Bertrand Charpentier

Technical University of Munich
Department of Informatics - I26
Boltzmannstr. 3
85748 Garching b. München
Germany

Room: 00.11.062
E-Mail: charpent [at] in.tum.de

Website: sharpenb.github.io

Research Focus

  • Uncertainty Estimation in Deep Learning
  • Causality and Hierarchy in Machine Learning
  • Efficient Machine Learning
  • Machine Learning for Graphs and Sequential Data

Publications

  • Emanuele Rossi, Bertrand Charpentier, Francesco di Giovianni, Stephan Günnemann, Michael Bronstein 
    Edge Directionality Improves Learning on Heterophilic Graphs
    Mining and Learning on Graphs (MLG - ECML PKDD), 2023.
    [Paper|Github|Publisher|Video]
  • Tom Wollschlager, Nicholas Gao, Bertrand Charpentier, Mohamed Amine Ketata, Stephan Gunnemann
    Uncertainty Estimation for Molecules: Desiderata and Methods
    International Conference on Machine Learning (ICML), 2023.
    [Paper|Github|Publisher|Video]
  • Joahnnes Getzner, Bertrand Charpentier, Stephan Günnemann 
    Accuracy is not the only Metric that matters: Estimating the Energy Consumption of Deep Learning Models
    Tackling Climate Change with Machine Learning: Global Perspectives and Local Challenges Workshop (TCCML - ICLR), 2023. Spotlight talk.
    [Paper|Github|Publisher|Video]
  • Bertrand Charpentier, Chenxiang Zhang, Stephan Günnemann 
    Training, Architecture, and Prior for Deterministic Uncertainty Methods
    Pitfalls of limited data and computation for Trustworthy ML Workshop (TrustML - ICLR), 2023
    [Paper|Github|Publisher|Video]
  • Bertrand Charpentier, Ransalu Senanayake, Mykel Kochenderfer, Stephan Günnemann 
    Disentangling Epistemic and Aleatoric Uncertainty in Reinforcement Learning
    Distribution-Free Uncertainty Quantification Workshop (DFUQ - ICML), 2022
    [Paper|Github|Publisher|Video]
  • John Rachwan, Daniel Zügner, Bertrand Charpentier, Simon Geisler, Morgane Ayle, Stephan Günnemann 
    Winning the Lottery Ahead of Time: Efficient Early Network Pruning
    International Conference on Machine Learning (ICML), 2022. Spotlight talk.
    [Paper|Github|Publisher|Video]
  • Morgane Ayle, Bertrand Charpentier, John Rachwan, Daniel Zügner, Simon Geisler, Stephan Günnemann 
    On the Robustness and Anomaly Detection of Sparse Neural Networks
    Sparsity in Neural Networks Workshop (SNN), 2022
    [Paper|Github]
  • Bertrand Charpentier*, Oliver Borchert*, Daniel Zügner, Simon Geisler, Stephan Günnemann 
    Natural Posterior Network: Deep Bayesian Predictive Uncertainty for Exponential Family Distributions
    International Conference on Learning Representations (ICLR), 2022Spotlight talk.
    [Paper|Github|Publisher|Video]
  • Bertrand Charpentier, Simon Kibler, Stephan Günnemann
    Differentiable DAG Sampling 
    International Conference on Learning Representations (ICLR), 2022.
    [Paper|Github|Publisher|Video]
  • Daniel Zügner, Bertrand Charpentier, Morgane Ayle, Sascha Geringer, Stephan Günnemann
    End-to-End Learning of Probabilistic Hierarchies on Graphs
    International Conference on Learning Representations (ICLR), 2022.
    [Paper|Github|Publisher|Video]
  • Maximilian Stadler*, Bertrand Charpentier*, Simon Geisler, Daniel Zügner, Stephan Günnemann
    Graph Posterior Network: Bayesian Predictive Uncertainty for Node Classification
    Conference on Neural Information Processing Systems (NeurIPS), 2021.
    [Paper|Github|Publisher|Video]
  • Anna-Kathrin Kopetzki*, Bertrand Charpentier*, Daniel Zügner, Sandhya Giri, Stephan Günnemann
    Evaluating Robustness of Predictive Uncertainty Estimation: Are Dirichlet-based Models Reliable?
    International Conference on Machine Learning (ICML), 2021. Spotlight talk.
    [Paper|Github|Publisher]
  • Sven Elflein, Bertrand Charpentier, Daniel Zügner, Stephan Günnemann
    On Out-of-distribution Detection with Energy-Based Models
    Uncertainty and Robustness in Deep Learning Workshop (UDL - ICML), 2021.
    [Paper|Github]
  • Bertrand Charpentier, Daniel Zügner, Stephan Günnemann
    Posterior Network: Uncertainty Estimation without OOD Samples via Density-Based Pseudo-Counts
    Conference on Neural Information Processing Systems (NeurIPS), 2020.
    [Paper|GitHub|Publisher|Video]
  • Thomas Bonald, Nathan de Lara, Quentin Lutz, Bertrand Charpentier
    Scikit-network: Graph Analysis in Python
    Journal of Machine Learning Research (JMLR), 2020.
    [Paper|GitHub|Docs|Publisher]
  • Marin Bilos*, Bertrand Charpentier*, Stephan Günnemann
    Uncertainty on Asynchronous Time Event Prediction
    Conference on Neural Information Processing Systems (NeurIPS), 2019. Spotlight talk.
    [Paper|GitHub|Publisher]
  • Bertrand Charpentier, Thomas Bonald
    Tree Sampling Divergence: An Information-Theoretic Metric for Hierarchical Graph Clustering
    International Joint Conferences on Artificial Intelligence (IJCAI), 2019.
    [Paper|GitHub|Publisher]
  • Bertrand Charpentier
    Multi-scale Clustering in Graphs using Modularity
    KTH Publication Library (DiVA), 2019.
    [Paper|GitHub|Publisher]
  • Thomas Bonald , Bertrand Charpentier, Alexis Galland, Alexandre Hollocou
    Hierarchical Graph Clustering using Node Pair Sampling
    Mining and Learning with Graphs Workshop (MLG - KDD), 2018.
    [Paper|GitHub|Publisher]

Education

  • 2016 - 2018: M.Sc. in Machine Learning (passed with high distinction), KTH Royal Institute of Technology
  • 2014 - 2018: M.Sc. & B.Sc. in Mathematics and Computer Science (passed with high distinction), Ensimag
  • 2012 - 2014: CPGE in Mathematics and Physics, Lycee Henri IV

Software

  • Scikit-Network (co-creator): Simple and efficient tools for the analysis of large graphs 
    [GitHub|Documention]