Short question
How can I implement Logic Inference with Deep Learning?
Long question
Based on Symbolic Logic, chaining multiple predicates (a short example is Syllogism) is a method of implementing Logic Inference. This programmatic way is suitable for many cases, however, if combined with NLP, too much programmatic way is involved.
If someone implements that Logic Inference with not programmatic way but machine learning, what model should be adopted, and what label data should be input to the model?
