In our system, noun phrases in the query are identified and classified into 4 types: proper names, dictionary phrases, simple phrases and complex phrases. A document has a phrase if all content words in a phrase are within a window of a certain size. The window sizes for different types of phrases are different. We consider phrases to be more important than individual terms. As a consequence, d...

This paper describes the methods we developed for the three tasks of the TREC Genomics Track, i.e., ad hoc retrieval, triage, and annotation tasks. For the ad hoc retrieval task, we used the classic vector space model and studied the use of query expansion and pseudorelevance feedback. Our submitted runs obtained a MAP of 0.183. For the triage task, we adopted a naı̈ve Bayes classifier trained o...

We provide a description of the QACTIS question-answering system and its application to and performance in the 2004 TREC question-answering evaluation. Since this was also QACTIS’s first year competing at TREC, we provide a complete overview of its purpose, development, structure, and its future directions.

This paper reports our results at DUC2004 and describes our approach, implemented in a system called Topiary. We will show that the combination of linguistically motivated sentence compression with statistically selected topic terms performs better than either alone, according to some automatic summary evaluation measures.

This paper describes the participation of the School of Informatics, University of Wales, Bangor in the 2004 Text Retrieval Conference. We present additions and modi cations to the QITEKAT system, initially developed as an entry for the 2003 QA evaluation, including automated regular expression induction, improved question matching, and application of our knowledge framework to the modi ed ques...

This paper describes the INEX 2004 participation of the Informatics Institute of the University of Amsterdam. We completely revamped our XML retrieval system, now implemented as a mixture language model on top of a standard search engine. To speed up structural reasoning, we indexed the collection’s structure in a separate database. We address three research questions. First, we investigate the...

We present the results of Michigan’s participation in DUC 2004. Our system, MEAD, ranked as one of the top systems in four of the five tasks. We introduce our new feature, LexPageRank, a new measure of sentence centrality inspired by the prestige concept in social networks. LexPageRank gave promising results in multi-document summarization. Our approach for Task 5, biographical summarization, w...

DIMACS participated in the text categorization and ad hoc retrieval tasks of the TREC 2004 Genomics track. For the categorization task, we tackled the triage and annotation hierarchy subtasks. 1. TEXT CATEGORIZATION TASK The Mouse Genome Informatics (MGI) project of the Jackson Laboratory provides data on the genetics, genomics, and biology of the laboratory mouse. In particular, the Mouse Geno...

A topic tracking system that combines elements from vector space and language modeling frameworks to compute document scores is described. The model is used for both the traditional TDT topic tracking evaluation design and the new supervised adaptive topic tracking evaluation. Results indicate that supervised adaptation and score normalization should be more closely coupled, and that current te...

Our last years Soccer Simulation team AT Humboldt 2003 was mainly designed as a research testbed for long-term deliberation and realtime reasoning, cooperative behavior in MAS, CBR-aided decision making and scalable behavior control mechanisms. The main goal was to implement and to evaluate the double pass architecture (DPA). This was always done with regard to the transition of the Simulation ...