Scholar
Joseph Salmon
Google Scholar ID: m7OEDmoAAAAJ
Inria
Citations & Impact
Citations
1,839
H-index
24
i10-index
41
Publications
20
Co-authors
30
list available
Publications
Resume (English only)
Academic Achievements
- - Publications:
- - Conformal Prediction for Long-Tailed Classification, ArXiv e-prints, 2025
- - Class conditional conformal prediction for multiple inputs by p-value aggregation, NeurIPS, 2025
- - A Two-Head Loss Function for Deep Average-K Classification, WACV, 2025
- - Cooperative learning of Pl@ntNet's Artificial Intelligence algorithm: how does it work and how can we improve it?, Methods in Ecology and Evolution, 2025
- - peerannot: A framework for label aggregation in crowdsourced datasets, JDS 2024, 2024
- - Local linear convergence of proximal coordinate descent algorithm, Optimization Letter, 2024
- - Awards:
- - IUF Nomination (junior member) in July 2021
- - Projects:
- - ANR VITE project (PI: B. Thirion, theme: variable importance/explainability) accepted in September 2023.
Research Experience
- - Senior Researcher at Inria (Team: Iroko) and collaborating with the Pl@nNet team since October 2024.
- - Full Professor at Université de Montpellier from 2018 to 2024, and Junior member of Institut Universitaire de France (IUF) from 2021 to 2024.
- - Visiting Assistant Professor at UW, Statistics Department during the spring and summer quarters of 2018.
- - Assistant Professor at Telecom Paris and associate member at INRIA Parietal Team (now Mind Team) from 2012 to 2018.
Education
- - Ph.D. in Statistics and Image Processing from Université Paris Diderot in 2010, supervised by Dominique Picard and Erwan Le Pennec.
- - Post-doctoral Associate at Duke University from 2011 to 2012, working with Rebecca Willett.
Background
- A statistician and applied mathematician with a strong interest in machine learning, optimization, and data science. Main research areas include citizen science, crowd-sourcing, and high-dimensional statistics.
Miscellany
- Blog posts on Isotonic regression, LaTeX for scientific writing, and Soft-max and Soft-argmax.
