349 Hits in 3.5 sec

Reranking answers for definitional QA using language modeling

Yi Chen, Ming Zhou, Shilong Wang
2006 Proceedings of the 21st International Conference on Computational Linguistics and the 44th annual meeting of the ACL - ACL '06  
To relax this assumption, this paper proposes a novel language model-based answer reranking method to improve the existing bag-ofwords model approach by considering the dependence of the words in the centroid  ...  The results on the TREC 2003 test set show that the reranking approach with biterm language model, significantly outperforms the one with the bag-ofwords model and unigram language model by 14.9% and 12.5%  ...  Cheng Niu, Yunbo Cao for their valuable suggestions on the draft of this paper. We are indebted to Shiqi Zhao, Shenghua Bao, Wei Yuan for their valuable discussions about this paper.  ... 
doi:10.3115/1220175.1220311 dblp:conf/acl/ChenZW06 fatcat:7if62jfzkbbkjfjprnz5at2vxm

Bio-AnswerFinder: a system to find answers to questions from biomedical texts

2019 Database: The Journal of Biological Databases and Curation  
The answer sentences are further ranked by a fine-tuned bidirectional encoder representation from transformers (BERT) classifier trained using 100 answer candidate sentences per question for 492 BioASQ  ...  Bio-AnswerFinder uses a weighted-relaxed word mover's distance based similarity on word/phrase embeddings learned from PubMed abstracts to rank answers after question focus entity type filtering.  ...  We also thank the anonymous reviewers for their comments.  ... 
doi:10.1093/database/baz137 pmid:31925435 pmcid:PMC7053013 fatcat:scxluxhtd5burcx2xkebte6die

Answering Any-hop Open-domain Questions with Iterative Document Reranking [article]

Ping Nie, Yuyu Zhang, Arun Ramamurthy, Le Song
2021 arXiv   pre-print
Existing approaches for open-domain question answering (QA) are typically designed for questions that require either single-hop or multi-hop reasoning, which make strong assumptions of the complexity of  ...  To improve the retrieval accuracy, we propose a graph-based reranking model that perform multi-document interaction as the core of our iterative reranking framework.  ...  [CLS] and [SEP] are special tokens used in pre-trained language models such as BERT [7] and AL-BERT [16] .  ... 
arXiv:2009.07465v5 fatcat:m3hosnzbwvdx3k3g7x7tlm5zhi

A Case Based Reasoning Approach for Answer Reranking in Question Answering [article]

Karl-Heinz Weis
2015 arXiv   pre-print
In this document I present an approach to answer validation and reranking for question answering (QA) systems.  ...  In the experiments based on QA@CLEF questions, the best learned models make heavy use of CBR features.  ...  Rank-optimizing decision trees [Gl09] are used for learning an answer reranking model from the existing answer validation features of LogAnswer [FG10] and the new CBR features.  ... 
arXiv:1503.02917v1 fatcat:tenrbjmd2zeu3ghxcrxtpreefq

Passage Reranking for Question Answering Using Syntactic Structures and Answer Types [chapter]

Elif Aktolga, James Allan, David A. Smith
2011 Lecture Notes in Computer Science  
Experimental results using the TREC QA 1999-2003 datasets show that our method significantly outperforms the baselines over all ranks in terms of the MRR measure.  ...  Whereas in previous work, passages are reranked only on the basis of syntactic structures of questions and answers, our method achieves a better ranking by aligning the syntactic structures based on the  ...  This work was supported in part by the Center for Intelligent Information Retrieval.  ... 
doi:10.1007/978-3-642-20161-5_62 fatcat:az6iraob2nggfhgtvcflpqchya

Building structures from classifiers for passage reranking

Aliaksei Severyn, Massimo Nicosia, Alessandro Moschitti
2013 Proceedings of the 22nd ACM international conference on Conference on information & knowledge management - CIKM '13  
Further comparison on the task restricted to the answer sentence reranking shows an improvement in MAP of more than 8% over the state of the art. 1 We use this question/answer passage (q/a) pair from TREC  ...  We conduct an extensive experimental evaluation of our models on well-known benchmarks from the question answer (QA) track of TREC challenges.  ...  Section 5 contains answer passage reranking experiments on TREC QA data, while Section 6 compares our models for answer sentence reranking to the current state-of-the-art systems.  ... 
doi:10.1145/2505515.2505688 dblp:conf/cikm/SeverynNM13 fatcat:pvkfvlgazrb5popdvxsbhuw644

Weakly Supervised Pre-Training for Multi-Hop Retriever [article]

Yeon Seonwoo, Sang-Woo Lee, Ji-Hoon Kim, Jung-Woo Ha, Alice Oh
2021 arXiv   pre-print
In multi-hop QA, answering complex questions entails iterative document retrieval for finding the missing entity of the question.  ...  as weak supervision for pre-training, and 3) a pre-training model structure based on dense encoders.  ...  In swer prediction and supporting fact prediction per- reranking experiments, we use the same rerank- formance of multi-hop QA models (Yang et al., ing model as MDR-rerank.  ... 
arXiv:2106.09983v1 fatcat:mhnlxpyp2rbsdisjbkfax7jpkq

Joint Models for Answer Verification in Question Answering Systems [article]

Zeyu Zhang, Thuy Vu, Alessandro Moschitti
2021 arXiv   pre-print
This paper studies joint models for selecting correct answer sentences among the top k provided by answer sentence selection (AS2) modules, which are core components of retrieval-based Question Answering  ...  We tested our models on WikiQA, TREC-QA, and a real-world dataset. The results show that our models obtain the new state of the art in AS2.  ...  For example, ASR achieves the best reported results, i.e., MAP values of 92.80% and 94.88, on WikiQA and TREC-QA, respectively.  ... 
arXiv:2107.04217v1 fatcat:z3vvruhxfzcqnbqgpu7u6f4p2e

An Integrated Machine Learning and Case-Based Reasoning Approach to Answer Validation

Ingo Glockner, Karl-Heinz Weis
2012 2012 11th International Conference on Machine Learning and Applications  
In our experiments on QA@CLEF questions, the best learned models make heavy use of CBR features.  ...  We propose a case-based reasoning (CBR) approach to answer validation/answer scoring and reranking in question answering (QA) systems, where annotated answer candidates for known questions provide evidence  ...  V uses data from QA@CLEF tasks for evaluating case retrieval and the achieved reranking quality. II.  ... 
doi:10.1109/icmla.2012.90 dblp:conf/icmla/GlocknerW12 fatcat:n5dbtsnv6fannorgtlc4hpio4m

Knowing More About Questions Can Help: Improving Calibration in Question Answering [article]

Shujian Zhang, Chengyue Gong, Eunsol Choi
2021 arXiv   pre-print
We study calibration in question answering, estimating whether model correctly predicts answer for each question.  ...  Furthermore, we present the first calibration study in the open retrieval setting, comparing the calibration accuracy of retrieval-based span prediction models and answer generation models.  ...  Acknowledgments We would like to thank UT Austin NLP group, especially Kaj Bostrom and Greg Durrett for feedback and suggestions.  ... 
arXiv:2106.01494v1 fatcat:ng2o6xdfevatbozasglpbke4xu

Multimedia answering

Liqiang Nie, Meng Wang, Zhengjun Zha, Guangda Li, Tat-Seng Chua
2011 Proceedings of the 34th international ACM SIGIR conference on Research and development in Information - SIGIR '11  
Our scheme investigates a rich set of techniques including question/answer classification, query generation, image and video search reranking, etc.  ...  However, for many questions, pure texts cannot provide intuitive information, while image or video contents are more appropriate.  ...  QA [26] , Definitional QA [11] and List QA [32] .  ... 
doi:10.1145/2009916.2010010 dblp:conf/sigir/NieWZLC11 fatcat:7o7vlyx3pvaynbx5dwidcik6uy

Dublin City University at QA@CLEF 2008 [chapter]

Sisay Fissaha Adafre, Josef van Genabith
2009 Lecture Notes in Computer Science  
We describe our participation in Multilingual Question Answering at CLEF 2008 using German and English as our source and target languages respectively.  ...  The system was built using UIMA (Unstructured Information Management Architechture) as underlying framework.  ...  Syntax based evidences have been used to rerank candidate answers in a number of QA Systems [6, 8, 12] .  ... 
doi:10.1007/978-3-642-04447-2_41 fatcat:del5ub26prea7k5zc6xjgw2rki

Pruning the Index Contents for Memory Efficient Open-Domain QA [article]

Martin Fajcik, Martin Docekal, Karel Ondrej, Pavel Smrz
2021 arXiv   pre-print
This work presents a simple approach for pruning the contents of a massive index such that the open-domain QA system altogether with index, OS, and library components fits into 6GiB docker image while  ...  Specifically, it proposes the novel R2-D2 (Rank twice, reaD twice) pipeline composed of retriever, passage reranker, extractive reader, generative reader and a simple way to combine them.  ...  The computation used the infrastructure supported by the Czech Ministry of Education, Youth and Sports from the Large Infrastructures for Research, Experimental Development and Innovations project "IT4Innovations  ... 
arXiv:2102.10697v2 fatcat:rbkfghoainb37asci2etnhyzse

Multilingual Answer Sentence Reranking via Automatically Translated Data [article]

Thuy Vu, Alessandro Moschitti
2021 arXiv   pre-print
Transformer model it is enough to rank answers in multiple languages; and (iii) mixed-language question/answer pairs can be used to fine-tune models to select answers from any language, where the input  ...  The main findings of this paper are: (i) the training data for AS2 translated into a target language can be used to effectively fine-tune a Transformer-based model for that language; (ii) one multilingual  ...  First and foremost, we described our approach for creating a large-scale dataset for QA of a total 120K question-answer pairs for English.  ... 
arXiv:2102.10250v1 fatcat:wyn2ayomkrf5bjlqkbk3ybloaq

KG-FiD: Infusing Knowledge Graph in Fusion-in-Decoder for Open-Domain Question Answering [article]

Donghan Yu, Chenguang Zhu, Yuwei Fang, Wenhao Yu, Shuohang Wang, Yichong Xu, Xiang Ren, Yiming Yang, Michael Zeng
2021 arXiv   pre-print
We initiate the passage node embedding from the FiD encoder and then use graph neural network (GNN) to update the representation for reranking.  ...  To improve the efficiency, we build the GNN on top of the intermediate layer output of the FiD encoder and only pass a few top reranked passages into the higher layers of encoder and decoder for answer  ...  method in two modules. w/ KG refers to using GNN for passage reranking as our current model while w/o KG refers to using MLP instead of GNN.  ... 
arXiv:2110.04330v1 fatcat:el6i64n2ojhn3evsgrxifkfmoy
« Previous Showing results 1 — 15 out of 349 results