Differentiable Representations For Multihop Inference Rules
This work addresses the challenge of efficient multi-hop inference for AI systems handling large-scale symbolic knowledge, though it is incremental as it builds on existing neural and symbolic methods.
The paper tackles the problem of scaling second-order multi-hop reasoning to large knowledge bases by introducing a differentiable operation for constructing reasoning templates, achieving competitive performance on tasks with millions of entities and tens of millions of triples.
We present efficient differentiable implementations of second-order multi-hop reasoning using a large symbolic knowledge base (KB). We introduce a new operation which can be used to compositionally construct second-order multi-hop templates in a neural model, and evaluate a number of alternative implementations, with different time and memory trade offs. These techniques scale to KBs with millions of entities and tens of millions of triples, and lead to simple models with competitive performance on several learning tasks requiring multi-hop reasoning.