Deep learning languages: a key fundamental shift from probabilities to weights?
Other Quantitative Biology
2019-08-05 v1 Computation and Language
Machine Learning
Abstract
Recent successes in language modeling, notably with deep learning methods, coincide with a shift from probabilistic to weighted representations. We raise here the question of the importance of this evolution, in the light of the practical limitations of a classical and simple probabilistic modeling approach for the classification of protein sequences and in relation to the need for principled methods to learn non-probabilistic models.
Keywords
Cite
@article{arxiv.1908.00785,
title = {Deep learning languages: a key fundamental shift from probabilities to weights?},
author = {François Coste},
journal= {arXiv preprint arXiv:1908.00785},
year = {2019}
}