Matthijs Douze

Matthijs Douze has been a research scientist at the Facebook AI Research (FAIR) lab in Paris since November 2015. At Facebook he is working on large-scale indexing (see the Faiss library), machine learning with graphs and similarity search on images and videos. He obtained a master’s degree from the ENSEEIHT engineering school and defended his PhD at University of Toulouse in 2004. From 2005-2015 Matthijs joined the LEAR team at INRIA Grenoble where he worked on a variety of topics, including image indexing, large-scale vector indexing, event recognition in videos and similar video search. Between 2010 and 2015 he also managed Kinovis, a large 3D motion capture studio at INRIA and developed high-performance geometric algorithms for constructive solid geometry operations. In addition to FAIR’s general research topics, Matthijs is interested in snappy algorithms that process images and produce graphical results.

Matthijs's Work

Matthijs's Publications

December 13, 2019

RESEARCH

ML APPLICATIONS

Lead2Gold: Towards exploiting the full potential of noisy transcriptions for speech recognition

The transcriptions used to train an Automatic Speech Recognition (ASR) system may contain errors. Usually, either a quality control stage discards transcriptions with too many errors, or the noisy transcriptions are used as is. We introduce…

Adrien Dufraux, Emmanuel Vincent, Awni Hannun, Armelle Brun, Matthijs Douze

December 13, 2019

December 09, 2019

RESEARCH

ML APPLICATIONS

Fixing the train-test resolution discrepancy

Data-augmentation is key to the training of neural networks for image classification. This paper first shows that existing augmentations induce a significant discrepancy between the size of the objects seen by the classifier at train and test…

Hugo Touvron, Andrea Vedaldi, Matthijs Douze, Hervé Jégou

December 09, 2019

September 09, 2018

RESEARCH

ML APPLICATIONS

Deep Clustering for Unsupervised Learning of Visual Features

Clustering is a class of unsupervised learning methods that has been extensively applied and studied in computer vision. Little work has been done to adapt it to the end-to-end training of visual features on large scale datasets. In this work, we present DeepCluster, a clustering method that jointly learns the parameters of a neural network and the cluster assignments of the resulting…

Mathilde Caron, Piotr Bojanowski, Armand Joulin, Matthijs Douze

September 09, 2018