Filters








31,857 Hits in 2.6 sec

Parallel computing for passage retrieval

A. MacFarlane, S.E. Robertson, J.A. McCann
2004 ASLIB Proceedings  
In this paper we examine methods for both speeding up passage processing and examining more passages using parallel computers.  ...  We describe this algorithm and our mechanism for applying parallel computing to speed up the processing.  ...  We are particularly grateful to David Hawking for making the arrangements for the visit to the ANU.  ... 
doi:10.1108/00012530410549231 fatcat:whwsqime3nfqngv7cwxip2xlne

Retrieval-Augmented Multilingual Keyphrase Generation with Retriever-Generator Iterative Training [article]

Yifan Gao, Qingyu Yin, Zheng Li, Rui Meng, Tong Zhao, Bing Yin, Irwin King, Michael R. Lyu
2022 arXiv   pre-print
Moreover, we develop a retriever-generator iterative training algorithm to mine pseudo parallel passage pairs to strengthen the cross-lingual passage retriever.  ...  Given a non-English passage, a cross-lingual dense passage retrieval module finds relevant English passages.  ...  Then we start a loop to mine pseudo parallel passage pairs for refining the passage retriever.  ... 
arXiv:2205.10471v2 fatcat:zl34fgf26zdqjp7zpfl6sbwg5y

Approximating the top-m passages in a parallel question answering system

Charles L. A. Clarke, Egidio L. Terra
2004 Proceedings of the Thirteenth ACM conference on Information and knowledge management - CIKM '04  
The paper is structured around a specific application -passage retrieval for question answering -but the primary results are more broadly applicable.  ...  However, if we are willing to accept a small probability that one or more of the top-m items may be missed, it is possible to reduce computation time by retrieving only the top k < m from each node.  ...  Parallel information retrieval systems are often based on a cluster-of-workstations architecture.  ... 
doi:10.1145/1031171.1031259 dblp:conf/cikm/ClarkeT04 fatcat:xyd5utwfbzgtzlqvnc54xl5hxy

NLEL-MAAT at ResPubliQA [chapter]

Santiago Correa, Davide Buscaldi, Paolo Rosso
2010 Lecture Notes in Computer Science  
The retrieved passages are ranked depending on the number, length and position of the question n-gram structures found in the passages.  ...  We used the JIRS passage retrieval system, which is based on redundancy, with the assumption that it is possible to find the response to a question in a large enough document collection.  ...  JIRS starts searching the candidate passage with a standard keyword search that retrieves an initial set of passages.  ... 
doi:10.1007/978-3-642-15754-7_24 fatcat:46ljdbotwzaovpvnc7blj6iy5y

PLIERS: A Parallel Information Retrieval System Using MPI [chapter]

A. MacFarlane, J. A. McCann, S. E. Robertson
1999 Lecture Notes in Computer Science  
The use of MPI in implementing algorithms for Parallel Information Retrieval Systems is outlined.  ...  Our description of Document Search includes that for Term Weighting, Boolean, Proximity and Passage Retrieval Operations. Document Update issues are centred on how partitioning methods are supported.  ...  Passage Retrieval Passage Retrieval search is the identification of a part of a document which may be relevant to a user e.g. in a multiple subject document.  ... 
doi:10.1007/3-540-48158-3_39 fatcat:lseaeq4crrfslpxhs36ehx6fy4

Learning Cross-Lingual IR from an English Retriever [article]

Yulong Li, Martin Franz, Md Arafat Sultan, Bhavani Iyer, Young-Suk Lee, Avirup Sil
2022 arXiv   pre-print
We present DR.DECR (Dense Retrieval with Distillation-Enhanced Cross-Lingual Representation), a new cross-lingual information retrieval (CLIR) system trained using multi-stage knowledge distillation (KD  ...  It is also the best single-model retriever on the XOR-TyDi benchmark at the time of this writing.  ...  Acknowledgements We thank Graeme Blackwood and Christoph Tillmann for providing the in-house parallel corpora. We also thank Akari Asai for her help submitting DR.DECR to the XOR-TyDi leaderboard.  ... 
arXiv:2112.08185v2 fatcat:sgkkdkzxn5hedmdbtby3wdd3qu

PLAID: An Efficient Engine for Late Interaction Retrieval [article]

Keshav Santhanam, Omar Khattab, Christopher Potts, Matei Zaharia
2022 arXiv   pre-print
Pre-trained language models are increasingly important components across multiple information retrieval (IR) paradigms.  ...  Without impacting quality, PLAID swiftly eliminates low-scoring passages using a novel centroid interaction mechanism that treats every passage as a lightweight bag of centroids.  ...  The CPU implementation instead parallelizes decompression at the granularity of individual passages.  ... 
arXiv:2205.09707v1 fatcat:upv3k2mchrgt5ohawli7dpq42m

Overview of the INEX 2008 Ad Hoc Track [chapter]

Jaap Kamps, Shlomo Geva, Andrew Trotman, Alan Woodley, Marijn Koolen
2009 Lecture Notes in Computer Science  
For this reason, the retrieval results were liberalized to arbitrary passages and measures were chosen to fairly compare systems retrieving elements, ranges of elements, and arbitrary passages.  ...  We discuss the results for the three tasks, examine the relative effectiveness of element and passage retrieval.  ...  Acknowledgments Eternal thanks to Benjamin Piwowarski for completely updating the X-RAI tools to ensure that all passage offsets can be mapped exactly.  ... 
doi:10.1007/978-3-642-03761-0_1 fatcat:exrtt2h6gzdjxmoqiodqhrhmhy

Question Answering in Spanish [chapter]

José L. Vicedo, Ruben Izquierdo, Fernando Llopis, Rafael Muñoz
2004 Lecture Notes in Computer Science  
University of Alicante for the CLEF 2003 Spanish monolingual QA evaluation task.  ...  This paper describes the architecture, operation and results obtained with the Question Answering prototype for Spanish developed in the Department of Language Processing and Information Systems at the  ...  Passage Retrieval The passage retrieval stage is accomplished in parallel using two different search engines: IR-n [5] and Google 3 .  ... 
doi:10.1007/978-3-540-30222-3_52 fatcat:vxcsr623gvabpp5pp7zezx7sci

Fine-tune the Entire RAG Architecture (including DPR retriever) for Question-Answering [article]

Shamane Siriwardhana, Rivindu Weerasekera, Elliott Wen, Suranga Nanayakkara
2021 arXiv   pre-print
In this paper, we illustrate how to fine-tune the entire Retrieval Augment Generation (RAG) architecture in an end-to-end manner.  ...  We also compare how end-to-end RAG architecture outperforms the original RAG architecture for the task of question answering.  ...  the indexed dataset. • Calculate the document scores by re-computing CLS embeddings for retrieved documents using the initialized passage BERT model and do the same thing for input using the question  ... 
arXiv:2106.11517v1 fatcat:svhpsod3erh6rpvtejcn6q3sra

Overview of the INEX 2009 Ad Hoc Track [chapter]

Shlomo Geva, Jaap Kamps, Miro Lethonen, Ralf Schenkel, James A. Thom, Andrew Trotman
2010 Lecture Notes in Computer Science  
For this reason, the retrieval results were liberalized to arbitrary passages and measures were chosen to fairly compare systems retrieving elements, ranges of elements, and arbitrary passages.  ...  We discuss the results for the three tasks, examine the relative effectiveness of element and passage retrieval.  ...  Acknowledgments Eternal thanks to Benjamin Piwowarski for completely updating the X-RAI tools to ensure that all passage offsets can be mapped exactly.  ... 
doi:10.1007/978-3-642-14556-8_4 fatcat:bdnyqr63bzdzxkxmbkma4mjpqm

Position-Aligned Translation Model for Citation Recommendation [chapter]

Jing He, Jian-Yun Nie, Yang Lu, Wayne Xin Zhao
2012 Lecture Notes in Computer Science  
It can be trained on a collection of query and document pairs, which are assumed to be parallel.  ...  The goal of a citation recommendation system is to suggest some references for a snippet in an article or a book, and this is very useful for both authors and the readers.  ...  For example, one may select the most similar passages using cosine similarity or any retrieval score.  ... 
doi:10.1007/978-3-642-34109-0_27 fatcat:t7sqeabc3jcrxhkkmlhruyrkii

Don't Read Too Much into It: Adaptive Computation for Open-Domain Question Answering [article]

Yuxiang Wu, Sebastian Riedel, Pasquale Minervini, Pontus Stenetorp
2020 arXiv   pre-print
To reduce this cost, we propose the use of adaptive computation to control the computational budget allocated for the passages to be read.  ...  However, they assume all retrieved passages are of equal importance and allocate the same amount of computation to them, leading to a substantial increase in computational cost.  ...  for reading the retrieved passages.  ... 
arXiv:2011.05435v1 fatcat:5zjlwpyoe5avfd4rv7wewhkydy

The CLEF 2003 Cross-Language Spoken Document Retrieval Track [chapter]

Marcello Federico, Gareth J. F. Jones
2004 Lecture Notes in Computer Science  
The current expansion in collections of natural language based digital documents in various media and languages is creating challenging opportunities for automatically accessing the information contained  ...  This paper describes the CLEF 2003 track investigation of Cross-Language Spoken Document Retrieval (CLSDR) combining information retrieval, cross-language translation and speech recognition.  ...  Querydocument retrieval scores are computed with two methods: a statistical language model and an Okapi derived formula.  ... 
doi:10.1007/978-3-540-30222-3_61 fatcat:vrufsw7bhvfwtefjfjazg6oxl4

Elhuyar-IXA: Semantic Relatedness and Cross-Lingual Passage Retrieval [chapter]

Eneko Agirre, Olatz Ansa, Xabier Arregi, Maddalen Lopez de Lacalle, Arantxa Otegi, Xabier Saralegi, Hugo Zaragoza
2010 Lecture Notes in Computer Science  
for the Basque to English retrieval when faced with the lack of parallel corpora for Basque in this domain, and (3) to check the contribution of semantic relatedness based on WordNet to expand the passages  ...  Our focus has been threefold: (1) to check to what extent IR can achieve good results in passage retrieval without question analysis and answer validation, (2) to check Machine Readable Dictionary techniques  ...  retrieval The purpose of the passage retrieval module is to retrieve passages from the document collection which are likely to contain an answer.  ... 
doi:10.1007/978-3-642-15754-7_31 fatcat:k2wh3gvvnndkbix3calvezvu7y
« Previous Showing results 1 — 15 out of 31,857 results