Extends PAC semantics to implicitly learn universally quantified first-order clauses over countably infinite domains via symmetries for query answering.
Partial observability and learnability
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.AI 2years
2019 2verdicts
UNVERDICTED 2representative citing papers
Backward proof search algorithms oblivious to the knowledge base can be modified to learn PAC-supporting rules for a query, with applicability to chaining and resolution.
citing papers explorer
-
Implicitly Learning to Reason in First-Order Logic
Extends PAC semantics to implicitly learn universally quantified first-order clauses over countably infinite domains via symmetries for query answering.
-
Query-driven PAC-Learning for Reasoning
Backward proof search algorithms oblivious to the knowledge base can be modified to learn PAC-supporting rules for a query, with applicability to chaining and resolution.