Learning probabilistic termination proofs
We present the first machine learning approach to the termination analysis of probabilistic programs. Ranking supermartingales (RSMs) prove that probabilistic programs halt, in expectation, within a finite number of steps. While previously RSMs were directly synthesised from source code, our method...
| Hlavní autoři: | , , |
|---|---|
| Médium: | Conference item |
| Jazyk: | English |
| Vydáno: |
Springer Nature
2021
|