Estimation of embedding vectors in high dimensions

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Azar, Golara Ahmadi, Emami, Melika, Fletcher, Alyson, Rangan, Sundeep
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912317790224384
author Azar, Golara Ahmadi
Emami, Melika
Fletcher, Alyson
Rangan, Sundeep
author_facet Azar, Golara Ahmadi
Emami, Melika
Fletcher, Alyson
Rangan, Sundeep
contents Embeddings are a basic initial feature extraction step in many machine learning models, particularly in natural language processing. An embedding attempts to map data tokens to a low-dimensional space where similar tokens are mapped to vectors that are close to one another by some metric in the embedding space. A basic question is how well can such embedding be learned? To study this problem, we consider a simple probability model for discrete data where there is some "true" but unknown embedding where the correlation of random variables is related to the similarity of the embeddings. Under this model, it is shown that the embeddings can be learned by a variant of low-rank approximate message passing (AMP) method. The AMP approach enables precise predictions of the accuracy of the estimation in certain high-dimensional limits. In particular, the methodology provides insight on the relations of key parameters such as the number of samples per value, the frequency of the terms, and the strength of the embedding correlation on the probability distribution. Our theoretical findings are validated by simulations on both synthetic data and real text data.
format Preprint
id arxiv_https___arxiv_org_abs_2312_07802
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Estimation of embedding vectors in high dimensions
Azar, Golara Ahmadi
Emami, Melika
Fletcher, Alyson
Rangan, Sundeep
Machine Learning
Information Theory
Embeddings are a basic initial feature extraction step in many machine learning models, particularly in natural language processing. An embedding attempts to map data tokens to a low-dimensional space where similar tokens are mapped to vectors that are close to one another by some metric in the embedding space. A basic question is how well can such embedding be learned? To study this problem, we consider a simple probability model for discrete data where there is some "true" but unknown embedding where the correlation of random variables is related to the similarity of the embeddings. Under this model, it is shown that the embeddings can be learned by a variant of low-rank approximate message passing (AMP) method. The AMP approach enables precise predictions of the accuracy of the estimation in certain high-dimensional limits. In particular, the methodology provides insight on the relations of key parameters such as the number of samples per value, the frequency of the terms, and the strength of the embedding correlation on the probability distribution. Our theoretical findings are validated by simulations on both synthetic data and real text data.
title Estimation of embedding vectors in high dimensions
topic Machine Learning
Information Theory
url https://arxiv.org/abs/2312.07802