My name is Eduardo Dadalto Câmara Gomes and I'm a PhD Candidate Machine Learning at Université Paris-Saclay. I am currently working on topics on Safety AI and I am committed to making AI a beneficial technology for all. Supervised by Pablo Piantanida.
Safety in AI
Out-of-distribution detection
Uncertainty estimation
Representation learning
Semantic search
| 2023-05 | Together with colleagues, released detectors: open-source library for out-of-distribution detection on computer vision applications. |
|---|---|
| 2023-01 | Together with colleagues, released Todd: open-source library for text out-of-distribution detection built on top of the Huggingface Transformers library. |
| 2022-09 | Our work entitled “Beyond Mahalanobis Distance for Textual OOD Detection” has been accepted to appear in the NeurIPS 2022 Conference. |
| 2022-12 | Will service as a reviewer for the CVPR 2023 Conference. |
| 2022-04 | Gave a talk at a poster session of the NeurIPS 2022 Conference. |
| 2022-11 | Will attend the NeurIPS 2022 Conference in New Orleans, LA, USA. |
| 2019-10 | I am happy to start a research internship at the International Laboratory on Learning Systems (ILLS) in partnership with McGill University, ÉTS Montreal, Mila – Quebec AI Institute, CNRS and Université Paris-Saclay. |
| 2022-04 | Gave a talk at a poster session of the ICLR 2022 Conference. |
| 2022-04 | I will be attending ICLR 2022 Conference virtually. |
| 2022-01 | Our work entitled “IGEOOD: An Information Geometry Approach to Out-of-Distribution Detection” has been accepted to appear in the ICLR 2022 Conference. |
| 2021-10 | Our short version work entitled “IGEOOD: An Information Geometry Approach to Out-of-Distribution Detection” has been accepted to appear in the NeurIPS 2021 DistShift Workshop. |
| 2021-08 | Our work entitled “CSI-Aided Robust Neural-Based Decoders” has been accepted to appear in the 11th International Symposium on Topics in Coding (ISTC). |
| 2021-05 | Will service as a reviewer for the NeurIPS 2021 Conference. |
| 2020-12 | I will be a teaching assistant at the course “Introduction to Deep Learning” at CentraleSupélec. |
| 2020-10 | I am excited to start my PhD in Machine Learning at CentraleSupélec. |
For a complete list, please check my Google Scholar page.
Eduardo Dadalto, Marco Romanelli, Georg Pichler, Pablo Piantanida. “A Data-Driven Measure of Relative Uncertainty for Misclassification Detection.” Preprint. 2023.
Eduardo Dadalto, Pierre Colombo, Guillaume Staerman, Nathan Noiry, Pablo Piantanida. “A Functional Perspective on Multi-Layer Out-of-Distribution Detection.” Preprint. 2023.
Eduardo Dadalto, Marco Romanelli, Federica Granese, Siddharth Garg, Pablo Piantanida. “Trusting the Untrustworthy: A Cautionary Tale on the Pitfalls of Training-based Rejection Option.” Preprint. 2023.
Pierre Colombo, Eduardo Dadalto, Guillaume Staerman, Nathan Noiry, and Pablo Piantanida. “Beyond Mahalanobis Distance for Textual OOD Detection.” In Advances in Neural Information Processing Systems 35 (NeurIPS 2022). 2022.
Eduardo Dadalto, Florence Alberge, Pierre Duhamel, and Pablo Piantanida. “Igeood: An Information Geometry Approach to Out-of-Distribution Detection.” In International Conference on Learning Representations. 2022.
Meryem Benammar, Eduardo Dadalto, and Pablo Piantanida. “CSI-Aided Robust Neural-Based Decoders.” In 2021 11th International Symposium on Topics in Coding (ISTC), 1–5. https://doi.org/10.1109/ISTC49272.2021.9594117. 2021.
You can contact me via email at edadaltocg@gmail.com. Happy to discuss new projects and collaborations.