Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2406.10256 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909224554987520 |
|---|---|
| author | Andonov, Jovan Ganea, Octavian Grnarova, Paulina Bécigneul, Gary Hofmann, Thomas |
| author_facet | Andonov, Jovan Ganea, Octavian Grnarova, Paulina Bécigneul, Gary Hofmann, Thomas |
| contents | Language Modelling has been a central part of Natural Language Processing for a very long time and in the past few years LSTM-based language models have been the go-to method for commercial language modeling. Recently, it has been shown that when looking at language modelling from a matrix factorization point of view, the final Softmax layer limits the expressiveness of the model, by putting an upper bound on the rank of the resulting matrix. Additionally, a new family of neural networks based called NeuralODEs, has been introduced as a continuous alternative to Residual Networks. Moreover, it has been shown that there is a connection between these models and Normalizing Flows. In this work we propose a new family of language models based on NeuralODEs and the continuous analogue of Normalizing Flows and manage to improve on some of the baselines. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_10256 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Explicit Word Density Estimation for Language Modelling Andonov, Jovan Ganea, Octavian Grnarova, Paulina Bécigneul, Gary Hofmann, Thomas Computation and Language Artificial Intelligence Machine Learning Language Modelling has been a central part of Natural Language Processing for a very long time and in the past few years LSTM-based language models have been the go-to method for commercial language modeling. Recently, it has been shown that when looking at language modelling from a matrix factorization point of view, the final Softmax layer limits the expressiveness of the model, by putting an upper bound on the rank of the resulting matrix. Additionally, a new family of neural networks based called NeuralODEs, has been introduced as a continuous alternative to Residual Networks. Moreover, it has been shown that there is a connection between these models and Normalizing Flows. In this work we propose a new family of language models based on NeuralODEs and the continuous analogue of Normalizing Flows and manage to improve on some of the baselines. |
| title | Explicit Word Density Estimation for Language Modelling |
| topic | Computation and Language Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2406.10256 |