DRESS

Curses and blessings of dimensionality

This project "Dimensionality Reduction for Efficient Similarity Search (DRESS)" is funded in part by EPSRC (EP/E037402/1) and Australian Research Council (DP0663272). A wide range of applications, such as multimedia retrieval, molecular biology, medical imaging, and so on, involve similarity search in high-dimensional feature spaces, which is often computational expensive if not prohibitive. Traditional index structures in relational databases and multi-dimensional index structures in spatial databases are not sufficient to index high-dimensional data. This project aims to improve the efficiency of similarity search by exploring linear and non-linear dimensionality reduction techniques and adapting them to different similarity measures, feature spaces and query processing strategies.

Lau, R., Bruza, P. and Song, D. (2007) Towards a Belief Revision Based Adaptive and Context Sensitive Information Retrieval System, Accepted by ACM Transactions on Information Systems (TOIS)

Huang, Z., Shen, H., Zhou, X., Song, D. and Rueger, S. (2007) Dimensionality Reduction for Dimension-specific Search, Poster at The 30th Annual International ACM SIGIR Conference (SIGIR'2007), pp. 849-850

Song, D., Lau, R., Bruza, P., Wong, K. and Chen, D. (2007) An Adaptive Information Agent for Document Title Classification and Filtering in Data Intensive Domains, Decision Support Systems, 44, pp. 251-265, Elsevier

Song, D., Cao, G., Bruza, P. and Lau, R. (2007) Concept induction via fuzzy C-means clustering in a high-dimensional semantic space, in eds. J. Valente de Oliverira and W. Pedrycz, Advances in Fuzzy Clustering and its Applications, pp. 393-403, John Wiley & Sons