
On the importance of pretraining data volume for compact language models
Recent advances in language modeling have led to computationally intensi...
read it

Fast Transformers with Clustered Attention
Transformers have been proven a successful model for a variety of tasks ...
read it

Transformers are RNNs: Fast Autoregressive Transformers with Linear Attention
Transformers achieve remarkable performance in several tasks but due to ...
read it

Taming GANs with Lookahead
Generative Adversarial Networks are notoriously challenging to train. Th...
read it

Gradient Alignment in Deep Neural Networks
One cornerstone of interpretable deep learning is the high degree of vis...
read it

Fair LatencyAware Metric for realtime video segmentation networks
As supervised semantic segmentation is reaching satisfying results, many...
read it

Multitask Reinforcement Learning with a Planning QuasiMetric
We introduce a new reinforcement learning approach combining a planning ...
read it

On the Tunability of Optimizers in Deep Learning
There is no consensus yet on the question whether adaptive gradient meth...
read it

Processing Megapixel Images with Deep AttentionSampling Models
Existing deep architectures cannot operate on very large signals such as...
read it

FullJacobian Representation of Neural Networks
Nonlinear functions such as neural networks can be locally approximated...
read it

Reducing Noise in GAN Training with Variance Reduced Extragradient
Using large minibatches when training generative adversarial networks (...
read it

Practical Deep Stereo (PDS): Toward applicationsfriendly deep stereo matching
Endtoend deeplearning networks recently demonstrated extremely good p...
read it

Not All Samples Are Created Equal: Deep Learning with Importance Sampling
Deep neural network training spends most of the computation on examples ...
read it

Knowledge Transfer with Jacobian Matching
Classical distillation methods transfer representations from a "teacher"...
read it

Geodesic Convolutional Shape Optimization
Aerodynamic shape optimization has many industrial applications. Existin...
read it

SGAN: An Alternative Training of Generative Adversarial Networks
The Generative Adversarial Networks (GANs) have demonstrated impressive ...
read it

The WILDTRACK MultiCamera Person Dataset
People detection methods are highly sensitive to the perpetual occlusion...
read it

Geometric calibration of Colour and Stereo Surface Imaging System of ESA's Trace Gas Orbiter
There are many geometric calibration methods for "standard" cameras. The...
read it

Semisupervised learning of deep metrics for stereo reconstruction
Deeplearning metrics have recently demonstrated extremely good performa...
read it

Globally Consistent MultiPeople Tracking using Motion Patterns
Many stateoftheart approaches to people tracking rely on detecting th...
read it

Social Scene Understanding: EndtoEnd MultiPerson Action Localization and Collective Activity Recognition
We present a unified framework for understanding human social behaviors ...
read it

MultiModal MeanFields via CardinalityBased Clamping
Mean Field inference is central to statistical physics. It has attracted...
read it

Predicting the dynamics of 2d objects with a deep residual network
We investigate how a residual network can learn to predict the dynamics ...
read it

A SubQuadratic Exact Medoid Algorithm
We present a new algorithm, trimed, for obtaining the medoid of a set, t...
read it

Scalable Metric Learning via Weighted Approximate Rank Component Analysis
We are interested in the largescale learning of Mahalanobis distances, ...
read it

Nested MiniBatch KMeans
A new algorithm is proposed which accelerates the minibatch kmeans alg...
read it

Fast KMeans with Accurate Bounds
We propose a novel accelerated exact kmeans algorithm, which performs b...
read it
François Fleuret
verfied profile
Full Professor of Computer Science, University of Geneva. Adjunct Professor, École Polytechnique Fédérale de Lausanne. Cofounder Neural Concept SA.