Publications associées à Modèle de fondation

Beyond fine-tuning: LoRA modules boost near-OOD detection and LLM security

Pierre Vandergheynst, Milos Vasic, Francesco Craighero, Renata Khasanova

Under resource constraints, LLMs are usually fine- tuned with additional knowledge using Parameter Efficient Fine-Tuning (PEFT), using Low-Rank Adaptation (LoRA) modules. In fact, LoRA injects a new set of small trainable matrices to adapt an LLM to a new ...

2024

Beyond fine-tuning: LoRA modules boost near-OOD detection and LLM security

Pierre Vandergheynst, Milos Vasic, Francesco Craighero, Renata Khasanova

Under resource constraints, LLMs are usually fine-tuned with additional knowledge using Parameter Efficient Fine-Tuning (PEFT), using Low-Rank Adaptation (LoRA) modules. In fact, LoRA injects a new set of small trainable matrices to adapt an LLM to a new t ...

2024

Performing and Detecting Backdoor Attacks on Face Recognition Algorithms

Alexander Carl Unnervik

The field of biometrics, and especially face recognition, has seen a wide-spread adoption the last few years, from access control on personal devices such as phones and laptops, to automated border controls such as in airports. The stakes are increasingly ...

EPFL2024

Social-Transmotion: Promptable Human Trajectory Prediction

Alexandre Massoud Alahi, Saeed Saadatnejad, Yang Gao, Kaouther Messaoud Ben Amor

Accurate human trajectory prediction is crucial for applications such as autonomous vehicles, robotics, and surveillance systems. Yet, existing models often fail to fully leverage the non-verbal social cues human subconsciously communicate when navigating ...

2024

Fly-Swat or Cannon? Cost-Effective Language Model Choice via Meta-Modeling

Robert West, Maxime Jean Julien Peyrard, Marija Sakota

Generative language models (LMs) have become omnipresent across data science. For a wide variety of tasks, inputs can be phrased as natural language prompts for an LM, from whose output the solution can then be extracted. LM performance has consistently be ...

Assoc Computing Machinery2024

Robust machine learning for neuroscientific inference

Steffen Schneider

Modern neuroscience research is generating increasingly large datasets, from recording thousands of neurons over long timescales to behavioral recordings of animals spanning weeks, months, or even years. Despite a great variety in recording setups and expe ...

EPFL2024

Detection of Settlements in Tanzania and Mozambique by Many Regional Few-Shot Models

Devis Tuia, Marc Conrad Russwurm, Lloyd Haydn Hughes

In this work, we propose an approach to aid in mapping small settlements, which are often misclassified by models trained on a large-scale context (global or regional). We leverage pre-trained land cover models and few-shot learning to enhance the detectio ...

2023

Monotonicity Reasoning in the Age of Neural Foundation Models

Zeming Chen, Qiyue Gao

The recent advance of large language models (LLMs) demonstrates that these large-scale foundation models achieve remarkable capabilities across a wide range of language tasks and domains. The success of the statistical learning approach challenges our unde ...

Springer2023

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

Self-Supervised Learning for Patient Stratification and Survival Analysis in Computational Pathology: An Application to Colorectal Cancer

Christian Robert Abbet

Over the years, clinical institutes accumulated large amounts of digital slides from resected tissue specimens. These digital images, called whole slide images (WSIs), are high-resolution tissue snapshots that depict the complex interaction of cells at the ...

EPFL2023

A generic diffusion-based approach for 3D human pose prediction in the wild

Alexandre Massoud Alahi, Saeed Saadatnejad, Taylor Ferdinand Mordan

Predicting 3D human poses in real-world scenarios, also known as human pose forecasting, is inevitably subject to noisy inputs arising from inaccurate 3D pose estimations and occlusions. To address these challenges, we propose a diffusion-based approach th ...

IEEE2023

4M: Massively Multimodal Masked Modeling

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

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

On matching data and model in LF-MMI-based dysarthric speech recognition

Enno Hermann

In light of steady progress in machine learning, automatic speech recognition (ASR) is entering more and more areas of our daily life, but people with dysarthria and other speech pathologies are left behind. Their voices are underrepresented in the trainin ...

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

REFINER: Reasoning Feedback on Intermediate Representations

Boi Faltings, Robert West, Maxime Jean Julien Peyrard, Antoine Bosselut, Beatriz Maria Borges Ribeiro, Debjit Paul, Mahammad Ismayilzada

Language models (LMs) have recently shown remarkable performance on reasoning tasks by explicitly generating intermediate inferences, e.g., chain-of-thought prompting. However, these intermediate inference steps may be in- appropriate deductions from the i ...

2023

Prompting Large Language Models to Power Educational Chatbots

Denis Gillet, Juan Carlos Farah, Sandy Ingram, Fanny Kim-Lan Lasne

The recent rise in both popularity and performance of large language models has garnered considerable interest regarding their applicability to education. Technologies like ChatGPT, which can engage in human-like dialog, have already disrupted educational ...

Springer2023

Diff3DHPE: A Diffusion Model for 3D Human Pose Estimation

Tong Zhang

Accurately estimating 3D human pose (3D HPE) and joint locations using only 2D keypoints is challenging. The noise in the predictions produced by conventional 2D human pose estimators often impeded the accuracy. In this paper, we present a diffusion-based ...

Ieee Computer Soc2023

Streamline RimNet: Tools for Automatic Classification of Paramagnetic Rim Lesions in MRI of Multiple Sclerosis

Meritxell Bach Cuadra, Cristina Granziera, Francesco La Rosa, Maxence Charles F Wynen

This site provides two software tools related to "RimNet: A deep 3D multimodal MRI architecture for paramagnetic rim lesion assessment in multiple sclerosis" by Barquero et al. NeuroImage: Clinical (2020). People using in part or f ...

Zenodo2023

Robustness of Local Predictions in Atomistic Machine Learning Models

Michele Ceriotti, Sanggyu Chong, Federico Grasselli, Chiheb Ben Mahmoud

Machine learning (ML) models for molecules and materials commonly rely on a decomposition of the global target quantity into local, atom-centered contributions. This approach is convenient from a computational perspective, enabling large-scale ML-driven si ...

Amer Chemical Soc2023

Social media and deep learning reveal specific cultural preferences for biodiversity

Devis Tuia

Social media has created new opportunities to map cultural ecosystem services (CES) related to biodiversity at large scales. However, using these novel data to understand people's preferences in relation to these CES remains a challenge. To address this, w ...

2023