S Bubeck, M Sellke - Advances in Neural Information …, 2021 - proceedings.neurips.cc
Classically, data interpolation with a parametrized model class is possible as long as the number of parameters is larger than the number of equations to be satisfied. A puzzling …
Cellular sheaves equip graphs with a``geometrical''structure by assigning vector spaces and linear maps to nodes and edges. Graph Neural Networks (GNNs) implicitly assume a graph …
What does a neural network encode about a concept as we traverse through the layers? Interpretability in machine learning is undoubtedly important, but the calculations of neural …
K Zhang, W Zuo, L Zhang - IEEE Transactions on Image …, 2018 - ieeexplore.ieee.org
Due to the fast inference and good performance, discriminative learning methods have been widely studied in image denoising. However, these methods mostly learn a specific model …
This paper seeks to answer the question: as the (near-) orthogonality of weights is found to be a favorable property for training deep convolutional neural networks, how can we enforce …
H Xia, V Suliafu, H Ji, T Nguyen… - Advances in …, 2021 - proceedings.neurips.cc
We propose heavy ball neural ordinary differential equations (HBNODEs), leveraging the continuous limit of the classical momentum accelerated gradient descent, to improve neural …
This paper develops the FastRNN and FastGRNN algorithms to address the twin RNN limitations of inaccurate training and inefficient prediction. Previous approaches have …
M Zhang, Y Liu, H Luan, M Sun - … of the 55th Annual Meeting of …, 2017 - aclanthology.org
Word embeddings are well known to capture linguistic regularities of the language on which they are trained. Researchers also observe that these regularities can transfer across …
M Lezcano-Casado… - … Conference on Machine …, 2019 - proceedings.mlr.press
We introduce a novel approach to perform first-order optimization with orthogonal and unitary constraints. This approach is based on a parametrization stemming from Lie group …