Saved in:
Bibliographic Details
Main Authors: Opper, Mattia, Siddharth, N.
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2404.01860
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916626369085440
author Opper, Mattia
Siddharth, N.
author_facet Opper, Mattia
Siddharth, N.
contents This paper presents two simple improvements to the Self-Structuring AutoEncoder (Self-StrAE). Firstly, we show that including reconstruction to the vocabulary as an auxiliary objective improves representation quality. Secondly, we demonstrate that increasing the number of independent channels leads to significant improvements in embedding quality, while simultaneously reducing the number of parameters. Surprisingly, we demonstrate that this trend can be followed to the extreme, even to point of reducing the total number of non-embedding parameters to seven. Our system can be pre-trained from scratch with as little as 10M tokens of input data, and proves effective across English, Spanish and Afrikaans.
format Preprint
id arxiv_https___arxiv_org_abs_2404_01860
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Self-StrAE at SemEval-2024 Task 1: Making Self-Structuring AutoEncoders Learn More With Less
Opper, Mattia
Siddharth, N.
Computation and Language
This paper presents two simple improvements to the Self-Structuring AutoEncoder (Self-StrAE). Firstly, we show that including reconstruction to the vocabulary as an auxiliary objective improves representation quality. Secondly, we demonstrate that increasing the number of independent channels leads to significant improvements in embedding quality, while simultaneously reducing the number of parameters. Surprisingly, we demonstrate that this trend can be followed to the extreme, even to point of reducing the total number of non-embedding parameters to seven. Our system can be pre-trained from scratch with as little as 10M tokens of input data, and proves effective across English, Spanish and Afrikaans.
title Self-StrAE at SemEval-2024 Task 1: Making Self-Structuring AutoEncoders Learn More With Less
topic Computation and Language
url https://arxiv.org/abs/2404.01860