Filters








77,576 Hits in 6.1 sec

What is wrong with style transfer for texts? [article]

Alexey Tikhonov, Ivan P. Yamshchikov
2018 arXiv   pre-print
A number of recent machine learning papers work with an automated style transfer for texts and, counter to intuition, demonstrate that there is no consensus formulation of this NLP task.  ...  Introduction Arguably, a growing interest to style transfer algorithms for texts is motivated by Gatys et al. (2016) , where style transfer was developed for images in a very convincing manner.  ...  In , for example, the authors estimate the quality of the style transfer with binary sentiment classifier trained on the same corpus of Yelp and Amazon reviews that is used for the training of the 'style  ... 
arXiv:1808.04365v1 fatcat:vaqciwg6gbftjiwb2za52kj6d4

Contextual Text Style Transfer [article]

Yu Cheng, Zhe Gan, Yizhe Zhang, Oussama Elachqar, Dianqi Li, Jingjing Liu
2020 arXiv   pre-print
We introduce a new task, Contextual Text Style Transfer - translating a sentence into a desired style with its surrounding context taken into account.  ...  To realize high-quality style transfer with natural context preservation, we propose a Context-Aware Style Transfer (CAST) model, which uses two separate encoders for each input sentence and its surrounding  ...  ., 2019) is another direction of text style transfer, aiming to change the style of a given sentence to more formal text. dos Santos et al.  ... 
arXiv:2005.00136v1 fatcat:azte6pnazbfotkbtxekttevlkq

Evaluating Style Transfer for Text

Remi Mir, Bjarke Felbo, Nick Obradovich, Iyad Rahwan
2019 Proceedings of the 2019 Conference of the North  
Research in the area of style transfer for text is currently bottlenecked by a lack of standard evaluation practices.  ...  Distance on style-masked texts, and adversarial classification for the respective aspects.  ...  Acknowledgments We would like to thank Juncen Li, Tianxiao Shen, and Junbo (Jake) Zhao for guidance in the use of their respective style transfer models.  ... 
doi:10.18653/v1/n19-1049 dblp:conf/naacl/MirFOR19 fatcat:l676z6smq5h2xaqiyqinuhc5ra

Evaluating Style Transfer for Text [article]

Remi Mir, Bjarke Felbo, Nick Obradovich, Iyad Rahwan
2019 arXiv   pre-print
Research in the area of style transfer for text is currently bottlenecked by a lack of standard evaluation practices.  ...  Distance on style-masked texts, and adversarial classification for the respective aspects.  ...  Acknowledgments We would like to thank Juncen Li, Tianxiao Shen, and Junbo (Jake) Zhao for guidance in the use of their respective style transfer models.  ... 
arXiv:1904.02295v1 fatcat:dkkqntp4gjfjxgyu7zk7oifcsy

Towards Universality in Multilingual Text Rewriting [article]

Xavier Garcia, Noah Constant, Mandy Guo, Orhan Firat
2021 arXiv   pre-print
In this work, we take the first steps towards building a universal rewriter: a model capable of rewriting text in any language to exhibit a wide variety of attributes, including styles and languages, while  ...  We then show that our model is able to modify multiple attributes at once, for example adjusting both language and sentiment jointly.  ...  One such task is text style transfer, which consists of transforming text from one style to another while preserving as much of the semantics as possible.  ... 
arXiv:2107.14749v1 fatcat:litw5uzxhzebvie4kifse53jnm

TextSETTR: Few-Shot Text Style Extraction and Tunable Targeted Restyling [article]

Parker Riley, Noah Constant, Mandy Guo, Girish Kumar, David Uthus, Zarana Parekh
2021 arXiv   pre-print
We present a novel approach to the problem of text style transfer.  ...  We adapt T5 (Raffel et al., 2020), a strong pretrained text-to-text model, to extract a style vector from text and use it to condition the decoder to perform style transfer.  ...  What's the weather like for the next day? What is the temperature and what is the humidity. Table 7 : Random augmentations of input text "What'll the weather be tomorrow?"  ... 
arXiv:2010.03802v3 fatcat:jcvk5s6ohnc2jh3keb4icgf3gm

Artistic text and its translation

O. Skliarenko
2021 Science and Education a New Dimension  
The given article reveals the problem of artistic texts translation. Peculiarities of the process have been highlighted and analyzed.  ...  It was stated that each translation is a creative process and must be marked by the translator's individuality, but the main task of the translator is still to transmit the characteristic features of the  ...  He is answered: Not you, sir. She is. What do you do in that case? Here's how the interpreter got out of that situation: Is it over? -Not for you, sir. For her.  ... 
doi:10.31174/send-ph2021-251ix74-16 fatcat:eambgoa2lnfrtlonfukhcw6qja

The Great Text Commentary

1891 Expository Times  
If we have been told that, we have been told wrong.  ...  But hope is thought of, at first sight, as a self-regarding quality, which throws forward its desires into the future, and dwells in what it imagines of happiness for itself.  ... 
doi:10.1177/001452469100200705 fatcat:txfszdqlazebtdepmaivtwmh4a

Deep Learning for Text Style Transfer: A Survey [article]

Di Jin, Zhijing Jin, Zhiting Hu, Olga Vechtomova, Rada Mihalcea
2021 arXiv   pre-print
Text style transfer is an important task in natural language generation, which aims to control certain attributes in the generated text, such as politeness, emotion, humor, and many others.  ...  In this paper, we present a systematic survey of the research on neural text style transfer, spanning over 100 representative articles since the first neural text style transfer work in 2017.  ...  Association for Computational Linguistics. 2018. What is wrong with style transfer for Wang, Ke, Hang Hua, and Xiaojun Wan. 2019. texts? CoRR, abs/1808.04365.  ... 
arXiv:2011.00416v5 fatcat:wfw3jfh2mjfupbzrmnztsqy4ny

Semi-supervised Text Style Transfer: Cross Projection in Latent Space

Mingyue Shang, Piji Li, Zhenxin Fu, Lidong Bing, Dongyan Zhao, Shuming Shi, Rui Yan
2019 Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP)  
Text style transfer task requires the model to transfer a sentence of one style to another style while retaining its original content meaning, which is a challenging problem that has long suffered from  ...  In this paper, we first propose a semi-supervised text style transfer model that combines the small-scale parallel data with the large-scale nonparallel data.  ...  Acknowledgement We would like to thank the anonymous reviewers for their constructive comments.  ... 
doi:10.18653/v1/d19-1499 dblp:conf/emnlp/ShangLFBZSY19 fatcat:juf5lcufojelhaj2hcdpde5vye

Semi-supervised Text Style Transfer: Cross Projection in Latent Space [article]

Mingyue Shang, Piji Li, Zhenxin Fu, Lidong Bing, Dongyan Zhao, Shuming Shi, Rui Yan
2019 arXiv   pre-print
Text style transfer task requires the model to transfer a sentence of one style to another style while retaining its original content meaning, which is a challenging problem that has long suffered from  ...  In this paper, we first propose a semi-supervised text style transfer model that combines the small-scale parallel data with the large-scale nonparallel data.  ...  Acknowledgement We would like to thank the anonymous reviewers for their constructive comments.  ... 
arXiv:1909.11493v1 fatcat:76udtzbdp5fafnuoac3pwygt5q

Style Transfer for Texts: Retrain, Report Errors, Compare with Rewrites [article]

Alexey Tikhonov, Viacheslav Shibaev, Aleksander Nagaev, Aigul Nugmanova, Ivan P. Yamshchikov
2019 arXiv   pre-print
This paper shows that standard assessment methodology for style transfer has several significant problems.  ...  the intuitive goal of the style transfer task.  ...  Content and style are deeply fused in natural language, but style transfer for texts is often addressed in the context of disentangled latent representations (Hu et al., 2017a; Shen et al., 2017; Fu et  ... 
arXiv:1908.06809v2 fatcat:hzwywebq35a4jatdn23fn63r5u

Text Detoxification using Large Pre-trained Neural Models [article]

David Dale, Anton Voronov, Daryna Dementieva, Varvara Logacheva, Olga Kozlova, Nikita Semenov, Alexander Panchenko
2021 arXiv   pre-print
Finally, we present the first large-scale comparative study of style transfer models on the task of toxicity removal. We compare our models with a number of methods for style transfer.  ...  We present two novel unsupervised methods for eliminating toxicity in text.  ...  We are grateful to the reviewers for their helpful suggestions which substantially improved this work.  ... 
arXiv:2109.08914v2 fatcat:umhz66ev3jep3e3uo73al5vqae

My Text in Your Handwriting

Tom S. F. Haines, Oisin Mac Aodha, Gabriel J. Brostow
2016 ACM Transactions on Graphics  
Rendering new text in someone's handwriting is difficult because natural handwriting is highly variable, yet follows both intentional and involuntary structure that makes a person's style self-consistent  ...  An annotated sample of the author's handwriting is required; the system is flexible enough that historical documents can usually be used with only a little extra effort.  ...  We would like to thank Aindri Chakraborty, Bryan O'Regan, Melissa Terras and the reviewers for their kind assistance.  ... 
doi:10.1145/2886099 fatcat:glm3mcyqgzg5dffjh2cxwkstdu

Text [chapter]

Ingo Gildenhard, Andrew Zissos
2016 Ovid, Metamorphoses, 3.511-733: Latin Text with Introduction, Commentary, Glossary of Terms, Vocabulary Aid and Study Questions  
Unfortunately, we cannot be sure which text (or texts) he is referencing here, not least since we do not have Pacuvius' play from which he took the name Acoetes (see Intro. §5b-iv). There is  ...  In a sense, then, Ovid is here outsourcing responsibility for the truth-value of his account, saying that he is covering material he has found elsewhere, without necessarily vouching for its veracity.  ...  Very much like Actaeon, who stumbled across Diana and her nymphs in the nude and paid for it with a canine sparagmos, the 'boar' is in the wrong place at the wrong time.  ... 
doi:10.11647/obp.0073.02 fatcat:ungvtafv6bdhjkn656xv24hlfy
« Previous Showing results 1 — 15 out of 77,576 results