- 📝 I'm a PhD candidate in ML security and privacy, co-supervised by École Polytechnique Paris and Crédit Agricole DataLab Groupe.
- 🎓 In 2023 I graduated from MVA master at ENS Paris-Saclay and engineering degree at École Polytechnique
- 🔭 For my research I'm interested in a wide range of ML topics, including ML privacy, language models, or AI safety and reliability. I'm also interested in interdisciplinary applications of ML, whether in biology, physics, or medicine.
- 🐍 My main IT language is Python. I also have strong knowledge in Java, Javascript, HTML, and some practice in PHP, C++ and D3.
- 🚀 I have a strong interest for the industrial challenges related to the developpment of high performance products, whether they are business challenges (agile development, user experience, etc) or technical challenges (software architecture, high performance computing, etc).
💬 Contact
- 📖 Here is the link to my ResearchGate profile, where I publish most of the academic work I do: https://www.researchgate.net/profile/Jeremie-Dentan
- ✉️ You can have a look at my LinkedIn profile and contact me from there!
🌱 Repositories
Topic | Paper | Comment | Repo |
---|---|---|---|
Distribution shift, Computer Vision, Forensic analysis | Using Error Level Analysis to remove Underspecification | Data challenge, computer vision on a highly underspecified task | age-underspecification |
Forgery Detection, Demosaicing | Towards a reliable detection of forgeries based on demosaicing | Evaluating the reliability of a forgery detection method based on colour demosaicing | demosaicing-detection |
NLP, Text classification | INF582 NLP Challenge: Summary Source Prediction | Data challenge, detecting computer-generated summaries | INF582-NLP-Challenge-Summary-Source-Prediction |
Adversarial attacks | Task-generalizable Adversarial Attack based on Perceptual Metric | Implementation of the NRDM attack of this paper from Naseer et al. | adversarialTransferts |
NLP, Graph mining, Protein classification | Cellular Component Ontology Prediction | Data challenge, protein classification using multimodal data (sequence + 3D structure) | Altegrad-Protein-Prediction |
Kernel methods, Graph Classification | Kernel methods for protein classification | Data challenge, using kernels e.g. Pyramid Match and Shortest Path for binary classification of proteins | kernel-proteins |
Feature Selection, Time Series classification | Laplacian Score for Feature Selection | Evaluating Laplacian Score and comparison with other methods | laplacian-score-4-time-series |
Reinforcement Learning | Cooperative Inverse Reinforcement Learning | Illustration that demonstration-by-expert can be suboptimal, as proven in this paper from Hadfield-Menell et al. | MAP578-Cooperative-Inverse-Reinforcement-Learning |
NLP, Textual entailment | Multi-lingual contradiction detection | Data challenge, feature augmentation of LLM prediction for multilingual textual entailment | contradictionDetection |
Optimal Transport on Random Graph | Paper review: Entropic Optimal Transport in Random Graphs | Test of some hypothesis in this paper of Nicolas keriven | otrg |
Reinforcement Learning, Finance | No paper associated | RL trading agent on the FX market | INF581-RL-Trading-agent |
Data Visualization | No paper associated | Visualizing some demographic indicators of France's regions | INF552-Departements-Data-Visualization |
Graph Transformers | A Generalization of Transformer Networks to Graphs | Testing some hypothesis of this paper from Dwivedi et al. on graph transformers | MAP583-Graph-Transformers |