I recently receive a PhD in deep learning at Consevatoire national des arts et métiers, Cnam, in Paris. My PhD was supervised by Nicolas Thome (Sorbonne Université), Nicolas Audebert (IGN) and Clément Rambour (Cnam). My PhD was co-financed by Coexya and my industrial supervisor was Xavier Bitot (Coexya). I investigated deep learning approaches to image retrieval. Specifically, I worked on designing appropriate losses to train deep neural networks to optimize ranking losses. I also contributed on collaborative filtering recommendation using graph neural networks and on OOD detection using energy-based models. During my PhD I have published three projects, ROADMAP at NeurIPS 2021, HAPPIER at ECCV 2022 and HEAT at ICML 2023. I have submitted an extension of ROADMAP and HAPPIER to TPAMI, which is under-review. I am now actively looking for full-time or post-doc position, ideally starting at the end of the summer or in september.

News

• I defended my PhD thesis on the 20th of March. I am now officialy hold a PhD 🎉
• My pre-print submitted to TPAMI is now online at https://arxiv.org/abs/2309.08250.
• I am going to ICML 2023. I will be presenting HEAT in a poster session.
• I am participating to the Internation Computer Vision Summer School (ICVSS, 2023) this summer in Sicily.
• I have released the first of its kind hierarchical landmark retrieval dataset as a part of a TPAMI submission. It is available at https://github.com/cvdfoundation/google-landmark.
• I am a reviewer for NeurIPS 2023.
• I am going to ORASIS 2023. I will be presenting HAPPIER in a poster session.
• Our paper on OOD detection, HEAT, using energy-based models has been accepted to ICML 2023.
• I am a reviewer for ICML 2023.
• I served as a sub-reviewer for CVPR 2023.
• I am going to present our ECCV 2022 paper, HAPPIER, to the 25th of October in Tel Aviv.
• Our paper on hierarchical image retrieval, HAPPIER, has been accepted to ECCV 2022.
• I will be presenting our ROADMAP paper at RFIAP 2022.
• Our paper on ranking metric optimization for image retrieval, ROADMAP, has been accepted to NeurIPS 2021.

Publications

Robust and Decomposable Average Precision for Image Retrieval

Elias Ramzi, Nicolas Thome, Clément Rambour, Nicolas Audebert, Xavier Bitot: Robust and decomposable average precision for image retrieval. Advances in Neural Information Processing Systems 34 (NeurIPS, 2021).

Hierarchical Average Precision Training for Pertinent Image Retrieval

Elias Ramzi, Nicolas Audebert, Nicolas Thome, Clément Rambour, Xavier Bitot: RHierarchical Average Precision Training for Pertinent Image Retrieval. In: European Conference on Computer Vision. Springer (ECCV, 2022).

Hybrid Energy Based Model in the Feature Space for Out-of-Distribution Detection

Marc Lafon, Elias Ramzi, Clément Rambour, Nicolas Thome: Hybrid Energy Based Model in the Feature Space for Out-of-Distribution Detection. International Conference on Machine Learning (ICML 2023).

Optimization of Rank Losses for Image Retrieval

Elias Ramzi, Nicolas Audebert, Clément Rambour, André Araujo, Xavier Bitot, Nicolas Thome: Optimization of Rank Losses for Image Retrieval. Submitted to IEEE Transactions on Pattern Analysis and Machine Intelligence (under-review TPAMI).