no code implementations • 16 Aug 2023 • Clara Macabiau, Thanh-Dung Le, Kevin Albert, Philippe Jouvet, Rita Noumeir
With a precision of 91%, a recall of 90% and an F1 score of 90% for the class without artifacts, the results demonstrate its effectiveness in labeling a medical dataset, even when clean samples are rare.
no code implementations • 22 Mar 2023 • Thanh-Dung Le, Philippe Jouvet, Rita Noumeir
In this study, we propose a simplified Switch Transformer framework and train it from scratch on a small French clinical text classification dataset at CHU Sainte-Justine hospital.
no code implementations • 26 Sep 2022 • Thanh-Dung Le, Rita Noumeir, Jerome Rambaud, Guillaume Sans, Philippe Jouvet
Goal: Our aim is therefore to access an alternative approach to tackle the sparsity by compressing the clinical representation feature space, where limited French clinical notes can also be dealt with effectively.
no code implementations • 8 Apr 2021 • Thanh-Dung Le, Rita Noumeir, Jerome Rambaud, Guillaume Sans, Philippe Jouvet
This study successfully applied learning representation and machine learning algorithms to detect heart failure from clinical natural language in a single French institution.
no code implementations • 8 Apr 2021 • Thanh-Dung Le, Rita Noumeir, Jerome Rambaud, Guillaume Sans, Philippe Jouvet
In the case without any feature selection, the proposed framework yielded an overall classification performance with acc, pre, rec, and f1 of 84% and 82%, 85%, and 83%, respectively.
no code implementations • 21 Jan 2021 • Alban Main de Boissiere, Rita Noumeir
Moreover, our networks use infrared from RGB-D cameras, which we are the first to use for online action detection, to our knowledge.
1 code implementation • submitted to IEEE Access 2020 • Alban Main de Boissiere, Rita Noumeir
Ablation studies show that using pre-trained networks on other large scale datasets as our modules and data augmentation yield considerable improvements on the action classification accuracy.
Ranked #18 on Action Recognition on NTU RGB+D