A closer look at adaptive regret
- Submitting institution
-
Royal Holloway and Bedford New College
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 29004100
- Type
- D - Journal article
- DOI
-
-
- Title of journal
- Journal of Machine Learning Research
- Article number
- -
- First page
- 1
- Volume
- 17
- Issue
- 23
- ISSN
- 1533-7928
- Open access status
- Out of scope for open access requirements
- Month of publication
- April
- Year of publication
- 2016
- 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
-
3
- Research group(s)
-
-
- Citation count
- 1
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This paper belongs to a strand of machine learning that is robust in the sense of making no statistical assumptions about the data. It establishes definitive results for the performance of prediction algorithms with low adaptive regret giving two very different constructions of the optimal algorithm (Fixed Share, which was designed earlier for a different purpose). Interestingly, these results are not only mathematically elegant but subsume the main results in the important field of tracking the best expert. They have been used and developed by numerous researchers in prediction with expert advice.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -