Browsing by Author "JAGADISH, H. V."

Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    Generic Inverted Index on the GPU
    (2015-11-23) ZHOU, Jingbo; GUO, Qi; JAGADISH, H. V.; LUAN, Wenhao; TUNG, Anthony K. H.; ZHENG, Yuxin
    Data variety, as one of the three Vs of the Big Data, is manifested by a growing number of complex data types such as documents, sequences, trees, graphs and high dimensional vectors. To perform similarity search on these data, existing works mainly choose to create customized indexes for different data types. Due to the diversity of customized indexes, it is hard to devise a general parallelization strategy to speed up the search. In this paper, we propose a generic inverted index on the GPU (called GENIE), which can support similarity search of multiple queries on various data types. GENIE can effectively support the approximate nearest neighbor search in different similarity measures through exerting Locality Sensitive Hashing schemes, as well as similarity search on original data such as short document data and relational data. Extensive experiments on different reallife datasets demonstrate the efficiency and effectiveness of our system.
  • No Thumbnail Available
    Item
    Object Semantics for XML Keyword Search
    (2013-05-21T01:18:54Z) LE, Thuy Ngoc; LING, Tok Wang; JAGADISH, H. V.; LIN, Chunbin; LU, Jiaheng
    We know that some XML elements correspond to objects (in the sense of object-orientation) and others do not. The question we consider in this paper is what benefits we can derive from paying attention to such object semantics, particularly for the problem of keyword queries. Keyword queries against XML data have been studied extensively in recent years, with several lowest-common-ancestor based schemes proposed for this purpose, including SLCA, MLCA, VLCA, and ELCA. It is easy to see that identifying objects can help each of these techniques return more meaningful answers than just the LCA node (or subtree). It is more interesting to see that object semantics can also be used to benefit the search itself. For this purpose, we introduce a novel nearest common object node semantics (NCON), which includes not just common ancestors but also common descendants and referenced objects in evaluating a query. We have developed XComplete, a system for our NCON-based approach, and used it in our extensive experimental evaluation. The experimental results show that our proposed approach outperforms the existing LCA-based approaches in terms of both effectiveness and efficiency.