Filters








92,756 Hits in 7.5 sec

RankQA: Neural Question Answering with Answer Re-Ranking [article]

Bernhard Kratzwald, Anna Eigenmann, Stefan Feuerriegel
2019 arXiv   pre-print
In contrast, this work proposes RankQA: RankQA extends the conventional two-stage process in neural QA with a third stage that performs an additional answer re-ranking.  ...  Here the answer re-ranking provides an effective remedy against the underlying noise-information trade-off due to a variable corpus size.  ...  We gratefully acknowledge the support of NVIDIA Corporation with the donation of the Titan Xp GPUs used for this research.  ... 
arXiv:1906.03008v2 fatcat:evb46nkwmzdw7hqjikl2mcvmdm

RankQA: Neural Question Answering with Answer Re-Ranking

Bernhard Kratzwald, Anna Eigenmann, Stefan Feuerriegel
2019 Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics  
The re-ranking leverages different features that are directly extracted from the QA pipeline, i. e., a combination of retrieval and comprehension features.  ...  In contrast, this work proposes RankQA 1 : RankQA extends the conventional two-stage process in neural QA with a third stage that performs an additional answer reranking.  ...  We gratefully acknowledge the support of NVIDIA Corporation with the donation of the Titan Xp GPUs used for this research.  ... 
doi:10.18653/v1/p19-1611 dblp:conf/acl/KratzwaldEF19 fatcat:2zj2fx56wfcubaxbx5iqrwfkuq

Evidence Aggregation for Answer Re-Ranking in Open-Domain Question Answering [article]

Shuohang Wang, Mo Yu, Jing Jiang, Wei Zhang, Xiaoxiao Guo, Shiyu Chang, Zhiguo Wang, Tim Klinger, Gerald Tesauro, Murray Campbell
2018 arXiv   pre-print
We propose two methods, namely, strength-based re-ranking and coverage-based re-ranking, to make use of the aggregated evidence from different passages to better determine the answer.  ...  Our models have achieved state-of-the-art results on three public open-domain QA datasets: Quasar-T, SearchQA and the open-domain version of TriviaQA, with about 8 percentage points of improvement over  ...  We formulate the above evidence aggregation as an answer re-ranking problem.  ... 
arXiv:1711.05116v2 fatcat:nrpnku76ffhapd4t3r66rpx2hm

Ranking Paragraphs for Improving Answer Recall in Open-Domain Question Answering [article]

Jinhyuk Lee, Seongjun Yun, Hyunjae Kim, Miyoung Ko, Jaewoo Kang
2018 arXiv   pre-print
In this paper, we introduce Paragraph Ranker which ranks paragraphs of retrieved documents for a higher answer recall with less noise.  ...  We show that ranking paragraphs and aggregating answers using Paragraph Ranker improves performance of open-domain QA pipeline on the four open-domain QA datasets by 7.8% on average.  ...  Although strength-based answer re-ranking showed good performances on some datasets, it is too complex to efficiently re-rank M answers.  ... 
arXiv:1810.00494v1 fatcat:ep35q2bb6vdvjnta4qnbwv7ge4

Ranking Paragraphs for Improving Answer Recall in Open-Domain Question Answering

Jinhyuk Lee, Seongjun Yun, Hyunjae Kim, Miyoung Ko, Jaewoo Kang
2018 Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing  
In this paper, we introduce Paragraph Ranker which ranks paragraphs of retrieved documents for a higher answer recall with less noise.  ...  We show that ranking paragraphs and aggregating answers using Paragraph Ranker improves performance of open-domain QA pipeline on the four opendomain QA datasets by 7.8% on average.  ...  Although strength-based answer re-ranking showed good performances on some datasets, it is too complex to efficiently re-rank M answers.  ... 
doi:10.18653/v1/d18-1053 dblp:conf/emnlp/LeeYKKK18 fatcat:23mylxd5ofbn3groxj7du5rram

Exploiting Sentence-Level Representations for Passage Ranking [article]

Jurek Leonhardt, Fabian Beringer, Avishek Anand
2021 arXiv   pre-print
We revisit the design decisions that govern the applicability of these models for the passage re-ranking task in open-domain question answering.  ...  In this paper, we explicitly model the sentence-level representations by using Dynamic Memory Networks (DMNs) and conduct empirical evaluation to show improvements in passage re-ranking over fine-tuned  ...  This work deals with the passage re-ranking step. BERT-based models have achieved high performance in passage re-ranking tasks.  ... 
arXiv:2106.07316v2 fatcat:eqok7kfsq5fqhpuhvs7o6uydmu

A Joint Model of Entity Linking and Predicate Recognition for Knowledge Base Question Answering

Yang Li, Qingliang Miao, ChenXin Yin, Chao Huo, Wenxiang Mao, Changjian Hu, Feiyu Xu
2018 China Conference on Knowledge Graph and Semantic Computing  
Finally, the paper selects the answer component from matched triple path based on heuristic rules.  ...  Second, we use a joint training entity linking and predicate recognition model to re-rank candidate triple paths for the question.  ...  In Predicate Recognition module, we pre-rank candidate triple paths with some features. Then BiMPM is utilized to select the matched triple paths.  ... 
dblp:conf/ccks/LiMYHMHX18 fatcat:qia6vzjiqzg2rh6md7zfhl5i6a

Learning to rank for robust question answering

Arvind Agarwal, Hema Raghavan, Karthik Subbian, Prem Melville, Richard D. Lawrence, David C. Gondek, James Fan
2012 Proceedings of the 21st ACM international conference on Information and knowledge management - CIKM '12  
We finally evaluate several rank aggregation techniques to combine these algorithms, and find that Supervised Kemeny aggregation is a robust technique that always beats the baseline ranking approach used  ...  This paper aims to solve the problem of improving the ranking of answer candidates for factoid based questions in a state-of-the-art Question Answering system.  ...  of rank-aggregation methods along with the learning to rank methods on Medical and Jeopardy data.  ... 
doi:10.1145/2396761.2396867 dblp:conf/cikm/AgarwalRSMLGF12 fatcat:k5nxdatwvrcavdb7w5huvl2zq4

Denoising Distantly Supervised Open-Domain Question Answering

Yankai Lin, Haozhe Ji, Zhiyuan Liu, Maosong Sun
2018 Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)  
Distantly supervised open-domain question answering (DS-QA) aims to find answers in collections of unlabeled text.  ...  Moreover, distant supervision data inevitably accompanies with the wrong labeling problem, and these noisy data will substantially degrade the performance of DS-QA.  ...  This paper is also partially funded by Microsoft Research Asia FY17-RES-THEME-017.  ... 
doi:10.18653/v1/p18-1161 dblp:conf/acl/SunLLJ18 fatcat:ht45gp4y7zdsvkwymw2f6edaeu

Recent Trends in Deep Learning Based Open-Domain Textual Question Answering Systems

Zhen Huang, Shiyi Xu, Minghao Hu, Xinyi Wang, Jinyan Qiu, Yongquan Fu, Yuncai Zhao, Yuxing Peng, Changjian Wang
2020 IEEE Access  
Open-domain textual question answering (QA), which aims to answer questions from large data sources like Wikipedia or the web, has gained wide attention in recent years.  ...  INDEX TERMS Open-domain textual question answering, deep learning, machine reading comprehension, information retrieval.  ...  FINAL ANSWER SELECTION Final answer selection mainly selects the final answer from multiple candidate answers using feature aggregation, aggregation methods can be divided into the following types. • Evidence  ... 
doi:10.1109/access.2020.2988903 fatcat:po4euxfronf3pob52qc2wcgrre

Information retrieval for label noise document ranking by bag sampling and group-wise loss [article]

Chunyu Li and Jiajia Ding and Xing hu and Fan Wang
2022 arXiv   pre-print
The retrieval data is divided into multiple bags at the ranking stage, and negative samples are selected in each bag. After sampling, two losses are combined. The first loss is LCE.  ...  Notably, our model shows excellent performance on the MS MARCO Long document ranking leaderboard.  ...  Since there is only one correctly labeled answer, there will be documents and topics similar to the correct answer near the right answer.  ... 
arXiv:2203.06408v1 fatcat:jc3k5s2dtneqjk2pjy6pt7e7fe

New York University at TREC 2018 Complex Answer Retrieval Track

Rodrigo Nogueira, Kyunghyun Cho
2018 Text Retrieval Conference  
In this framework, an agent consists of multiple specialized subagents and a meta-agent that learns to aggregate the answers from sub-agents to produce a final answer.  ...  We use a method introduced by Nogueira et al. ( 2018 ) to efficiently learn diverse strategies in reinforcement learning for query reformulation and focus minimally on the ranking function.  ...  The K lists of ranked documents returned by the environment are then merged into a single list and re-ranked by the Aggregator.  ... 
dblp:conf/trec/NogueiraC18 fatcat:pvmlkjpm4jhq7cpyirqtqgo2yq

The Stability of Gene Selection in Microarray Experiments

Magdalena Wietlicka-Piszcz
2013 Studies in Logic, Grammar and Rhetoric  
The similarities be- tween gene rankings yielded by various gene selection methods performed with resampled datasets were assessed.  ...  The mean percentage of overlapping genes for two rankings varied from 10 to 90% depending on the applied gene selection method and the size of the list.  ...  To assess the stability of gene rankings produced by a particular gene selection method, the similarities between the rankings obtained with the original dataset and the rankings obtained with the re-sampled  ... 
doi:10.2478/slgr-2013-0040 fatcat:g5i4jlqvufcy5lvb6x5wq2ozbu

Adversarial Training for Community Question Answer Selection Based on Multi-scale Matching [article]

Xiao Yang, Madian Khabsa, Miaosen Wang, Wei Wang, Madian Khabsa, Ahmed Awadallah, Daniel Kifer, C. Lee Giles
2018 arXiv   pre-print
We evaluate the proposed method on SemEval 2016 and SemEval 2017 datasets and achieves state-of-the-art or similar performance.  ...  Community-based question answering (CQA) websites represent an important source of information.  ...  As a result, each question is associated with 100 candidate answers, and the ultimate goal is to re-rank these 100 answers according to their relevance to the original question.  ... 
arXiv:1804.08058v2 fatcat:5vkvemfydjdlhf4krt2wg4kq5m

Adversarial Training for Community Question Answer Selection Based on Multi-Scale Matching

Xiao Yang, Madian Khabsa, Miaosen Wang, Wei Wang, Ahmed Hassan Awadallah, Daniel Kifer, C. Lee Giles
2019 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
We evaluate the proposed method on SemEval 2016 and SemEval 2017 datasets and achieves state-of-the-art or similar performance.  ...  Community-based question answering (CQA) websites represent an important source of information.  ...  As a result, each question is associated with 100 candidate answers, and the ultimate goal is to re-rank these 100 answers according to their relevance to the original question.  ... 
doi:10.1609/aaai.v33i01.3301395 fatcat:zj6u4dgsw5ectfbofa4qi2h37a
« Previous Showing results 1 — 15 out of 92,756 results