By studying the effect of varying context on phrasal indexing techniques, she hopes to get a better understanding of the interaction between context and the linguistic phenomena of semantic compositionality and collocational constraint. Her work is also aimed at enhancing the more simplistic word-based techniques currently used in Information Retrieval.
Bill Hart presented a poster titled Efficient Global Optimization using Local Search with the GA. The Genetic Algorithm (GA) performs an adaptive global sampling of a search space. Researchers have found that the addition of local search operators can improve the rate at which the GA discoveries better solutions.
Bill's research examines factors which affect the efficiency of GA's that use local search operators. These factors include the frequency with which local search is used, the length of the local search and the sophistication of the local search. He presented results which illustrated how the local search frequency can affect the efficiency of the GA on a difficult testbed of functions.
Rik Belew -- Assistant Professor, CSE
Go back to June, 1993 Table of Contents