Zobrazeno 1 - 10
of 324
pro vyhledávání: '"Hjelm, R. P."'
Autor:
Prato, Gabriele, Song, Yale, Rajendran, Janarthanan, Hjelm, R Devon, Joshi, Neel, Chandar, Sarath
Transformers have become one of the dominant architectures in the field of computer vision. However, there are yet several challenges when applying such architectures to video data. Most notably, these models struggle to model the temporal patterns o
Externí odkaz:
http://arxiv.org/abs/2211.14449
Autor:
Fedorov, Alex, Geenjaar, Eloy, Wu, Lei, Sylvain, Tristan, DeRamus, Thomas P., Luck, Margaux, Misiura, Maria, Hjelm, R Devon, Plis, Sergey M., Calhoun, Vince D.
Recent neuroimaging studies that focus on predicting brain disorders via modern machine learning approaches commonly include a single modality and rely on supervised over-parameterized models.However, a single modality provides only a limited view of
Externí odkaz:
http://arxiv.org/abs/2209.02876
Autor:
Hjelm, R Devon, Mazoure, Bogdan, Golemo, Florian, Kahou, Samira Ebrahimi, Braga, Pedro, Frujeri, Felipe, Jalobeanu, Mihai, Kolobov, Andrey
A broad challenge of research on generalization for sequential decision-making tasks in interactive environments is designing benchmarks that clearly landmark progress. While there has been notable headway, current benchmarks either do not provide su
Externí odkaz:
http://arxiv.org/abs/2203.10351
Autor:
Chuang, Ching-Yao, Hjelm, R Devon, Wang, Xin, Vineet, Vibhav, Joshi, Neel, Torralba, Antonio, Jegelka, Stefanie, Song, Yale
Contrastive learning relies on an assumption that positive pairs contain related views, e.g., patches of an image or co-occurring multimodal signals of a video, that share certain underlying information about an instance. But what if this assumption
Externí odkaz:
http://arxiv.org/abs/2201.04309
A highly desirable property of a reinforcement learning (RL) agent -- and a major difficulty for deep RL approaches -- is the ability to generalize policies learned on a few tasks over a high-dimensional observation space to similar tasks not seen du
Externí odkaz:
http://arxiv.org/abs/2106.02193
Autor:
Hosseini, Arian, Reddy, Siva, Bahdanau, Dzmitry, Hjelm, R Devon, Sordoni, Alessandro, Courville, Aaron
Negation is a core construction in natural language. Despite being very successful on many tasks, state-of-the-art pre-trained language models often handle negation incorrectly. To improve language models in this regard, we propose to augment the lan
Externí odkaz:
http://arxiv.org/abs/2105.03519
Zero-shot classification is a generalization task where no instance from the target classes is seen during training. To allow for test-time transfer, each class is annotated with semantic information, commonly in the form of attributes or text descri
Externí odkaz:
http://arxiv.org/abs/2010.13320
Autor:
Baratin, Aristide, George, Thomas, Laurent, César, Hjelm, R Devon, Lajoie, Guillaume, Vincent, Pascal, Lacoste-Julien, Simon
We approach the problem of implicit regularization in deep learning from a geometrical viewpoint. We highlight a regularization effect induced by a dynamical alignment of the neural tangent features introduced by Jacot et al, along a small number of
Externí odkaz:
http://arxiv.org/abs/2008.00938
Autor:
Hjelm, R Devon, Bachman, Philip
Self-supervised learning has made unsupervised pretraining relevant again for difficult computer vision tasks. The most effective self-supervised methods involve prediction tasks based on features extracted from diverse views of the data. DeepInfoMax
Externí odkaz:
http://arxiv.org/abs/2007.13278
Autor:
Schwarzer, Max, Anand, Ankesh, Goel, Rishab, Hjelm, R Devon, Courville, Aaron, Bachman, Philip
While deep reinforcement learning excels at solving tasks where large amounts of data can be collected through virtually unlimited interaction with the environment, learning from limited interaction remains a key challenge. We posit that an agent can
Externí odkaz:
http://arxiv.org/abs/2007.05929