Published on:

8 February 2024

Primary Category:

Logic

Paper Authors:

Nikolay Bazhenov,

Ekaterina Fokina,

Dino Rossegger,

Alexandra Soskova,

Stefan Vatev

•

Defines textual learning of algebraic structures

•

Gives a model-theoretic characterization of textual learnability

•

Connects to prior work on learning from informant

•

Introduces hierarchy of positive infinitary sentences

•

Proves equivalence of textual learning and distinguishability by positive Sigma_2 sentences

Learning algebraic structure types from text

This paper introduces a formal framework to define learning algebraic structures from text. The key result shows a class of algebraic structures is learnable from text if and only if the structures can be distinguished by positive infinitary Sigma_2 sentences in their theories. This provides a model-theoretic characterization of textual learnability.

No comments yet, be the first to start the conversation...

Sign up to comment on this paper