Publications related to On the Limitations of Cross-lingual Encoders as Exposed by Reference-Free Machine Translation Evaluation

Advancing Self-Supervised Deep Learning for 3D Scene Understanding

Recent advancements in deep learning have revolutionized 3D computer vision, enabling the extraction of intricate 3D information from 2D images and video sequences. This thesis explores the application of deep learning in three crucial challenges of 3D com ...

EPFL2024

Dense Image-based Predictions for Comics Analysis

Deblina Bhattacharjee

Dense image-based prediction methods have advanced tremendously in recent years. Their remarkable development has been possible due to the ample availability of real-world imagery. While these methods work well on photographs, their abilities do not genera ...

EPFL2023

Stop Pre-Training: Adapt Visual-Language Models to Unseen Languages

Karl Aberer, Rémi Philippe Lebret, Negar Foroutan Eghlidi

Vision-Language Pre-training (VLP) has advanced the performance of many visionlanguage tasks, such as image-text retrieval, visual entailment, and visual reasoning. The pre-training mostly utilizes lexical databases and image queries in English. Previous w ...

Assoc Computational Linguistics-Acl2023

Interpretable Representation Learning and Evaluation for Abstractive Summarization

Andreas Thomas Marfurt

Abstractive summarization has seen big improvements in recent years, mostly due to advances in neural language modeling, language model pretraining, and scaling models and datasets. While large language models generate summaries that are fluent, coherent, ...

EPFL2023

Transformer Models for Vision

Jean-Baptiste Francis Marie Juliette Cordonnier

The recent developments of deep learning cover a wide variety of tasks such as image classification, text translation, playing go, and folding proteins. All these successful methods depend on a gradient-based learning algorithm to train a model on massive ...

EPFL2023

Text as a Richer Source of Supervision in Semantic Segmentation Tasks

Devis Tuia, Valérie Zermatten, Lloyd Haydn Hughes, Javiera Francisca Castillo Navarro

This paper introduces TACOSS a text-image alignment approach that allows explainable land cover semantic segmentation by directly integrating semantic concepts encoded from texts. TACOSS combines convolutional neural networks for visual feature extraction ...

The Institute of Electrical and Electronics Engineers, Inc2023

Class Specific Feature Disentanglement and Text Embeddings for Multi-label Generalized Zero Shot CXR Classification

Seyedbehzad Bozorgtabar

Robustness of medical image classification models is limited by its exposure to the candidate disease classes. Generalized zero shot learning (GZSL) aims at correctly predicting seen and unseen classes and most current GZSL approaches have focused on the s ...

Springer International Publishing Ag2023

Modeling Structured Data in Attention-based Models

Alireza Mohammadshahi

Natural language processing has experienced significant improvements with the development of Transformer-based models, which employ self-attention mechanism and pre-training strategies. However, these models still present several obstacles. A notable issue ...

EPFL2023

Improving Generalization of Pretrained Language Models

Rabeeh Karimi Mahabadi

In this dissertation, we propose multiple methods to improve transfer learning for pretrained language models (PLMs). Broadly, transfer learning is a powerful technique in natural language processing, where a language model is first pre-trained on a data-r ...

EPFL2023

4M: Massively Multimodal Masked Modeling

Amir Roshan Zamir, Roman Christian Bachmann, Shuqing Teresa Yeo, David Mizrahi, Oguzhan Fatih Kar

Current machine learning models for vision are often highly specialized and limited to a single modality and task. In contrast, recent large language models exhibit a wide range of capabilities, hinting at a possibility for similarly versatile models in co ...

Neural Information Processing Systems (Nips)2023

A multi-modal pre-training transformer for universal transfer learning in metal-organic frameworks

Berend Smit, Hyunsoo Park

Metal-organic frameworks (MOFs) are a class of crystalline porous materials that exhibit a vast chemical space owing to their tunable molecular building blocks with diverse topologies. An unlimited number of MOFs can, in principle, be synthesized. Machine ...

NATURE PORTFOLIO2023

BMAT: An open-source BIDS managing and analysis tool

Meritxell Bach Cuadra, Benoît Macq, Francesco La Rosa, Maxence Charles F Wynen

Magnetic Resonance Imaging (MRI) is an established technique to study in vivo neurological disorders such as Multiple Sclerosis (MS). To avoid errors on MRI data organization and automated processing, a standard called Brain Imaging Data Structure (BIDS) h ...

ELSEVIER SCI LTD2022

Biochemistry of Aminoacyl tRNA Synthetase and tRNAs and Their Engineering for Cell-Free and Synthetic Cell Applications

Sebastian Maerkl, Ragunathan Bava Ganesh

Cell-free biology is increasingly utilized for engineering biological systems, incorporating novel functionality, and circumventing many of the complications associated with cells. The central dogma describes the information flow in biology consisting of t ...

FRONTIERS MEDIA SA2022

Unsupervised Term Extraction for Highly Technical Domains

Diego Matteo Antognini

Term extraction is an information extraction task at the root of knowledge discovery platforms. Developing term extractors that are able to generalize across very diverse and potentially highly technical domains is challenging, as annotations for domains r ...

2022

Continual Test-Time Domain Adaptation

Qin Wang, Olga Fink

Test-time domain adaptation aims to adapt a source pretrained model to a target domain without using any source data. Existing works mainly consider the case where the target domain is static. However, real-world machine perception systems are running in n ...

IEEE COMPUTER SOC2022

Efficient Transformer-Based Speech Recognition

Apoorv Vyas

Training deep neural network based Automatic Speech Recognition (ASR) models often requires thousands of hours of transcribed data, limiting their use to only a few languages. Moreover, current state-of-the-art acoustic models are based on the Transformer ...

EPFL2022

Improved framework to estimate travel time and derived distributions in hydrological control volumes

Mitra Asadollahi

Crossing properties of soil saturation, defined as the duration and excursion of soil saturation below and above certain thresholds, are key variables to ecosystem functioning and evolution by primarily influencing the plant and soil microbes physiological ...

EPFL2022

Personalized Productive Engagement Recognition in Robot-Mediated Collaborative Learning

Jauwairia Nasir, Barbara Bruno

In this paper, we propose and compare personalized models for Productive Engagement (PE) recognition. PE is defined as the level of engagement that maximizes learning. Previously, in the context of robot-mediated collaborative learning, a framework of prod ...

2022

Discourse Phenomena in Machine Translation

Lesly Sadiht Miculicich Werlen

Machine Translation (MT) has made considerable progress in the past two decades, particularly after the introduction of neural network models (NMT). During this time, the research community has mostly focused on modeling and evaluating MT systems at the se ...

EPFL2021

System for planning and/or providing neuromodulation

Vincent Delattre, Grégoire Courtine, Marco Capogrosso, Fabien Bertrand Paul Wagner, Karen Minassian

The present invention relates to systems and methods for planning and/or providing neuromodulation One example system includesa stimulation related basic data storage module for storing stimulation related basic data,a stimulation related response data sto ...

2021