Details of Grant 

EPSRC Reference:	GR/R86188/01
Title:	Unsupervised learning for classification problems: A theoretical approach
Principal Investigator:	Goldberg, Professor P
Other Investigators:
Researcher Co-Investigators:
Project Partners:
Department:	Computer Science
Organisation:	University of Warwick
Scheme:	Standard Research (Pre-FEC)
Starts:	14 October 2002	Ends:	13 July 2006	Value (£):	61,200
EPSRC Research Topic Classifications:	Artificial Intelligence Image & Vision Computing Information & Knowledge Mgmt
EPSRC Industrial Sector Classifications:
Related Grants:
Panel History:	Panel Date Panel Name Outcome 30 Jan 2002 Software Technologies 30 & 31 January 2002 Deferred
Summary on Grant Application Form
Classification problems are about finding good rules for assigning labels to inputs, for example in hand-written digit recognition the problem is to label an image of a hand-written digit with the number it is intended to represent. In the context of machine learning, there are advantages to dividing labelled training data into sets of inputs having the same label, then fitting probability distributions to those sets, and then using maximum likelihood as a rule for assigning class labels. There is also a drawback - the practice of partitioning the data in this way ignores relevant information, in particular, for each label, set, we no longer know what other sets we are trying to distinguish it from.Computational learning theory addresses questions of where lie the fundamental obstacles (computational hardness, insufficiency of information) in machine earning problems. Negative results can be interpreted in a constructive way as indicators of what assumptions should be made about the data generation process in order to learning to be feasible. We propose to address the question of what kinds of classification problem become much harder if the separate classes are made available to separate algorithms whose results are to be combined subsequently. We also consider naturally associated topics such as algorithm design in cases where the problem appears to be tractable.
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.warwick.ac.uk