Operational Latent Spaces

TL;DR

This paper explores the creation of semantically meaningful latent spaces through self-supervised learning, enabling operations like clustering and transformations, with applications including musical symmetries.

Contribution

It introduces the concept of operational latent spaces and a novel FiLMR layer for intentional construction of such spaces with specific symmetries.

Findings

Operational latent spaces exhibit semantic clustering and transformations.

Unintended properties can emerge in self-supervised learning spaces.

The FiLMR layer enables ring-like symmetries in latent spaces.

Abstract

We investigate the construction of latent spaces through self-supervised learning to support semantically meaningful operations. Analogous to operational amplifiers, these "operational latent spaces" (OpLaS) not only demonstrate semantic structure such as clustering but also support common transformational operations with inherent semantic meaning. Some operational latent spaces are found to have arisen "unintentionally" in the progress toward some (other) self-supervised learning objective, in which unintended but still useful properties are discovered among the relationships of points in the space. Other spaces may be constructed "intentionally" by developers stipulating certain kinds of clustering or transformations intended to produce the desired structure. We focus on the intentional creation of operational latent spaces via self-supervised learning, including the introduction of rotation operators via a novel "FiLMR" layer, which can be used to enable ring-like symmetries found in some musical constructions.