| EPSRC Reference: |
GR/L52093/01 |
| Title: |
LEARNING FIXED EXAMPLE SETS IN MULTI-LAYER NEURAL NETWORKS |
| Principal Investigator: |
Saad, Professor D |
| Other Investigators: |
|
| Researcher Co-Investigators: |
|
| Project Partners: |
|
| Department: |
Information Engineering |
| Organisation: |
Aston University |
| Scheme: |
Standard Research (Pre-FEC) |
| Starts: |
01 February 1998 |
Ends: |
30 April 2000 |
Value (£): |
106,613
|
| EPSRC Research Topic Classifications: |
|
| EPSRC Industrial Sector Classifications: |
| No relevance to Underpinning Sectors |
|
|
| Related Grants: |
|
| Panel History: |
|
|
Summary on Grant Application Form |
|
The successful application of artificial neural networks to various tasks has been overshadowed by the lack of clear theoretical analysis of the training process. In a previous study we investigated analytically different aspects of on-line learning, on of the most common and applicable training algorithms, presenting the only existing framework for analysing fully and exactly a learning scenario in multilayer networks. One of the most important aspects of on-line training that have not been addressed is the question of learning from fixed example sets, which play an important role in realistic training scenarios, resulting in overfitting and the need for regularising the training process. These phenomena do not show up in the current framework, limiting the impact of the analysis on realistic training scenarios. This project will focus on extending the current analysis to include on-line learning from fixed example sets, providing a link between equilibrium and non-equilibrium techniques and making a whole new set of questions related to real world learning accessible. We will concentrate on the effects of fixed training sets on local minima and overfitting, on the analytical investigation of current heuristic techniques for alleviating these problems and on devising new methods for improving training performance.
|
|
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.aston.ac.uk |