jpcorb20
Research scientist @ Microsoft specialized in NLP/NLU and LLMs, with experience in deep learning, mathematical modelling and software development.
Pinned Repositories
bet-backtranslation-paraphrase-experiment
Code for experiments done for EMNLP2020.
covid19-transmission-ukf
With this repository, I derive the time-dependent R0 coefficient of the COVID-19 with the Unscented Kalman Filter from the data gathered by John Hopkins assuming the SEIR model.
google-translate-backtranslation-da
Backtranslation for NLP data augmentation done with Google API.
INF8808_exercices
Ce répertoire contient les exercices pour le cours INF8808.
iryonlp-mediqa-corr-2024
meditron
Meditron is a suite of open-source medical Large Language Models (LLMs).
pure-matrix
Pure python matrix code to do algebra with PCA (naive power iteration) and KMean (random initialization) implementations.
toxic-comment-server
Models to detect hateful comments served with flask trained on Kaggle's Toxic Comment Classification Challenge dataset.
wikipedia-lang-families
These scripts scrape and visualize data about language families from linguistics.
orsum2020_collaborative_datasets
Anonymized train set and test set used in RecSys2020 experiment to optimize the hyperparameters.
jpcorb20's Repositories
jpcorb20/bet-backtranslation-paraphrase-experiment
Code for experiments done for EMNLP2020.
jpcorb20/toxic-comment-server
Models to detect hateful comments served with flask trained on Kaggle's Toxic Comment Classification Challenge dataset.
jpcorb20/covid19-transmission-ukf
With this repository, I derive the time-dependent R0 coefficient of the COVID-19 with the Unscented Kalman Filter from the data gathered by John Hopkins assuming the SEIR model.
jpcorb20/iryonlp-mediqa-corr-2024
jpcorb20/pure-matrix
Pure python matrix code to do algebra with PCA (naive power iteration) and KMean (random initialization) implementations.
jpcorb20/google-translate-backtranslation-da
Backtranslation for NLP data augmentation done with Google API.
jpcorb20/INF8808_exercices
Ce répertoire contient les exercices pour le cours INF8808.
jpcorb20/meditron
Meditron is a suite of open-source medical Large Language Models (LLMs).
jpcorb20/montrealfintechecosystemmap
jpcorb20/webrtc.io-demo
webrtc.io multi user chat demo. Highly experimental technology
jpcorb20/wikipedia-lang-families
These scripts scrape and visualize data about language families from linguistics.