New Bounds on Compressive Linear Least Squares Regression
- Submitting institution
-
The University of Birmingham
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 42748003
- Type
- E - Conference contribution
- DOI
-
-
- Title of conference / published proceedings
- 17th International Conference on Artificial Intelligence and Statistics (AISTATS)
- First page
- 448
- Volume
- -
- Issue
- -
- ISSN
- -
- Open access status
- -
- Month of publication
- April
- Year of publication
- 2014
- 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
-
0
- Research group(s)
-
-
- Citation count
- 16
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This paper introduces a new strategy to bound the error of learning from compressive data, which allows better theoretical guarantees under more general conditions than previous work. This is significant, as the main idea of exploiting the structure of the problem is widely applicable. It helped winning an EPSRC fellowship grant; related results appear in (Kaban, AAAI'19) and (Kaban & Durrant, Journal of Artificial Intelligence Research 2020). It is a theoretical paper which instigated new work in both theoretical and practical areas from well-known research
groups, including Cambridge Statistical Laboratory, Mark Rudelson in Michigan, Meinshausen's group in Switzerland, and others.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -