no code implementations • 30 Aug 2022 • Pierrick Pochelu, Clara Erard, Philippe Cordier, Serge G. Petiton, Bruno Conche
First, it automatically performs the weakly-supervised bounding box annotation using the motion from multiple frames.
no code implementations • 30 Aug 2022 • Pierrick Pochelu, Serge G. Petiton, Bruno Conche
Deep Reinforcement Learning (or just "RL") is gaining popularity for industrial and research applications.
no code implementations • 30 Aug 2022 • Pierrick Pochelu, Serge G. Petiton, Bruno Conche
Experiments show the flexibility and efficiency under extreme scenarios: It successes to serve an ensemble of 12 heavy DNNs into 4 GPUs and at the opposite, one single DNN multi-threaded into 16 GPUs.
no code implementations • 30 Aug 2022 • Pierrick Pochelu, Serge G. Petiton, Bruno Conche
Finally, we propose a novel algorithm to optimize the inference of the DNNs ensemble in a GPU cluster based on allocation optimization.
no code implementations • 29 Sep 2021 • Pierrick Pochelu, Serge G. Petiton, Bruno Conche
Automated Machine Learning with ensembling seeks to automatically build ensembles of Deep Neural Networks (DNNs) to achieve qualitative predictions.
no code implementations • 1 Jan 2021 • Pierrick Pochelu, Bruno Conche, Serge G. Petiton
Due to the lack of consensus to design a successful deep learning ensemble, we introduce Hyperband-Dijkstra, a new workflow that automatically explores neural network designs with Hyperband and efficiently combines them with Dijkstra's algorithm.
no code implementations • 20 Aug 2019 • Maxime Soler, Martin Petitfrere, Gilles Darche, Melanie Plainchault, Bruno Conche, Julien Tierny
Different metrics, based on optimal transport, for comparing time-varying persistence diagrams in this specific applicative case are introduced.
no code implementations • 17 Aug 2018 • Maxime Soler, Mélanie Plainchault, Bruno Conche, Julien Tierny
First, we revisit the seminal assignment algorithm by Kuhn and Munkres which we specifically adapt to the problem of matching persistence diagrams in an efficient way.
no code implementations • 8 Feb 2018 • Maxime Soler, Melanie Plainchault, Bruno Conche, Julien Tierny
However, in many scenarios it is desirable to control in a similar way the preservation of higher-level notions, such as topological features , in order to provide guarantees on the outcome of post-hoc data analyses.