Referential Uncertainty and Word Learning in High-dimensional, Continuous Meaning Spaces

TL;DR

This paper evaluates how machine learning algorithms handle high-dimensional continuous meaning spaces in word learning, showing that increased dimensions linearly affect performance and referential uncertainty has minimal impact.

Contribution

It demonstrates that current machine learning techniques effectively manage high-dimensional meaning spaces and that referential uncertainty from word sensitivity is negligible.

Findings

Performance scales linearly with dimensionality

Referential uncertainty has minimal impact

ML techniques are effective in high-dimensional spaces

Abstract

This paper discusses lexicon word learning in high-dimensional meaning spaces from the viewpoint of referential uncertainty. We investigate various state-of-the-art Machine Learning algorithms and discuss the impact of scaling, representation and meaning space structure. We demonstrate that current Machine Learning techniques successfully deal with high-dimensional meaning spaces. In particular, we show that exponentially increasing dimensions linearly impact learner performance and that referential uncertainty from word sensitivity has no impact.