired machine learning, ...);
      - Applications of the above to modelling/predicting real-world systems in science and engineering (finance, biology, physics, chemistry, engineering, ...); 
      - Use of machine learning to model systems described by differential equations (finance, biology, physics, chemistry, engineering, ...);
   - Approaches to extract physical knowledge out of learned differential equations for explainable AI (SINDy, ...);
   - Computational efficiency of DE solvers involved in ML algorithms (ODE solvers, ...).

For more information, visit our website: \href{}{ML-DE Workshop ECAI 2024.

Credits and Sources

[1] ML-DE@ECAI 2024 : Machine Learning Meets Differential Equations: From Theory to Applications

Check other Conferences, Workshops, Seminars, and Events