Publications associées à Leverage Your Local and Global Representations: A New Self-Supervised Learning Strategy

Publications associées (31)

Exploiting Representation Similarities in Self-Supervised Learning for Vision Tasks

Recent progress in computer vision has been driven by a simple yet transformative observation: the potential of 30-year-old neural networks can be unlocked by massively scaling their parameters and the amount of human-labeled training data. This breakthrou ...

EPFL2025

Weighting Pseudo-labels via High-Activation Feature Index Similarity and Object Detection for Semi-supervised Segmentation

Minh Hieu Lê

Semi-supervised semantic segmentation methods leverage unlabeled data by pseudo-labeling them. Thus the success of these methods hinges on the reliability of the pseudo-labels. Existing methods mostly choose high-confidence pixels in an effort to avoid err ...

Springer Science and Business Media Deutschland GmbH2025

Plasma state monitoring and disruption characterization using multimodal VAEs

Giulio Romanelli, Olivier Sauter, Alessandro Pau, Yoeri Renier Johannes Poels, Cristina Venturini

When a plasma disrupts in a tokamak, significant heat and electromagnetic loads are deposited onto the surrounding device components. These forces scale with plasma current and magnetic field strength, making disruptions one of the key challenges for futur ...

IOP Publishing Ltd2025

CHAMELEON: A Flexible Data-mixing Framework for Language Model Pretraining and Finetuning

Volkan Cevher, Wanyun Xie

Training data mixtures greatly impact the generalization performance of large language models. Existing domain reweighting methods often rely on costly weight computations and require retraining when new data is introduced. To this end, we introduce a flex ...

2025

Overcoming Distribution Shifts and Imbalance Challenges in Representation Learning for Deep Regression Models

Ismail Nejjar

Deep learning has shown remarkable potential for industrial applications, particularly in predictive maintenance and condition monitoring. A key challenge in this domain is prognostics, specifically predicting the Remaining Useful Life (RUL) of assets - a ...

EPFL2025

Test-time adaptation for 6D pose tracking

Andrea Cavallaro

We propose a test -time adaptation for 6D object pose tracking that learns to adapt a pre -trained model to track the 6D pose of novel objects. We consider the problem of 6D object pose tracking as a 3D keypoint detection and matching task and present a mo ...

Elsevier Sci Ltd2024

Using Motion Cues to Supervise Single-Frame Body Pose and Shape Estimation in Low Data Regimes

Pascal Fua, Mathieu Salzmann, Andrey Davydov

When enough annotated training data is available, supervised deep-learning. algorithms excel at estimating human body pose and shape using a single camera. The effects of too little such data being available can be mitigated by using other information sour ...

Journal of Machine Learning Research Inc.2024

Asymptotics of feature learning in two-layer networks after one gradient-step

Lenka Zdeborová, Florent Gérard Krzakala, Bruno Loureiro, Hugo Chao Cui, Yatin Dandi, Luca Pesce, Yue Lu

In this manuscript, we investigate the problem of how two-layer neural networks learn features from data, and improve over the kernel regime, after being trained with a single gradient descent step. Leveraging the insight from (Ba et al., 2022), we model t ...

2024

Macroeconomic Conditioned Synthetic Financial Markets

Alexander Michael Rusnak

The creation of high fidelity synthetic data has long been an important goal in machine learning, particularly in fields like finance where the scarcity of available training and test data make it difficult to utilize many of the deep learning techniques w ...

ACM2024

Understanding generalization and robustness in modern deep learning

Maksym Andriushchenko

In this thesis, we study two closely related directions: robustness and generalization in modern deep learning. Deep learning models based on empirical risk minimization are known to be often non-robust to small, worst-case perturbations known as adversari ...

EPFL2024

On the Utility of Speech and Audio Foundation Models for Marmoset Call Analysis

Mathew Magimai Doss, Eklavya Sarkar

Marmoset monkeys encode vital information in their calls and serve as a surrogate model for neuro-biologists to understand the evolutionary origins of human vocal communication. Traditionally analyzed with signal processing-based features, recent approache ...

2024

Acoustic emission signature of martensitic transformation in laser powder bed fusion of Ti6Al4V-Fe, supported by operando X-ray diffraction

Roland Logé, Jamasp Jhabvala, Mathijs Pieter van der Meer, Milad Hamidi Nasab, Reza Esmaeilzadeh, Claire Navarre, Charlotte Léopoldine de Formanoir de la Cazerie, Lucas Maximilian Schlenger

This study focuses on investigating Acoustic Emission (AE) monitoring in the Laser Powder Bed Fusion (LPBF) process, using premixed Ti6Al4V-(x wt%) Fe, where x = 0, 3, and 6. By employing a structure-borne AE sensor, we analyze AE data statistically, uncov ...

2024

MULTI-TASK CURRICULUM LEARNING FOR PARTIALLY LABELED DATA

Jiancheng Yang

Incomplete labels are common in multi-task learning for biomedical applications due to several practical difficulties, e.g., expensive annotation efforts by experts, limit of data collection, different sources of data. A naive approach to enable joint lear ...

IEEE2023

Template Inversion Attack against Face Recognition Systems using 3D Face Reconstruction

Hatef Otroshi Shahreza

Face recognition systems are increasingly being used in different applications. In such systems, some features (also known as embeddings or templates) are extracted from each face image. Then, the extracted templates are stored in the system's database dur ...

Ieee Computer Soc2023

Multi-task Learning for Joint Re-identification, Team Affiliation, and Role Classification for Sports Visual Tracking

Vladimir Dominic K. Somers

Effective tracking and re-identification of players is essential for analyzing soccer videos. But, it is a challenging task due to the non-linear motion of players, the similarity in appearance of players from the same team, and frequent occlusions. Theref ...

Assoc Computing Machinery2023

DrapeNet: Garment Generation and Self-Supervised Draping

Pascal Fua, Mathieu Salzmann, Benoît Alain René Guillard, Ren Li

Recent approaches to drape garments quickly over arbitrary human bodies leverage self-supervision to eliminate the need for large training sets. However, they are designed to train one network per clothing item, which severely limits their generalization a ...

IEEE2023

A structured prediction approach for robot imitation learning

Aude Billard, Iason Batzianoulis

We propose a structured prediction approach for robot imitation learning from demonstrations. Among various tools for robot imitation learning, supervised learning has been observed to have a prominent role. Structured prediction is a form of supervised le ...

Sage Publications Ltd2023

How to Boost Face Recognition with StyleGAN?

Nikita Durasov

State-of-the-art face recognition systems require vast amounts of labeled training data. Given the priority of privacy in face recognition applications, the data is limited to celebrity web crawls, which have issues such as limited numbers of identities. O ...

Ieee Computer Soc2023

ChatGPT takes on the European Exam in Core Cardiology: an artificial intelligence success story?

Emmanuel Abbé, Thabodhan Mahendiran

Chat Generative Pre-trained Transformer (ChatGPT) is currently a trending topic worldwide triggering extensive debate about its predictive power, its potential uses, and its wider implications. Recent publications have demonstrated that ChatGPT can correct ...

Oxford Univ Press2023

Multi-layer state evolution under random convolutional design

Lenka Zdeborová, Florent Gérard Krzakala

Signal recovery under generative neural network priors has emerged as a promising direction in statistical inference and computational imaging. Theoretical analysis of reconstruction algorithms under generative priors is, however, challenging. For generati ...

2023