Unified Inductive Logic: From Formal Learning to Statistical Inference to Supervised Learning

TL;DR

This paper proposes a Peircean alternative to Carnapian inductive logic, unifying formal learning, statistics, and supervised machine learning under a common inferential principle.

Contribution

It introduces a unifying principle for evaluating non-deductive inferences across formal learning, statistics, and supervised learning, based on Peircean philosophy.

Findings

Unifies formal learning theory, statistics, and supervised learning

Provides a common justification for non-deductive inference standards

Bridges philosophical perspectives with practical machine learning methods

Abstract

While the traditional conception of inductive logic is Carnapian, I develop a Peircean alternative and use it to unify formal learning theory, statistics, and a significant part of machine learning: supervised learning. Some crucial standards for evaluating non-deductive inferences have been assumed separately in those areas, but can actually be justified by a unifying principle.