We suggest a simple Gaussian mixture model for data generation that comp...
We choose random points in the hyperbolic disc and claim that these poin...
Softmax is the de facto standard in modern neural networks for language
...
There is an ongoing debate in the NLP community whether modern language
...
We show analytically that removing sigmoid transformation in the SGNS
ob...
We show that the skip-gram embedding of any word can be decomposed into ...
We critically review the smooth inverse frequency sentence embedding met...
We perform an empirical evaluation of several methods of low-rank
approx...
This paper takes a step towards theoretical analysis of the relationship...
We review neural network architectures which were motivated by Fourier s...
We propose several ways of reusing subword embeddings and other weights ...
Words in some natural languages can have a composite structure. Elements...
Syllabification does not seem to improve word-level RNN language modelin...