Abstract We introduce ART, a new corpus-level autoencoding approach for training dense retrieval models that does not require any labeled data. Dense is central challenge open-domain tasks, such as Open QA, where state-of-the-art methods typically large supervised datasets with custom hard-negative mining and denoising of positive examples. in contrast, only requires access to unpaired inputs o...

Nowadays there is a big amount of biomedical literature which uses complex nouns and acronyms of biological entities thus complicating the task of retrieval specific information. The Genomics Track works for this goal and this paper describes the approach we used to take part of this track of TREC 2007. As this is the first time we participate in this track, we configurated a new system consist...

We compare the effect of different segmentation strategies for passage retrieval of user generated internet video. We consider retrieval of passages for rather abstract and complex queries that go beyond finding a certain object or constellation of objects in the visual channel. Hence the retrieval methods have to rely heavily on the recognized speech. Passage retrieval has mainly been studied ...

We describe a commercial question-answering system that uses AI – specifically, the Cyc system – to support passage retrieval and perform deductive QA, to produce results superior to what each question-answering technique could produce alone.

One of the NLM experimental approaches to the 2007 Genomics track question answering task followed the track evaluation design: we attempted identifying exact answers in the form of semantic relations between biomedical entities named in questions and the potential answer types and then marked the passages containing the relations as containing the answers. The goal of this knowledgebased appro...

ii I hereby declare that I am the sole author of this thesis. I authorize the University of Waterloo to lend this thesis to other institutions or individuals for the purpose of scholarly research. Signature I further authorize the University of Waterloo to reproduce this thesis by photocopying or by other means, in total or in part, at the request of other institutions or individuals for the pu...

Information retrieval systems can attempt to answer the user’s query directly, by extracting an appropriate passage of text from a corpus and presenting it on the results page. However, sometimes the passage of text contains extraneous information, or multiple passages are needed to form an answer. In cases like these, some sort of answer distillation system could be useful, taking as input the...

TREC’06 genomics track was focusing on text mining and passage retrieval. WIM lab participated in this year’s TREC genomics track. Our system consists of five parts: preprocessing, sentence generation, document retrieval, answer extraction and answer fusion. And we developed two different method: a automated profile matchingbased method and a text categorizationbased method to do the text minin...

We propose a framework for discriminative Information Retrieval (IR) atop linguistic features, trained to improve the recall of tasks such as answer candidate passage retrieval, the initial step in text-based Question Answering (QA). We formalize this as an instance of linear feature-based IR (Metzler and Croft, 2007), illustrating how a variety of knowledge discovery tasks are captured under t...

This paper describes the architecture, operation and results obtained with the Question Answering prototype developed in the Department of Language Processing and Information Systems at the University of Alicante. Our system is based on our TREC-9 approach where different improvements have been introduced. Essentially these modifications are twofold: the introduction of a passage retrieval modu...