Mateusz Pach, Jessica Bader, Quentin Bouniot, Serge Belongie, Zeynep Akata International Conference on Machine Learning (ICML), 2026 Text-to-image generation models have advanced rapidly, yet achieving fine-grained control over generated im…
Khalid Oublal, Quentin Bouniot, Qi Gan, Stephan Clémençon, Zeynep Akata International Conference on Machine Learning (ICML), 2026 As black box models and pretrained models gain traction in time series applications, understanding and explain…
Yiran Huang, Karsten Roth, Quentin Bouniot, Wenjia Xu, Zeynep Akata International Conference on Machine Learning (ICML), 2026 Transformer-based multimodal large language models often exhibit in-context learning (ICL) abilities. Motivated by…
Simon Roschmann, Paul Krzakala, Sonia Mazelet, Quentin Bouniot, Zeynep Akata International Conference on Machine Learning (ICML), 2026 The Platonic Representation Hypothesis posits that neural networks trained on different modalities conver…
Vasilii Feofanov, Songkang Wen, Shifeng Xie, Simon Roschmann, Marius Alonso, Hongbo Guo, Romain Ilbert, Malik Tiomoko, Quentin Bouniot, Zeynep Akata, Lujia Pan, Jianfeng Zhang, Ievgen Redko International Conference on Machine Learning (ICML…
Merve Tapli, Quentin Bouniot, Wolfgang Stammer, Zeynep Akata, Emre Akbas Concept Bottleneck Models (CBMs) ground predictions in human-understandable concepts but face fundamental limitations: the absence of a metric to pre-evaluate concept
Mateusz Pach, Shyamgopal Karthik, Quentin Bouniot, Serge Belongie, Zeynep Akata Given that interpretability and steerability are crucial to AI safety, Sparse Autoencoders (SAEs) have emerged as a tool to enhance them in Large Language Model…
Quentin Bouniot, Pavlo Mozharovskyi, Florence d'Alché-Buc International Conference on Learning Representations (ICLR), 2025 Among all data augmentation techniques proposed so far, linear interpolation of training samples, also called Mixup,…
Learning something new in real life does not necessarily mean going through a lot of examples in order to capture the essence of it. Humans are able to build upon prior experience and have the ability to adapt, allowing to combine previous
pdftk is a tool in the command line that allows you to manipulate pdf documents, such as concatenating documents or extracting pages. pdftk INPUT_FILES OPERATION output OUTPUT_FILES where OPERATION corresponds to the desired manipulation on…
In this thesis, we develop theoretical, algorithmic and experimental contributions for Machine Learning with limited labels, and more specifically for the tasks of Image Classification and Object Detection in Computer Vision. In a first con…
Quentin Bouniot, Romaric Audigier, Angelique Loesch, Amaury Habrard The use of pretrained deep neural networks represents an attractive way to achieve strong results with few data available. When specialized in dense problems such as object…