Nos travaux de recherche

La recherche en IA, Data & Privacy

La recherche en intelligence artificielle et la protection des données sont au cœur des valeurs de notre équipe. Retrouvez nos publications en IA, NLP et Data Privacy, en collaboration notamment avec l'ENS, l'INRIA, l'Imperial College London, l'université McGill et la fondation OpenMined.

Certificat HDS

Certifié Hébergeur de Données de Santé

Norme Iso 27001

Certifié
ISO 27001

RGDP

En conformité
avec le RGPD

CNIL

En conformité avec la MR-004 et le référentiel EDS

Publications

End-to-end privacy preserving deep learning on multi-institutional medical imaging...

En savoir plus

Nature Machine Intelligence, May 2021

Georgios Kaissis, Alexander Ziller, Jonathan Passerat-Palmbach, Théo Ryffel, et al.

TESA: A Task in Entity Semantic Aggregation for Abstractive Summarization

En savoir plus

EMNLP 2020

Clément Jumel, Annie Louis, Jackie C. K. Cheung
🧑💻 github.com/clementjumel/tesa

Pre-training Is (Almost) All You Need: An Application to Commonsense Reasoning

En savoir plus

ArXiv Preprint 2020

Alexandre Tamborrino, Nicola Pellicano, Baptiste Pannier, Pascal Voitot, Louise Naudin

Health research and innovation: Can we optimize the interface between startups/pharmaceutical companies and academic health care institutions or not?

En savoir plus

ArXiv Preprint 2020

Jean-François Dhainaut, Olivier Blin, ...,Corneliu Malciu

ARIANN: Low-Interaction Privacy-Preserving Deep Learning via Function Secret Sharing

En savoir plus

ArXiv Preprint 2020

Théo Ryffel, David Pointcheval, Francis Bach

Privacy-preserving medical image analysis

En savoir plus

Med-NeurIPS 2020

Alexander Ziller, Jonathan Passerat-Palmbach, Théo Ryffel, Dmitrii Usynin, Andrew Trask, Ionésio Da Lima Costa Junior, Jason Mancuso, Marcus Makowski, Daniel Rueckert, Rickmer Braren, Georgios Kaissis,
🧑💻 github.com/OpenMined/PySyft

Toward trustworthy AI development: mechanisms for supporting verifiable claims

En savoir plus

ArXiv Preprint 2020

Miles Brundage, Shahar Avin, Jasmine Wang et al.

Partially Encrypted Machine Learning using Functional Encryption

En savoir plus

NeurIPS 2019

Théo Ryffel, Edouard Dufour-Sans, Romain Gay, Francis Bach, David Pointcheval
🧑💻 github.com/LaRiffle/collateral-learning

A Generic Framework for Privacy Preserving Deep Learning

En savoir plus

NeurIPS 2018 Workshop on Privacy-Preserving Machine Learning

Théo Ryffel, Andrew Trask, Morten Dahl, Bobby Wagner, Jason Mancuso, Daniel Rueckert, Jonathan Passerat-Palmbach
🧑💻 github.com/OpenMined/PySyft

Icône flèche gauche
Icône flèche droite

Articles de blog

PySyft + Opacus: Federated Learning with Differential Privacy

Read More

By Théo Ryffel on september 30th, 2020

Encrypted inference with ResNet-18 using PyTorch + PySyft on ants & bees images

Read More

by Théo Ryffel on September 15th, 2020

Anonymisation vs pseudonymisation: don't be fooled anymore!

Read More

by Théo Ryffel on April 6th, 2020

Encrypted Deep Learning Training with Multi-Party Computation

Read More

by Théo Ryffel on August 5th, 2019

Encrypted Deep Learning Classification with PyTorch & PySyft

Read More

by Théo Ryffel on April 16th, 2019

Deep Learning & Federated Learning in 10 Lines of PyTorch + PySyft

Read More

by Théo Ryffel on March 1st, 2019

Icône flèche gauche
Icône flèche droite

Talks

Poster : on our publication "Large Language Models as Instructors: A Study on Multilingual Clinical Entity Extraction" in association with Inria at the ACL Meeting (Canada)

Meetup FHIR France #8: OSIRIS on FHIR: How to model large sets of clinical and genomic data for multi-centric oncology research.

The Majors Originals #32: Corneliu Malciu from Arkhn.

Talk : on interactions between Differential Privacy and Multi-Party Computation at at ENS Paris (internal).

Presentation : of new tools for automated text analysis and structuration, at AP-HP.

Presentation : about data architectures in healthcare facilities and privacy enhancing techniques, at Roche.

Podcast : "Recruiting a technical team with expertise in data and healthcare". Listen the podcast.

Talk : about building a multi-Purpose stack using FHIR as a persistence layer FHIR Dev Days 2020.

Talk : at the OpenMined Privacy Conference about concrete applications of privacy in healthcare.

Presentation : of privacy-preserving demos at Paris OpenMined Meetup.

Presentation : on Federated Analytics on Real-life Healthcare Data at the Federated Learning Conference.

Talk : at FHIR Dev Days: "Pyrog: an open-source mapping tool and ETL for FHIR".

Keynote : on Data Anonymization at the BNP Paribas - Plug And Play Deep Dive.

Presentation : "Tools for Safe AI" at Laboratoire de Sciences Cognitives et Psycholinguistiques (BabyCloud team).

Presentation : of Federated Learning Techniques at ENS Paris.

Talk : at Paris Meetup OpenMined on Secure & Federated Learning at Arkhn

Talk : at Paris Meetup OpenMined to present PySyft.

Communauté Open-Source

Les projets ambitieux ont besoin d'une communauté pour les soutenir et ils doivent être accessibles au plus grand nombre

Icône Github

Github
Rejoignez-nous sur Github pour contribuer au développement d'Arkhn.

Icône twitter

Twitter
Restez au courant en nous suivant sur Twitter