Inference of Abstraction for a Unified Account of Reasoning and Learning
This work addresses a foundational challenge in AI by aiming to integrate reasoning and learning, but it appears incremental as it builds on existing Bayesian and logical frameworks without clear SOTA advancements.
The paper tackles the problem of unifying reasoning and learning by proposing a probabilistic inference theory based on abstraction, where data cause symbolic knowledge through satisfiability in formal logic. It demonstrates theoretical correctness using logical consequence relations and empirical correctness on the MNIST dataset, though no concrete performance numbers are provided.
Inspired by Bayesian approaches to brain function in neuroscience, we give a simple theory of probabilistic inference for a unified account of reasoning and learning. We simply model how data cause symbolic knowledge in terms of its satisfiability in formal logic. The underlying idea is that reasoning is a process of deriving symbolic knowledge from data via abstraction, i.e., selective ignorance. The logical consequence relation is discussed for its proof-based theoretical correctness. The MNIST dataset is discussed for its experiment-based empirical correctness.