Doubly robust Bayesian inference for non-stationary streaming data with β-divergences
- Submitting institution
-
The University of Warwick
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 5966
- Type
- E - Conference contribution
- DOI
-
-
- Title of conference / published proceedings
- Thirty-second Conference on Neural Information Processing Systems
- First page
- 64
- Volume
- 31
- Issue
- -
- ISSN
- -
- Open access status
- -
- Month of publication
- September
- Year of publication
- 2018
- 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
-
2
- Research group(s)
-
D - Data Science, Systems and Security
- Citation count
- -
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Published in the top machine learning conference, this paper led to invited talks at leading universities (Columbia, Cornell, Duke, NYU, Oxford, Sydney) and industry (Amazon, Facebook). Together with follow-up work by the same authors (arXiv:1904.02063, provisionally accepted by JMLR), it has opened a new research direction on robustness at the intersection of information geometry and Bayesian inference (e.g. Kuśmierczyk et al., NeurIPS'19; Loaiza-Ganem and Cunningham, NeurIPS'19; Jankowiak et al., ICML'20). It has also impacted UAI and ICPP work by Wild’s group at the Argonne National Laboratory, and been adopted by Uber AI Labs in their GPyTorch codebase (Martin Jankowiak, mjankowi@broadinstitute.org).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -