Related: a recent paper from Antoine Bosselut and Yejin Choi explores dynamically constructing a context-specific, common sense knowledge graph using a transformer, in the context of question answering.
Big fan of this work. The headline is only slightly misleading: you'll still need a CUDA-enabled GPU to generate text at a reasonably interactive speed. But you don't have to fine tune a language model beforehand, which is a huge win. Plus, you can use whatever attribute model you'd like, provided you can get a gradient from it.
https://arxiv.org/abs/1911.03876