Publications associées à Self-training Improves Pre-training for Few-shot Learning in Task-oriented Dialog Systems

The multimodality cell segmentation challenge: toward universal solutions

Xin Yang, Bo Wang, Yixin Wang, Jun Ma, Lin Han, Maxime Emmanuel Scheder, Marco Labagnara, Sahand Jamal Rahi, Vojislav Gligorovski, Yao Zhang

Cell segmentation is a critical step for quantitative single-cell analysis in microscopy images. Existing cell segmentation methods are often tailored to specific modalities or require manual interventions to specify hyper-parameters in different experimen ...

Nature Portfolio2024

Infusing structured knowledge priors in neural models for sample-efficient symbolic reasoning

Mattia Atzeni

The ability to reason, plan and solve highly abstract problems is a hallmark of human intelligence. Recent advancements in artificial intelligence, propelled by deep neural networks, have revolutionized disciplines like computer vision and natural language ...

EPFL2024

Augmenting large language models with chemistry tools

Philippe Schwaller, Oliver Tobias Schilter, Andres Camilo Marulanda Bran, Carlo Baldassari

Large language models (LLMs) have shown strong performance in tasks across domains but struggle with chemistry-related problems. These models also lack access to external knowledge sources, limiting their usefulness in scientific applications. We introduce ...

Nature Portfolio2024

Advancing Self-Supervised Deep Learning for 3D Scene Understanding

Seyed Mohammad Mahdi Johari

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

Accelerator-driven Data Arrangement to Minimize Transformers Run-time on Multi-core Architectures

David Atienza Alonso, Giovanni Ansaloni, Alireza Amirshahi

The increasing complexity of transformer models in artificial intelligence expands their computational costs, memory usage, and energy consumption. Hardware acceleration tackles the ensuing challenges by designing processors and accelerators tailored for t ...

2024

Driving and suppressing the human language network using large language models

Martin Schrimpf

Transformer models such as GPT generate human-like language and are predictive of human brain responses to language. Here, using functional-MRI-measured brain responses to 1,000 diverse sentences, we first show that a GPT-based encoding model can predict t ...

Nature Portfolio2024

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

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

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

TeSLA: Test-Time Self-Learning With Automatic Adversarial Augmentation

Jean-Philippe Thiran, Guillaume Marc Georges Vray, Seyedbehzad Bozorgtabar, Devavrat Tomar

Most recent test-time adaptation methods focus on only classification tasks, use specialized network architectures, destroy model calibration or rely on lightweight information from the source domain. To tackle these issues, this paper proposes a novel Tes ...

Los Alamitos2023

VETIM: Expanding the Vocabulary of Text-to-Image Models only with Text

Sabine Süsstrunk, Radhakrishna Achanta, Mahmut Sami Arpa, Martin Nicolas Everaert

Text-to-image models, such as Stable Diffusion, can generate high-quality images from simple textual prompts. With methods such as Textual Inversion, it is possible to expand the vocabulary of these models with additional concepts, by learning the vocabula ...

BMVA2023

Revisiting Offline Compression: Going Beyond Factorization-based Methods for Transformer Language Models

Karl Aberer, Rémi Philippe Lebret, Mohammadreza Banaei

Recent transformer language models achieve outstanding results in many natural language processing (NLP) tasks. However, their enormous size often makes them impractical on memory-constrained devices, requiring practitioners to compress them to smaller net ...

Assoc Computational Linguistics-Acl2023

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

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

DISCO: Distilling Counterfactuals with Large Language Models

Antoine Bosselut, Zeming Chen, Qiyue Gao

Models trained with counterfactually augmented data learn representations of the causal structure of tasks, enabling robust generalization. However, high-quality counterfactual data is scarce for most tasks and not easily generated at scale. When crowdsour ...

Assoc Computational Linguistics-Acl2023

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

Dual-frequency spectral radar retrieval of snowfall microphysics: a physics-driven deep-learning approach

Alexis Berne, Gionata Ghiggi, Anne-Claire Marie Billault--Roux

The use of meteorological radars to study snowfall microphysical properties and processes is well established, in particular via a few distinct techniques: the use of radar polarimetry, of multi-frequency radar measurements, and of the radar Doppler spectr ...

COPERNICUS GESELLSCHAFT MBH2023

Yes but.. Can ChatGPT Identify Entities in Historical Documents?

Ahmed Hamdi, Emanuela Boros

Large language models (LLMs) have been leveraged for several years now, obtaining state-of-the-art performance in recognizing entities from modern documents. For the last few months, the conversational agent ChatGPT has "prompted" a lot of interest in the ...

Assoc Computing Machinery2023

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