Three Modern Roles for Logic in AI
arXiv:2004.08599v158 citations
AI Analysis
This work addresses foundational challenges in AI for researchers and practitioners, but it appears incremental as it builds on existing theory without new empirical gains.
The paper tackles the problem of integrating logic into AI by proposing three roles based on tractable Boolean circuits, including computation, learning from data and knowledge, and reasoning about ML systems, but does not report specific results or numbers.
We consider three modern roles for logic in artificial intelligence, which are based on the theory of tractable Boolean circuits: (1) logic as a basis for computation, (2) logic for learning from a combination of data and knowledge, and (3) logic for reasoning about the behavior of machine learning systems.