Style versus Content: A distinction without a (learnable) difference?

Jafaritazehjani, Somayeh; Lecorvé, Gwénolé; Lolive, Damien; Kelleher, John D.

doi:10.18653/v1/2020.coling-main.197

Cited by 10 publications

(8 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There has been a lot of attention to the problems of evaluation metrics of TST and potential improvements (Pang and Gimpel 2019;Tikhonov and Yamshchikov 2018;Mir et al 2019;Fu et al 2019;Yamshchikov et al 2021;Jafaritazehjani et al 2020). Recently, Gehrmann et al (2021 has proposed a new framework which is a live environment to evaluate NLG in a principled and reproducible manner.…”

Section: Improving Evaluation Metricsmentioning

confidence: 99%

Deep Learning for Text Style Transfer: A Survey

Jin

et al. 2022

Computational Linguistics

116

View full text Add to dashboard Cite

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. It has a long history in the field of natural language processing, and recently has re-gained significant attention thanks to the promising performance brought by deep neural models. 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. We discuss the task formulation, existing datasets and subtasks, evaluation, as well as the rich methodologies in the presence of parallel and non-parallel data. We also provide discussions on a variety of important topics regarding the future development of this task.

show abstract

Section: Improving Evaluation Metricsmentioning

confidence: 99%

Deep Learning for Text Style Transfer: A Survey

Jin

et al. 2022

Computational Linguistics

116

View full text Add to dashboard Cite

show abstract

“…Content vs. Style Attributes (3%) It is unclear whether style and content can truly be separated as some content features are important for style or profiling an author (Jafaritazehjani et al, 2020;Bischoff et al, 2020;Patel et al, 2022). Even after filtering, 3% of dimensions of LISA still represent content.…”

Section: Error Analysismentioning

confidence: 99%

“…Prior work has treated the style of a text as separable from the content. Stylistic attributes have included, but are not limited to, linguistic choices in syntax, grammar, spelling, vocabulary, and punctuation (Jafaritazehjani et al, 2020). Style representations should represent two texts with similar stylistic attributes more closely than texts with different attributes independent of what content is present in the texts.…”

Section: Introductionmentioning

confidence: 99%

Learning Interpretable Style Embeddings via Prompting LLMs

Patel,

Rao,

Kothary

et al. 2023

Findings of the Association for Computational Linguistics: EMNLP 2023

View full text Add to dashboard Cite

Style representation learning builds contentindependent representations of author style in text. To date, no large dataset of texts with stylometric annotations on a wide range of style dimensions has been compiled, perhaps because the linguistic expertise to perform such annotation would be prohibitively expensive. Therefore, current style representation approaches make use of unsupervised neural methods to disentangle style from content to create style vectors. These approaches, however, result in uninterpretable representations, complicating their usage in downstream applications like authorship attribution where auditing and explainability is critical. In this work, we use prompting to perform stylometry on a large number of texts to generate a synthetic stylometry dataset. We use this synthetic data to then train humaninterpretable style representations we call LISA embeddings. We release our synthetic dataset (STYLEGENOME) and our interpretable style embedding model (LISA) as resources.

show abstract

“…Many limitations of disentanglement were pointed out in other sentiment-based style transfer studies (e.g., using fix-sized vectors for the latent representations might fail to retain the rich semantic information characterizing long texts), with some of them casting doubt on the feasibility of the style-to-content separation (e.g., Jafaritazehjani et al 2020). As an alternative to the manipulation of latent representations, Dai et al (2019) added a style embedding as an input to their transformer encoder, while Li et al (2020a) directly proposed a novel architecture composed of two generators and no discriminator.…”

Section: Datamentioning

confidence: 99%

From theories on styles to their transfer in text: Bridging the gap with a hierarchical survey

2022

View full text Add to dashboard Cite

Humans are naturally endowed with the ability to write in a particular style. They can, for instance, rephrase a formal letter in an informal way, convey a literal message with the use of figures of speech or edit a novel by mimicking the style of some well-known authors. Automating this form of creativity constitutes the goal of style transfer. As a natural language generation task, style transfer aims at rewriting existing texts, and specifically, it creates paraphrases that exhibit some desired stylistic attributes. From a practical perspective, it envisions beneficial applications, like chatbots that modulate their communicative style to appear empathetic, or systems that automatically simplify technical articles for a non-expert audience. Several style-aware paraphrasing methods have attempted to tackle style transfer. A handful of surveys give a methodological overview of the field, but they do not support researchers to focus on specific styles. With this paper, we aim at providing a comprehensive discussion of the styles that have received attention in the transfer task. We organize them in a hierarchy, highlighting the challenges for the definition of each of them and pointing out gaps in the current research landscape. The hierarchy comprises two main groups. One encompasses styles that people modulate arbitrarily, along the lines of registers and genres. The other group corresponds to unintentionally expressed styles, due to an author’s personal characteristics. Hence, our review shows how these groups relate to one another and where specific styles, including some that have not yet been explored, belong in the hierarchy. Moreover, we summarize the methods employed for different stylistic families, hinting researchers towards those that would be the most fitting for future research.

show abstract

Style versus Content: A distinction without a (learnable) difference?

Cited by 10 publications

References 20 publications

Deep Learning for Text Style Transfer: A Survey

Deep Learning for Text Style Transfer: A Survey

Learning Interpretable Style Embeddings via Prompting LLMs

From theories on styles to their transfer in text: Bridging the gap with a hierarchical survey

Contact Info

Product

Resources

About