Discovering Universal Geometry in Embeddings with ICA

05/22/2023
by   Hiroaki Yamagiwa, et al.
0

This study employs Independent Component Analysis (ICA) to uncover universal properties of embeddings of words or images. Our approach extracts independent semantic components of embeddings, enabling each embedding to be represented as a composition of intrinsic interpretable axes. We demonstrate that embeddings can be expressed as a combination of a few axes and that these semantic axes are consistent across different languages, modalities, and embedding algorithms. This discovery of universal properties in embeddings contributes to model interpretability, potentially facilitating the development of highly interpretable models and the compression of large-scale models.

READ FULL TEXT

page 1

page 2

page 6

page 7

page 21

research
12/19/2022

Independent Components of Word Embeddings Represent Semantic Features

Independent Component Analysis (ICA) is an algorithm originally develope...
research
01/21/2018

A Universal Semantic Space

Multilingual embeddings build on the success of monolingual embeddings a...
research
07/06/2021

Enhanced Universal Dependency Parsing with Automated Concatenation of Embeddings

This paper describes the system used in submission from SHANGHAITECH tea...
research
04/29/2015

On the universal structure of human lexical semantics

How universal is human conceptual structure? The way concepts are organi...
research
10/24/2022

On Universality of the S Combinator

In combinatory logic it is known that the set of two combinators K and S...
research
10/23/2020

Adversarial Learning of Feature-based Meta-Embeddings

Certain embedding types outperform others in different scenarios, e.g., ...
research
08/17/2022

Visual Comparison of Language Model Adaptation

Neural language models are widely used; however, their model parameters ...

Please sign up or login with your details

Forgot password? Click here to reset