A Computational Theory for Life-Long Learning of Semantics

TL;DR

This paper proposes a unified, online framework for incrementally learning semantic vectors across multiple data types using binary vectors, aiming to bridge the gap between data-driven and supervised learning models.

Contribution

It introduces a novel framework for dynamic, multi-medium, online semantic learning that integrates different learning paradigms into a unified model.

Findings

Framework enables incremental semantic learning across data types

Supports online, real-time updates of semantic vectors

Lays groundwork for future research in unified semantic models

Abstract

Semantic vectors are learned from data to express semantic relationships between elements of information, for the purpose of solving and informing downstream tasks. Other models exist that learn to map and classify supervised data. However, the two worlds of learning rarely interact to inform one another dynamically, whether across types of data or levels of semantics, in order to form a unified model. We explore the research problem of learning these vectors and propose a framework for learning the semantics of knowledge incrementally and online, across multiple mediums of data, via binary vectors. We discuss the aspects of this framework to spur future research on this approach and problem.