| EPSRC Reference: |
GR/L57586/01 |
| Title: |
STATISTICAL LEARNING METHODOLOGY IN NEURAL COMPUTING PROBLEMS |
| Principal Investigator: |
Titterington, Professor DM |
| Other Investigators: |
|
| Researcher Co-Investigators: |
|
| Project Partners: |
|
| Department: |
Statistics |
| Organisation: |
University of Glasgow |
| Scheme: |
Standard Research (Pre-FEC) |
| Starts: |
01 November 1997 |
Ends: |
31 October 2000 |
Value (£): |
131,032
|
| EPSRC Research Topic Classifications: |
| Statistics & Appl. Probability |
|
|
| EPSRC Industrial Sector Classifications: |
| No relevance to Underpinning Sectors |
|
|
| Related Grants: |
|
| Panel History: |
|
|
Summary on Grant Application Form |
|
The project concerns likelihood and Bayesian methods for developing learning rules for models involving hidden variables. Such models are currently popular in neural-computing but are also familiar in the statistical literature. In practice, there are difficulties with standard versions of procedures such as the EM algorithm, and one approximating method is to use Mean-field approximations in the E-step. It is intended to provide theoretical underpinning of such modifications on a more secure footing. It is also proposed to investigate variations on the basic Mean-field approach that should improve the performance and efficiency of the algorithms.Mean-field approximations have also been used to create approximating functions for likelihoods and posterior distributions in some neural-computing contexts. It is proposed to investigate these implementations systematically, to compare their efficacy with competing approximations and to implement the approach in a wider class of statistical problems. In some hidden-variable problems for which maximum likelihood is intractable there exist alternative, inefficient estimators that are easily calculated. It is proposed to investigate the properties of estimators generated from these inefficient ones by a one-stage iterative process.
|
|
Key Findings |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
|
Potential use in non-academic contexts |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
|
Impacts |
|
| Description |
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk |
| Summary |
|
| Date Materialised |
|
|
|
|
Sectors submitted by the Researcher |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
| Project URL: |
|
| Further Information: |
|
| Organisation Website: |
http://www.gla.ac.uk |