Denotational validation of higher-order Bayesian inference
- Submitting institution
-
University of Edinburgh
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 84913447
- Type
- D - Journal article
- DOI
-
10.1145/3158148
- Title of journal
- Proceedings of the ACM on Programming Languages
- Article number
- 60
- First page
- -
- Volume
- 2
- Issue
- POPL
- ISSN
- 2475-1421
- Open access status
- Compliant
- Month of publication
- December
- Year of publication
- 2017
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
9
- Research group(s)
-
C - Foundations of Computation
- Citation count
- -
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Develops semantics for Bayesian inference algorithms underlying probabilistic programming languages, that are, for the first time, both rigorous and modular. POPL'18 acceptance rate 24%. Validates rigorously sophisticated inference algorithms commonly used in data science and machine learning, using compositional building blocks: sequential Monte Carlo, Markov Chain Monte Carlo. Enabled our later work (ICFP'18) as foundation for a performant modular implementation of inference representations. Enabled our later work (POPL'19) which reuses both the semantics and synthetic measure theory. The approach and model we developed was later used by others, e.g. Lew et al. (POPL'20) to validate some programmable inference algorithms.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -