Some recent work on neural formal theorem proving (already mentioned in the question, but the examples give a sense of the state of the art):
Steve Huntsman
- 15.4k
- 7
- 75
- 130