TSTBench: A Comprehensive Benchmark for Text Style Transfer.

We introduce a simple method for text style transfer that frames style transfer as denoising: we synthesize a noisy corpus and treat the source style as a noisy version of the target style. To control for aspects such as preserving meaning while modifying style, we propose a reranking approach in the data synthesis phase. We evaluate our method on three novel style transfer tasks: transferring between British and American varieties, text genres (formal vs. casual), and lyrics from different musical genres. By measuring style transfer quality, meaning preservation, and the fluency of generated outputs, we demonstrate that our method is able both to produce high-quality output while maintaining the flexibility to suggest syntactically rich stylistic edits.

Existing text style transfer (TST) methods rely on style classifiers to disentangle the text’s content and style attributes for text style transfer. While the style classifier plays a critical role in existing TST methods, there is no known investigation on its effect on the TST methods. In this paper, we conduct an empirical study on the limitations of the style classifiers used in existing TST methods. We demonstrated that the existing style classifiers cannot learn sentence syntax effectively and ultimately worsen existing TST models’ performance. To address this issue, we propose a novel Syntax-Aware Controllable Generation (SACG) model, which includes a syntax-aware style classifier that ensures learned style latent representations effectively capture the sentence structure for TST. Through extensive experiments on two popular text style transfer tasks, we show that our proposed method significantly outperforms twelve state-of-the-art methods. Our case studies have also demonstrated SACG’s ability to generate fluent target-style sentences that preserved the original content.

The study presents a scientometric analysis of the development of Malay linguistics research over the past three decades. Due to rapid technological changes in linguistics and policy needs, this study examines publication trends, institution and country contributions, themes, and co-citation cluster networks to provide an overview of the field's intellectual structure. The study utilises CiteSpace version 6.3. R3 Advance to analyse records from the Scopus database (1994-2024). The findings indicate substantial growth in publication output, with Malaysia publishing the most papers, followed by Indonesia, Australia, Singapore, and Brunei Darussalam. Universiti Kebangsaan Malaysia, Universiti Sains Malaysia, Universiti Malaya, Universiti Putra Malaysia, Universiti Teknologi Mara, and International Islamic University of Malaysia are the top contributing institutions to research in Malay linguistics. The emerging themes indicate a shift in research focus from traditional linguistic studies toward computational linguistics, digital text processing, and AI-driven applications. The key research clusters were identified, spanning dialectology, historical linguistics, sociolinguistics, and computational linguistics. Among influential contributors, Asmah Haji Omar, with a citation count of 55 and burstness of 4.08, is the most influential author, significantly shaping the discipline. The GEMA Online Journal of Language Studies has the highest citation bursts (1.81), which are crucial in advancing research in computational linguistics, sociolinguistics, and language identity. This study provides critical insights into the evolution of Malay linguistics research, offering valuable evidence to guide researchers and education policymakers in formulating strategic policies for academic innovation and disciplinary advancement.

The dramatic expansion of modern linguistic research and enhanced accuracy of linguistic analysis have become a reality due to the ability of artificial neural networks not only to learn and adapt, but also carry out automate linguistic analysis, select, modify and compare texts of various types and genres. The purpose of this article and the journal issue as a whole is to present modern areas of research in computational linguistics and linguistic complexology, as well as to define a solid rationale for the new interdisciplinary field, i.e. discourse complexology. The review of trends in computational linguistics focuses on the following aspects of research: applied problems and methods, computational linguistic resources, contribution of theoretical linguistics to computational linguistics, and the use of deep learning neural networks. The special issue also addresses the problem of objective and relative text complexity and its assessment. We focus on the two main approaches to linguistic complexity assessment: “parametric approach” and machine learning. The findings of the studies published in this special issue indicate a major contribution of computational linguistics to discourse complexology, including new algorithms developed to solve discourse complexology problems. The issue outlines the research areas of linguistic complexology and provides a framework to guide its further development including a design of a complexity matrix for texts of various types and genres, refining the list of complexity predictors, validating new complexity criteria, and expanding databases for natural language.

Text style transfer and paraphrasing of texts are actively growing areas of NLP, dozens of methods for solving these tasks have been recently introduced. In both tasks, the system is supposed to generate a text which should be semantically similar to the input text. Therefore, these tasks are dependent on methods of measuring textual semantic similarity. However, it is still unclear which measures are the best to automatically evaluate content preservation between original and generated text. According to our observations, many researchers still use BLEU-like measures, while there exist more advanced measures including neural-based that significantly outperform classic approaches. The current problem is the lack of a thorough evaluation of the available measures. We close this gap by conducting a large-scale computational study by comparing 57 measures based on different principles on 19 annotated datasets. We show that measures based on cross-encoder models outperform alternative approaches in almost all cases. We also introduce the Mutual Implication Score (MIS), a measure that uses the idea of paraphrasing as a bidirectional entailment and outperforms all other measures on the paraphrase detection task and performs on par with the best measures in the text style transfer task.

Text style transfer (TST) without parallel data has achieved some practical success. However, most of the existing unsupervised text style transfer methods suffer from (i) requiring massive amounts of non-parallel data to guide transferring different text styles. (ii) colossal performance degradation when fine-tuning the model in new domains. In this work, we propose DAML-ATM (Domain Adaptive Meta-Learning with Adversarial Transfer Model), which consists of two parts: DAML and ATM. DAML is a domain adaptive meta-learning approach to learning general knowledge in multiple heterogeneous source domains, capable of adapting to new unseen domains with a small amount of data. Moreover, we propose a new unsupervised TST approach Adversarial Transfer Model (ATM), composed of a sequence-to-sequence pre-trained language model and uses adversarial style training for better content preservation and style transfer. Results on multi-domain datasets demonstrate that our approach generalizes well on unseen low-resource domains, achieving state-of-the-art results against ten strong baselines.

Text style transfer (TST) involves altering the linguistic style of a text while preserving its style-independent content.This paper focuses on sentiment transfer, a popular TST subtask, across a spectrum of Indian languages: Hindi, Magahi, Malayalam, Marathi, Punjabi, Odia, Telugu, and Urdu, expanding upon previous work on English-Bangla sentiment transfer (Mukherjee et al., 2023a).We introduce dedicated datasets of 1,000 positive and 1,000 negative style-parallel sentences for each of these eight languages.We then evaluate the performance of various benchmark models categorized into parallel, non-parallel, cross-lingual, and shared learning approaches, including the Llama2 and GPT-3.5 large language models (LLMs).Our experiments highlight the significance of parallel data in TST and demonstrate the effectiveness of the Masked Style Filling (MSF) approach (Mukherjee et al., 2023a) in non-parallel techniques.Moreover, crosslingual and joint multilingual learning methods show promise, offering insights into selecting optimal models tailored to the specific language and task requirements.To the best of our knowledge, this work represents the first comprehensive exploration of the TST task as sentiment transfer across a diverse set of languages.

William C. Mann passed away on August 13, 2004, after a long struggle with leukemia. He is known to readers of Computational Linguistics as a keen supporter and past president (1987) of ACL, a pioneer in the development of text generation as a field of research in computational linguistics, the originator of rhetorical structure theory (RST) and dialogue game theory (DGT), and the developer (together with David Weber) of dialect adaptation as a technique within machine translation. And he’ll be remembered for these and many other contributions, but above all, in this context, he’ll be remembered as a unique visionary in computational linguistics (CL) and as a truly wonderful and extraordinarily generous colleague. Bill Mann made huge contributions to many people’s intellectual development, to their careers, and to their lives in general. What follows are my own recollections of him, selected from a crowded jungle of memories. I will leave it to others to fill out the picture, but I shall attempt to give an indication of the intellectual and historical context in which Bill was working and also of the network of researchers that he created; his ideas have radiated throughout this network of scholars around the world. I met Bill in April 1980 and worked as a research linguist on a succession of projects he directed, or codirected, at the Information Sciences Institute (ISI) at the University of Southern California (USC), until I left the institute to take up a position at the University of Sydney in August 1988. Through that period (which happened to coincide, more or less, with the Reagan years), not only was Bill a very wise and wonderful project leader, but he also became a mentor and guide and a true friend. I followed from afar the further developments in his life: his departure from ISI and move to Kenya with his family to take up a position with the Summer Institute of Linguistics (SIL) as a consultant responsible for work on discourse and computational matters, their return to the United States, and his new phase of research with initiatives in RST and DGT.

Unsupervised text style transfer aims to alter text styles while preserving the content, without aligned data for supervision. Existing seq2seq methods face three challenges: 1) the transfer is weakly interpretable, 2) generated outputs struggle in content preservation, and 3) the trade-off between content and style is intractable. To address these challenges, we propose a hierarchical reinforced sequence operation method, named Point-Then-Operate (PTO), which consists of a high-level agent that proposes operation positions and a low-level agent that alters the sentence. We provide comprehensive training objectives to control the fluency, style, and content of the outputs and a mask-based inference algorithm that allows for multi-step revision based on the single-step trained agents. Experimental results on two text style transfer datasets show that our method significantly outperforms recent methods and effectively addresses the aforementioned challenges.

The style transfer task (here style is used in a broad “authorial” sense with many aspects including register, sentence structure, and vocabulary choice) takes text input and rewrites it in a specified target style preserving the meaning, but altering the style of the source text to match that of the target. Much of the existing research on this task depends on the use of parallel datasets. In this work we employ recent results in unsupervised cross-lingual language modeling (XLM) and machine translation to effect style transfer while treating the input data as unaligned. First, we show that adding “content embeddings” to the XLM which capture human-specified groupings of subject matter can improve performance over the baseline model. Evaluation of style transfer has often relied on metrics designed for machine translation which have received criticism of their suitability for this task. As a second contribution, we propose the use of a suite of classical stylometrics as a useful complement for evaluation. We select a few such measures and include these in the analysis of our results.

Given that context-free grammars (CFG) cannot adequately describe natural languages, grammar formalisms beyond CFG that are still computationally tractable are of central interest for computational linguists. This book provides an extensive overview of the formal language landscape between CFG and PTIME, moving from Tree Adjoining Grammars to Multiple Context-Free Grammars and then to Range Concatenation Grammars while explaining available parsing techniques for these formalisms. Although familiarity with the basic notions of parsing and formal languages is helpful when reading this book, it is not a strict requirement. The presentation is supported with many illustrations and examples relating to the different formalisms and algorithms, and chapter summaries, problems and solutions. The book will be useful for students and researchers in computational linguistics and in formal language theory.

Categorization is a central capability of human cognition, and a number of theories have been developed to account for properties of categorization. Even though many tasks in semantics also involve categorization of some kind, theories of categorization do not play a major role in contemporary research in computational linguistics. This paper follows the idea that embedding-based models of semantics lend themselves well to being formulated in terms of classical categorization theories. The benefit is a space of model families that enables (a) the formulation of hypotheses about the impact of major design decisions, and (b) a transparent assessment of these decisions. We instantiate this idea on the task of frame-semantic frame identification. We define four models that cross two design variables: (a) the choice of prototype vs. exemplar categorization, corresponding to different degrees of generalization applied to the input; and (b) the presence vs. absence of a fine-tuning step, corresponding to generic vs. task-adaptive categorization. We find that for frame identification, generalization and task-adaptive categorization both yield substantial benefits. Our prototype-based, fine-tuned model, which combines the best choices for these variables, establishes a new state of the art in frame identification.

The PAP900 dataset centers on the semantic relationship between affective words in Portuguese. It contains 900 word pairs, each annotated by at least 30 human raters for both semantic similarity and semantic relatedness. In addition to the semantic ratings, the dataset includes the word categorization used to build the word pairs and detailed sociodemographic information about annotators, enabling the analysis of the influence of personal factors on the perception of semantic relationships. Furthermore, this article describes in detail the dataset construction process, from word selection to agreement metrics.Data was collected from Portuguese university psychology students, who completed two rounds of questionnaires. In the first round annotators were asked to rate word pairs on either semantic similarity or relatedness. The second round switched the relation type for most annotators, with a small percentage being asked to repeat the same relation. The instructions given emphasized the differences between semantic relatedness and semantic similarity, and provided examples of expected ratings of both.There are few semantic relations datasets in Portuguese, and none focusing on affective words. PAP900 is distributed in distinct formats to be easy to use for both researchers just looking for the final averaged values and for researchers looking to take advantage of the individual ratings, the word categorization and the annotator data. This dataset is a valuable resource for researchers in computational linguistics, natural language processing, psychology, and cognitive science.

Much research in computational linguistics and artificial intelligence has gone into characterizing natural language discourse structures, developing discourse models, and using these to develop more useful natural language interfaces to databases and expert systems. However, this knowledge about the nature of discourse has not been used to any advantage in graphical expert system interfaces. This paper proposes that discourse elements and discourse models used in the world of natural language understanding can be extended to graphical expert system interfaces, promoting continuity of interactions and allowing relationships among interactions to be exploited, redundancy reduced, operations tracked, and higher-level operations recognized. fj-^-, V' 1 f, T 1 4^ ^'' Jj-v' *. - fci L

The intrinsic value of Indian traditional knowledge sources in modern sciences is examined by evaluating 2431 publications of the last 200 years culled from the Scopus database. The results showed growing interest among scholars and researchers from about 60 countries who have researched these knowledge sources. Although major publications belong to the fields of humanities and social sciences, researchers in the areas of computer science, engineering, medicine, and environmental science have shown research interests in these domains. A shift of research interest has been noticed as researchers in computational linguistics, semantics, cultural heritage, and sustainable development have identified emerging themes. The analysis of secondary sources showed that publications appeared between 1700 and 1950 on various traditional sources, mostly by non-Indic authors. This research concludes that the fusion of traditional knowledge in fields such as medicine, agriculture, and environmental sciences may be helpful for the study of contemporary problems, achieving sustainable goals, and enhancing both academic inquiry and practical applications. It is urged that the Governments in India should support researchers to work in the areas of social welfare using the Indian traditional knowledge sources.