Learning Logical Rules using Minimum Message Length
This work addresses a key challenge in AI for researchers and practitioners by providing a more data-efficient and robust method for learning logical rules, though it appears incremental as it builds on existing minimum description length approaches.
The paper tackles the challenge of unifying probabilistic and logical learning by introducing a Bayesian inductive logic programming approach that learns minimum message length programs from noisy data, achieving significant performance improvements over previous methods in domains like game playing and drug design.
Unifying probabilistic and logical learning is a key challenge in AI. We introduce a Bayesian inductive logic programming approach that learns minimum message length programs from noisy data. Our approach balances hypothesis complexity and data fit through priors, which explicitly favour more general programs, and a likelihood that favours accurate programs. Our experiments on several domains, including game playing and drug design, show that our method significantly outperforms previous methods, notably those that learn minimum description length programs. Our results also show that our approach is data-efficient and insensitive to example balance, including the ability to learn from exclusively positive examples.