Filters








62 Hits in 2.2 sec

Abstractive Unsupervised Multi-Document Summarization using Paraphrastic Sentence Fusion

Mir Tafseer Nayeem, Tanvir Ahmed Fuad, Yllias Chali
2018 International Conference on Computational Linguistics  
We design a paraphrastic sentence fusion model which jointly performs sentence fusion and paraphrasing using skip-gram word embedding model at the sentence level.  ...  We conduct our experiments on the human-generated multi-sentence compression datasets and evaluate our system on several newly proposed Machine Translation (MT) evaluation metrics.  ...  Acknowledgements We would like to thank the anonymous reviewers for their useful comments.  ... 
dblp:conf/coling/NayeemFC18 fatcat:qndwhochmrffnoavsmhlamljoi

Learning from Lexical Perturbations for Consistent Visual Question Answering [article]

Spencer Whitehead, Hui Wu, Yi Ren Fung, Heng Ji, Rogerio Feris, Kate Saenko
2020 arXiv   pre-print
regularization tool for VQA models.  ...  We also present VQA Perturbed Pairings (VQA P2), a new, low-cost benchmark and augmentation pipeline to create controllable linguistic variations of VQA questions.  ...  For Q3R, λ = 1.0 and r = 0.2 for all experiments and we use a L1 loss for our distance metric. Please see Appendix A for details.  ... 
arXiv:2011.13406v2 fatcat:poyfejkn4nbx3h3t6pkzdlbzly

ParaBank: Monolingual Bitext Generation and Sentential Paraphrasing via Lexically-constrained Neural Machine Translation [article]

J. Edward Hu, Rachel Rudinger, Matt Post, Benjamin Van Durme
2019 arXiv   pre-print
Finally, we use ParaBank to train a monolingual NMT model with the same support for lexically-constrained decoding for sentence rewriting tasks.  ...  We present ParaBank, a large-scale English paraphrase dataset that surpasses prior work in both quantity and quality.  ...  Acknowledgments We thank Chu-Cheng Lin and three anonymous reviewers for their thoughtful feedback. This research was supported in part by DARPA AIDA.  ... 
arXiv:1901.03644v1 fatcat:xycxwrlgnfep3joaukmcweb3w4

Debugging Frame Semantic Role Labeling [article]

Alexandre Kabbach
2019 arXiv   pre-print
We propose a quantitative and qualitative analysis of the performances of statistical models for frame semantic structure extraction.  ...  We report on the robustness of a recent statistical classifier for frame semantic parsing to lexical configurations of predicate-argument structures, relying on an artificially augmented dataset generated  ...  They then used a conditional log-linear model over spans for each role of each evoked where the kappa statistic is used, FrameNet annotators do not have to choose among a fixed pool of label for each annotated  ... 
arXiv:1901.07475v1 fatcat:7sktiu6yjnawnomsqvh5sgde7u

PARABANK: Monolingual Bitext Generation and Sentential Paraphrasing via Lexically-Constrained Neural Machine Translation

J. Edward Hu, Rachel Rudinger, Matt Post, Benjamin Van Durme
2019 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
Finally, we use PARABANK to train a monolingual NMT model with the same support for lexically-constrained decoding for sentence rewriting tasks.  ...  We present PARABANK, a large-scale English paraphrase dataset that surpasses prior work in both quantity and quality.  ...  Acknowledgments We thank Chu-Cheng Lin and three anonymous reviewers for their thoughtful feedback. This research was supported in part by DARPA AIDA.  ... 
doi:10.1609/aaai.v33i01.33016521 fatcat:ge234vb4wvgvdlseirzkqdhiga

Iterative Paraphrastic Augmentation with Discriminative Span Alignment [article]

Ryan Culkin, J. Edward Hu, Elias Stengel-Eskin, Guanghui Qin, Benjamin Van Durme
2020 arXiv   pre-print
We introduce a novel paraphrastic augmentation strategy based on sentence-level lexically constrained paraphrasing and discriminative span alignment.  ...  Our approach allows for the large-scale expansion of existing resources, or the rapid creation of new resources from a small, manually-produced seed corpus.  ...  Using Paraphrastic Data on a Downstream Task We have demonstrated the usefulness of iterative paraphrastic augmentation for expanding lexical Table 3 : Human evaluation of system outputs across several  ... 
arXiv:2007.00320v1 fatcat:cjx6kkzpi5gujhjpqu65zxfxc4

Generating Phrasal and Sentential Paraphrases: A Survey of Data-Driven Methods

Nitin Madnani, Bonnie J. Dorr
2010 Computational Linguistics  
In this article, we attempt to conduct a comprehensive and application-independent survey of data-driven phrasal and sentential paraphrase generation methods, while also conveying an appreciation for the  ...  We also discuss the strategies used for evaluating paraphrase generation techniques and briefly explore some future trends in paraphrase generation.  ...  Looking towards the future, Dagan (2008) suggests that the textual entailment task provides a comprehensive framework for semantic inference and argues for building a concrete inference engine that not  ... 
doi:10.1162/coli_a_00002 fatcat:deov6ypw4bb7hfkkkvjxsyt6fy

NEURAL SYMBOLIC ARABIC PARAPHRASING WITH AUTOMATIC EVALUATION

Fatima Al-Raisi, Abdelwahab Bourai, Weijian Lin
2018 Figshare  
Additionally, we propose a novel automatic evaluation metric for paraphrasing that correlates highly with human judgement.  ...  We share a large coverage phrase dictionary for Arabic and contribute a large parallel monolingual corpus that can be used in developing new seq-to-seq models for paraphrasing.  ...  We also plan to use it at word subunits such as morphemes and even character level especially for surface distance comparison in morphology rich languages like Arabic.  ... 
doi:10.6084/m9.figshare.6700718 fatcat:jcztjf7bdbc4xdpkc7kz7v4dsi

ParaPhraser: Russian Paraphrase Corpus and Shared Task [chapter]

Lidia Pivovarova, Ekaterina Pronoza, Elena Yagunova, Anton Pronoza
2017 Communications in Computer and Information Science  
Research in the area of paraphrase extraction, detection and generation has been successfully developing for a long time while there has been only a recent surge of interest towards the problem in the  ...  We try to overcome this gap by introducing the project ParaPhraser.ru dedicated to the collection of Russian paraphrase corpus and organizing a Paraphrase Detection Shared Task, which uses the corpus as  ...  Since a comprehensive Russian WordNet is not currently available we used a synonym dictionary instead of WordNet; we also introduced several modifications into Fernando and Stevenson similarity metric  ... 
doi:10.1007/978-3-319-71746-3_18 fatcat:rnplovd5dzcqhovmfglechaftu

On the Translation and Use of the Psalms for the Public Worship of the Church

W. Robertson Smith
1904 Expository Times  
But what business has he, as a student of science, to be content to be mistaken with any man? If his science makes immortality impossible, let him say so and reject the belief in immortality.  ...  For by substituting an entirely different set of circumstances, we, as a rule, do away with all the fitness and harmony of the Psalmist's utterance. (2) The foundation of the paraphrast must be historic  ...  And even the common and necessary mental substitutions, by which we understand the Heavenly Zion to stand for Jerusalem, and the Church for Israel, may be abused, and form the first step towards an allegorical  ... 
doi:10.1177/001452460401600302 fatcat:gujqhb7tsng6tlxnwz7rxkbw2u

Using Semantic Similarity as Reward for Reinforcement Learning in Sentence Generation

Go Yasui, Yoshimasa Tsuruoka, Masaaki Nagata
2019 Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics: Student Research Workshop  
While cross-entropy loss has convenient properties for supervised learning, it is unable to evaluate sentences as a whole, and lacks flexibility.  ...  Traditional model training for sentence generation employs cross-entropy loss as the loss function.  ...  Acknowledgments We would like to thank Kazuma Hashimoto and anonymous reviewers for helpful comments and suggestions.  ... 
doi:10.18653/v1/p19-2056 dblp:conf/acl/YasuiTN19 fatcat:i3wr35vlv5gmhpgqqep52av4iq

Beyond BLEU:Training Neural Machine Translation with Semantic Similarity

John Wieting, Taylor Berg-Kirkpatrick, Kevin Gimpel, Graham Neubig
2019 Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics  
In this paper, we introduce an alternative reward function for optimizing NMT systems that is based on recent work in semantic similarity.  ...  However, training with BLEU has some limitations: it doesn't assign partial credit, it has a limited range of output values, and it can penalize semantically correct hypotheses if they differ lexically  ...  Tuning for metrics other than BLEU has not (to our knowledge) been extensively examined for NMT, and we hope this paper provides a first step towards broader consideration of training metrics for NMT.  ... 
doi:10.18653/v1/p19-1427 dblp:conf/acl/WietingBGN19 fatcat:ckylq5pjtbfhhpswp2lkgpplwe

Beyond BLEU: Training Neural Machine Translation with Semantic Similarity [article]

John Wieting, Taylor Berg-Kirkpatrick, Kevin Gimpel, Graham Neubig
2019 arXiv   pre-print
In this paper, we introduce an alternative reward function for optimizing NMT systems that is based on recent work in semantic similarity.  ...  However, training with BLEU has some limitations: it doesn't assign partial credit, it has a limited range of output values, and it can penalize semantically correct hypotheses if they differ lexically  ...  Tuning for metrics other than BLEU has not (to our knowledge) been extensively examined for NMT, and we hope this paper provides a first step towards broader consideration of training metrics for NMT.  ... 
arXiv:1909.06694v1 fatcat:3xapqc54u5ggfomuklzwhrk6wq

C Authorship and the Development of the English Metrical Psalter [chapter]

2016 The Reformation in Rhyme  
Wynken de Worde's Sarum Hours, for example, offer a metrical calendar toward the beginning, a long verse litany in the middle, and a crudely rhymed version of the Ten Commandments at the conclusion.  ...  The immediate and continued popularity of these versifications encouraged Sternhold to broaden his audience in promotion of the evangelical agenda in the psalms he composed toward a complete metrical psalter  ... 
doi:10.4324/9781315237435-17 fatcat:bpkh4vg3hjhphefrylomj47xgy

Not Far Away, Not So Close: Sample Efficient Nearest Neighbour Data Augmentation via MiniMax [article]

Ehsan Kamalloo, Mehdi Rezagholizadeh, Peyman Passban, Ali Ghodsi
2021 arXiv   pre-print
Inspired by this paradigm, we introduce Minimax-kNN, a sample efficient data augmentation strategy tailored for Knowledge Distillation (KD).  ...  Recently, leveraging kNN such that augmented examples are retrieved from large repositories of unlabelled sentences has made a step toward interpretable augmentation.  ...  Acknowledgments We thank MindSpore 5 -a new deep learning framework-for partially supporting this work.  ... 
arXiv:2105.13608v2 fatcat:gfw524e6xvcxtmiwnr5am3jnva
« Previous Showing results 1 — 15 out of 62 results