Zobrazeno 1 - 10
of 52
pro vyhledávání: '"Bouchacourt, Diane"'
Neural networks can fail when the data contains spurious correlations. To understand this phenomenon, researchers have proposed numerous spurious correlations benchmarks upon which to evaluate mitigation methods. However, we observe that these benchm
Externí odkaz:
http://arxiv.org/abs/2409.04188
Autor:
Al-Tahan, Haider, Garrido, Quentin, Balestriero, Randall, Bouchacourt, Diane, Hazirbas, Caner, Ibrahim, Mark
Significant research efforts have been made to scale and improve vision-language model (VLM) training approaches. Yet, with an ever-growing number of benchmarks, researchers are tasked with the heavy burden of implementing each protocol, bearing a no
Externí odkaz:
http://arxiv.org/abs/2408.04810
Autor:
Sobal, Vlad, Ibrahim, Mark, Balestriero, Randall, Cabannes, Vivien, Bouchacourt, Diane, Astolfi, Pietro, Cho, Kyunghyun, LeCun, Yann
Learning good representations involves capturing the diverse ways in which data samples relate. Contrastive loss - an objective matching related samples - underlies methods from self-supervised to multimodal learning. Contrastive losses, however, can
Externí odkaz:
http://arxiv.org/abs/2407.18134
Autor:
Kitouni, Ouail, Nolte, Niklas, Bouchacourt, Diane, Williams, Adina, Rabbat, Mike, Ibrahim, Mark
Today's best language models still struggle with hallucinations: factually incorrect generations, which impede their ability to reliably retrieve information seen during training. The reversal curse, where models cannot recall information when probed
Externí odkaz:
http://arxiv.org/abs/2406.05183
Autor:
Bordes, Florian, Pang, Richard Yuanzhe, Ajay, Anurag, Li, Alexander C., Bardes, Adrien, Petryk, Suzanne, Mañas, Oscar, Lin, Zhiqiu, Mahmoud, Anas, Jayaraman, Bargav, Ibrahim, Mark, Hall, Melissa, Xiong, Yunyang, Lebensold, Jonathan, Ross, Candace, Jayakumar, Srihari, Guo, Chuan, Bouchacourt, Diane, Al-Tahan, Haider, Padthe, Karthik, Sharma, Vasu, Xu, Hu, Tan, Xiaoqing Ellen, Richards, Megan, Lavoie, Samuel, Astolfi, Pietro, Hemmat, Reyhane Askari, Chen, Jun, Tirumala, Kushal, Assouel, Rim, Moayeri, Mazda, Talattof, Arjang, Chaudhuri, Kamalika, Liu, Zechun, Chen, Xilun, Garrido, Quentin, Ullrich, Karen, Agrawal, Aishwarya, Saenko, Kate, Celikyilmaz, Asli, Chandra, Vikas
Following the recent popularity of Large Language Models (LLMs), several attempts have been made to extend them to the visual domain. From having a visual assistant that could guide us through unfamiliar environments to generative models that produce
Externí odkaz:
http://arxiv.org/abs/2405.17247
Vision-language models enable open-world classification of objects without the need for any retraining. While this zero-shot paradigm marks a significant advance, even today's best models exhibit skewed performance when objects are dissimilar from th
Externí odkaz:
http://arxiv.org/abs/2404.16717
Autor:
Kirichenko, Polina, Ibrahim, Mark, Balestriero, Randall, Bouchacourt, Diane, Vedantam, Ramakrishna, Firooz, Hamed, Wilson, Andrew Gordon
Data augmentation (DA) encodes invariance and provides implicit regularization critical to a model's performance in image classification tasks. However, while DA improves average accuracy, recent studies have shown that its impact can be highly class
Externí odkaz:
http://arxiv.org/abs/2401.01764
Autor:
Eastwood, Cian, von Kügelgen, Julius, Ericsson, Linus, Bouchacourt, Diane, Vincent, Pascal, Schölkopf, Bernhard, Ibrahim, Mark
Self-supervised representation learning often uses data augmentations to induce some invariance to "style" attributes of the data. However, with downstream tasks generally unknown at training time, it is difficult to deduce a priori which attributes
Externí odkaz:
http://arxiv.org/abs/2311.08815
Autor:
Pezeshki, Mohammad, Bouchacourt, Diane, Ibrahim, Mark, Ballas, Nicolas, Vincent, Pascal, Lopez-Paz, David
Environment annotations are essential for the success of many out-of-distribution (OOD) generalization methods. Unfortunately, these are costly to obtain and often limited by human annotators' biases. To achieve robust generalization, it is essential
Externí odkaz:
http://arxiv.org/abs/2309.16748
Autor:
Bordes, Florian, Shekhar, Shashank, Ibrahim, Mark, Bouchacourt, Diane, Vincent, Pascal, Morcos, Ari S.
Synthetic image datasets offer unmatched advantages for designing and evaluating deep neural networks: they make it possible to (i) render as many data samples as needed, (ii) precisely control each scene and yield granular ground truth labels (and c
Externí odkaz:
http://arxiv.org/abs/2308.03977