Revisiting direct neuralisation of first-order logic




There is a long history of direct translation of propositional Horn-clause logic programs into neural networks. The possibility of translating first-order logical syntax in the same way has been largely overlooked, perhaps due to a “propositional fixation” fixation! We briefly revise the possibility and advantage of translating existentially and universally quantified clauses into a neural form that follows the first-order syntax in a natural way

This project has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No 731593.