We report our participation to the TREC 2017 Core Track. Our objective is to systematically investigate the use of weak baselines, namely off-the-shelf open source IR systems, and understand how they compare with respect to TREC state-of-the-art ones. We are also interested in understanding how IR components – namely stop lists, stemmers/n-grams, and IR models – contribute to the overall perfor...

This paper reports on East Normal China University’s participation in the TREC 2015 LiveQA track. An overview is presented to introduce our community question answer system and discuss the technologies. This year, the Trec LiveQA track expands the traditional QA track, focusing on “live” question answering for the real-user questions. At this challenge, we built a real-time community question a...

Two sets of parameters are important in the retrieval model. One is the translation probabilities P(tq|td). In TREC 2001, we used model 1 of Brown’s statistical MT work (Brown et al, 1993) for estimating term translation probabilities from a parallel corpus due to efficiency considerations. With more computer power at disposal, for TREC 2002 we used the more complex but potentially more accurat...

This paper describes the main retrieval problems when facing with blogs. Using the classical tf idf vector-space model together with three probabilistic and one statistical language model, we evaluate them using a TREC test-collections composed of 100 topics. Using two performance measures, we show that ignoring a stemming approach results in a better performance than other indexing strategies ...

In the TREC-8 Interactive Track, our results indicated that the better performance obtained in batch searching evaluation do not translate into better performance by users in an instance recall task. This year we pursued this investigation further by performing the same experiments using the new questionanswering task adopted in the TREC-9 Interactive Track. Our results once again show that bet...

This paper presents a new version of the DIOGENE Question Answering (QA) system developed at ITCIrst. With respect to our first participation to the TREC QA main task (TREC-2001), the system presents both improvements and extensions. On one hand, significant improvements rely on the substitution of basic components (e.g. the search engine and the tool in charge of the named entities recognition...

This working note reports our experiences with TREC-8 on four tracks: Ad Hoc, Filtering, Web, and QA. The Ad Hoc retrieval engine, SCAIR, has been used for the Web and QA experiments, and the ltering experiments were based on it's own engine. As a second entry to TREC, we focused this year on exploring possibilities of applying machine learning techniques to TREC tasks. The Ad Hoc track employe...

Experiments performed on small collections suggest that expanding query vectors with words that are lexically related to the original query words can improve retrieval eeectiveness. Prior experiments using WordNet to automatically expand vectors in the large TREC-1 collection were inconclusive regarding eeec-tiveness gains from lexically related words since any such eeects were dominated by the...

A frozen 18.5 million page snapshot of part of the Web has been created to enable and encourage meaningful and reproducible evaluation of Web search systems and techniques. This collection is being used in an evaluation framework within the Text Retrieval Conference (TREC) and will hopefully provide convincing answers to questions such as, “Can link information result in better rankings?”, “Do ...

Web users often issue a series of related queries, which form a search session, to fulfill complicated information needs. Query aggregation utilizes and combines multiple queries for session search. Prior research has demonstrated that query aggregation is an effective technique for session search that can generally achieve a 15% ∼ 20% nDCG@10 improvement on finding relevant documents for the l...