Hello there, I am Romain Égelé and I love science!
I am a doctor in computer science on the “Optimization of Learning Workflows at Large Scale on High-Performance Computing Platforms”. I obtained my Ph.D. in June 2024 under the joint supervision of Prof. Isabelle Guyon and Dr. Prasanna Balaprakash. I have been working on the topics Neural Architecture Search and Hyperparameter optimization since 2018. For a few years now, I have been interested in Uncertainty Quantification in Machine Learning.
Projects
DeepHyper
Scale and Automate the Developement of Your Machine Learning Workflows
An Automated Machine Learning tool-box with hyperparameter optimization, neural architecture search, uncertainty quantification, multi-fidelity, warm start. Exectuable from a single-laptop to the largest supercomputer in the world (i.e., used on Frontier at Oak Ridge National Laboratory in 2024).
Data-Centric AI
The science of systematically engineering the data of an AI system.
- Romain Egele, Julio C. S. Jacques Junior, Jan N. van Rijn, Isabelle Guyon, Xavier Baró, Albert Capés, Prasanna Balaprakash, Sergio Escalera, Thomas Moeslund, Jun Wan. “AI Competitions and Benchmarks: Dataset Development” arXiv preprint arXiv:2404.09703 (2024). (ArXiv Link)
research
2024
- [preprint] Sun, Yixuan, Ololade Sowunmi, Romain Egele, Sri Hari Krishna Narayanan, Luke Van Roekel, and Prasanna Balaprakash. “Streamlining Ocean Dynamics Modeling with Fourier Neural Operators: A Multiobjective Hyperparameter and Architecture Optimization Approach.” arXiv preprint arXiv:2404.05768 (2024).
- [preprint] Egele, Romain, Felix Mohr, Tom Viering, and Prasanna Balaprakash. “The Unreasonable Effectiveness Of Early Discarding After One Epoch In Neural Network Hyperparameter Optimization.” arXiv preprint arXiv:2404.04111 (2024).
2023
- [Best Paper Award] Egelé, Romain, Isabelle Guyon, Venkatram Vishwanath, and Prasanna Balaprakash. “Asynchronous decentralized bayesian optimization for large scale hyperparameter optimization.” In 2023 IEEE 19th International Conference on e-Science (e-Science), pp. 1-10. IEEE, 2023.
- Maulik, Romit, Romain Egele, Krishnan Raghavan, and Prasanna Balaprakash. “Quantifying uncertainty for deep learning based forecasting and flow-reconstruction using neural architecture search ensembles.” Physica D: Nonlinear Phenomena 454 (2023): 133852.
- Egele, Romain, Isabelle Guyon, Yixuan Sun, and Prasanna Balaprakash. “Is One Epoch All You Need For Multi-Fidelity Hyperparameter Optimization?.” In the 31st European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN), 2023.
- [preprint] Egele, Romain, Tyler Chang, Yixuan Sun, Venkatram Vishwanath, and Prasanna Balaprakash. “Parallel Multi-Objective Hyperparameter Optimization with Uniform Normalization and Bounded Objectives.” arXiv preprint arXiv:2309.14936 (2023).
- [preprint] Dash, Sajal, Isaac Lyngaas, Junqi Yin, Xiao Wang, Romain Egele, Guojing Cong, Feiyi Wang, and Prasanna Balaprakash. “Optimizing Distributed Training on Frontier for Large Language Models.” arXiv preprint arXiv:2312.12705 (2023).
2022
- Dorier, Matthieu, Romain Egele, Prasanna Balaprakash, Jaehoon Koo, Sandeep Madireddy, Srinivasan Ramesh, Allen D. Malony, and Rob Ross. “Hpc storage service autotuning using variational-autoencoder-guided asynchronous bayesian optimization.” In 2022 IEEE International Conference on Cluster Computing (CLUSTER), pp. 381-393. IEEE, 2022.
- Egele, Romain, Romit Maulik, Krishnan Raghavan, Bethany Lusch, Isabelle Guyon, and Prasanna Balaprakash. “Autodeuq: Automated deep ensemble with uncertainty quantification.” In 2022 26th International Conference on Pattern Recognition (ICPR), pp. 1908-1914. IEEE, 2022.
2021
- Egele, Romain, Prasanna Balaprakash, Isabelle Guyon, Venkatram Vishwanath, Fangfang Xia, Rick Stevens, and Zhengying Liu. “AgEBO-tabular: joint neural architecture and hyperparameter search with autotuned data-parallel training for tabular data.” In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, pp. 1-14. 2021.
2020
- Maulik, Romit, Romain Egele, Bethany Lusch, and Prasanna Balaprakash. “Recurrent neural network architecture search for geophysical emulation.” In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, pp. 1-14. 2020.
2019
- Balaprakash, Prasanna, Romain Egele, Misha Salim, Stefan Wild, Venkatram Vishwanath, Fangfang Xia, Tom Brettin, and Rick Stevens. “Scalable reinforcement-learning-based neural architecture search for cancer deep learning research.” In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, pp. 1-33. 2019.