Poetry Generation Research Articles

Romanian Style Chinese Modern Poetry Generation with Pre-Trained Model and Direct Preference Optimization

The poetry of distant country with different culture and language is always distinctive and fascinating. Chinese and Romanian belong to Sinitic languages of the Sino-Tibetan language family and Romance languages of the Indo-European language family, which have relatively different syntax and general imagery of literature. Therefore, in this study, we make an attempt that was rarely involved in previous poetry generation research, using modern Chinese as the carrier, and generating modern poetry with Romanian style based on pre-trained model and direct preference optimization. Using a 5-point grading system, human evaluators awarded scores ranging from 3.21 to 3.83 across seven evaluation perspectives for the generated poems, achieving 76.2% to 91.6% of the comparable scores for the Chinese translations of authentic Romanian poems. The coincidence of the 30th to the 50th most frequently occurring poetic images in both generated poems and Romanian poems can reach 58.0–63.3%. Human evaluation and comparative statistical results on poetic imagery show that direct preference optimization is of great help in improving the degree of stylization, and the model can successfully create Chinese modern poems with Romanian style.

UPON: Urdu Poetry Generation Using Deep Learning: A Novel Approach and Evaluation

Poetry represents the oldest and most esteemed literary form, allowing poets to convey ideas while carefully attending to elements such as meaning, coherence, poetic quality, and fluency. Notably, the creation of good poetry entails considerations of rhyme and meter. With the advent of artificial intelligence (AI), significant advancements have been made in automatic text generation, primarily within languages such as English and Chinese. However, the generation of Urdu poetry presents a unique challenge due to the language’s inherent ambiguity, cultural and historical nuances, and the demand for creativity. The existing body of literature has only marginally explored Urdu prose and has almost entirely overlooked the domain of Urdu poetry generation, primarily due to the scarcity of comprehensive training data. In response to this deficiency, this research endeavor addresses this challenge. It begins by introducing a specialized Urdu poetry dataset adhering to a specific meter, ’behr-e-khafeef,’ which incorporates approximately 20,000 couplets from the Rekhta repository. Subsequently, a character-based encoding methodology is proposed to transform these couplets into a numerical representation, assigning a distinct identifier to each character. The generation process initiates with the creation of the first verse through a character-level LSTM, followed by the application of a machine translation technique, specifically sequence-to-sequence learning, to formulate the second verse based on the first. The generated poetry is subjected to evaluation based on metrics, including BLEU scores. Additionally, an expert panel of Urdu poets is engaged to conduct a human assessment of the generated couplets, with the evaluation encompassing critical dimensions such as meaning, coherence, poetic quality, and fluency. Our findings are juxtaposed with existing poetry generation systems, demonstrating a notable advancement in the state-of-the-art, as evidenced by a BLEU score of 0.23. The research culminates with the presentation of prospective avenues for further exploration, aimed at inspiring the scholarly community to enhance the domain of poetry generation and augment existing contributions in this field.

Copiously Quote Classics: Improving Chinese Poetry Generation with historical allusion knowledge

Integrating allusions into poems is an advanced form of human poetry writing, which could clearly express the author’s thoughts and arouse the resonance of readers. However, existing poetry generation works mainly focus on improving the coherence and fluency of poetry, while generating poems with allusion knowledge is rarely considered. To solve this issue, we propose an Allusion-aware Chinese Poetry Generation (ACPG) framework in this study. Concretely, we first release an Allusion-Enriched Poetry (AEP) dataset by linking poems with historical allusions, which might enable a new research direction for poetry generation. Based on this dataset, we design a three-stage learning mechanism to encourage the training stage under a low-resource setting, which can effectively exploit the knowledge of large-scale poetry and allusion data to generate informative allusive poems. Extensive experiments demonstrate the effectiveness of ACPG among a series of proposed baselines. Moreover, the proposed ACPG framework can also be applied to lyrics generation or other controlled text generation tasks, which can incorporate allusion knowledge into the generated results and enhance the meaning and quality of the texts.

Automatic Generation and Evaluation of French-Style Chinese Modern Poetry

Literature has a strong cultural imprint and regional color, including poetry. Natural language itself is part of the poetry style. It is interesting to attempt to use one language to present poetry in another language style. Therefore, in this study, we propose a method to fine-tune a pre-trained model in a targeted manner to automatically generate French-style modern Chinese poetry and conduct a multi-faceted evaluation of the generated results. In a five-point scale based on human evaluation, judges assigned scores between 3.29 and 3.93 in seven dimensions, which reached 80.8–93.6% of the scores of the Chinese versions of real French poetry in these dimensions. In terms of the high-frequency poetic imagery, the consistency of the top 30–50 high-frequency poetic images between the poetry generated by the fine-tuned model and the French poetry reached 50–60%. In terms of the syntactic features, compared with the poems generated by the baseline model, the distribution frequencies of three special types of words that appear relatively frequently in French poetry increased by 12.95%, 15.81%, and 284.44% per 1000 Chinese characters in the poetry generated by the fine-tuned model. The human evaluation, poetic image distribution, and syntactic feature statistics show that the targeted fine-tuned model is helpful for the spread of language style. This fine-tuned model can successfully generate modern Chinese poetry in a French style.

Ancient poetry generation based on bidirectional LSTM model neural network

Automatic generation of ancient poems has been a research hotspot in the field of artificial intelligence, which is of great significance in cultural inheritance and literary creation. Due to the strict tonal patterns and complex structural rules of classical Chinese poetry, generating classical Chinese poetry has been a challenging task for both human poets and computer programs. In this paper, we propose a neural network model based on the introduction of an attention mechanism, which combines a bidirectional LSTM model and an attention mechanism to solve some problems in the traditional automatic generation model of ancient poems. The model is trained using the dataset of ancient poems studied by previous researchers, and the performance of the model is evaluated and analyzed by evaluating the metrics BLEU, Perplexity, Generation Effect and Linguistic Coherence. The experimental results show that the model exhibits good performance and excellent results on the task of automatic generation of ancient poems.

Exploring the potential of data augmentation in poetry generation with small-scale corpora

Poetry generation is a complex task in the field of natural language processing, especially when working with small datasets. Data augmentation techniques have been shown to be an effective way to improve the performance of deep learning models in various tasks, including image classification and speech recognition. Therefore, this study focuses on exploring the impact of four different data augmentation methods - Synonym Replacement, Random Insertion, Random Swap, and Random Deletion - on the performance of poetry generation with a small poetry dataset. The results of the study reveal that Random Insertion performed well in terms of Bilingual Evaluation Understudy (BLEU), Recall-Oriented Understudy for Gisting Evaluation (ROUGE), and manual evaluation when compared to other data augmentation techniques. Synonym Replacement performed poorly in all three evaluations. This study confirms the potential value of data augmentation technology in poetry generation tasks and provides innovative perspectives and directions for future research in this area. Data augmentation can be employed to help address the problem of limited data in poetry generation tasks and enhance the efficiency of deep learning models. Future research could focus on exploring more advanced data augmentation techniques and their impact on poetry generation tasks.

PoetryDiffusion: Towards Joint Semantic and Metrical Manipulation in Poetry Generation

Controllable text generation is a challenging and meaningful field in natural language generation (NLG). Especially, poetry generation is a typical one with well-defined and strict conditions for text generation which is an ideal playground for the assessment of current methodologies. While prior works succeeded in controlling either semantic or metrical aspects of poetry generation, simultaneously addressing both remains a challenge. In this paper, we pioneer the use of the Diffusion model for generating sonnets and Chinese SongCi poetry to tackle such challenges. In terms of semantics, our PoetryDiffusion model, built upon the Diffusion model, generates entire sentences or poetry by comprehensively considering the entirety of sentence information. This approach enhances semantic expression, distinguishing it from autoregressive and large language models (LLMs). For metrical control, its constraint control module which can be trained individually enables us to flexibly incorporate a novel metrical controller to manipulate and evaluate metrics (format and rhythm). The denoising process in PoetryDiffusion allows for the gradual enhancement of semantics and flexible integration of the metrical controller which can calculate and impose penalties on states that stray significantly from the target control distribution. Experimental results on two datasets demonstrate that our model outperforms existing models in terms of automatic evaluation of semantic, metrical, and overall performance as well as human evaluation. Codes are released to https://github.com/ChorlingLau/PoetryDiffusion.

A Polishing Model for Machine-Generated Ancient Chinese Poetry

Machine poetry generation has been studied for decades, among which ancient Chinese poetry is still challenging in the field of poetry generation due to its unique regularity and rhythm. The quality improvement of ancient Chinese poetries is one of the most promising research areas of ancient Chinese Natural Language Processing. This paper proposes an ancient Chinese poetry polishing model, which is used for polishing to obtain high-quality ancient Chinese poetry. The model consists of a detection network and a correction network. The detection network based on BiLSTM and CRF is used to detect different types of low-quality words in poems. The correction network based on the BERT model is used to modify the detected low-quality words in the global context. The polishing process is iteratively performed until the model judges that there are no low-quality words in the poem. The results show that the polished poems are improved in multiple evaluations. Compared with existing polishing models, the model proposed in this paper performs better in both automatic evaluation and human evaluation when the number of parameters is reduced.

A Systematic Literature Review on Text Generation using Deep Neural Networks

Text generation using deep neural networks has become an exciting area of research. Deep neural networks, such as recurrent neural networks (RNNs) and transformers, have shown remarkable capabilities in generating coherent and contextually relevant text. By training these models on large text corpora, they learn to capture the underlying patterns and structures of the language. This enables them to generate new text that resembles the style and content of the training data. Text generation using deep neural networks has a wide range of applications, including chatbots, language translation, poetry generation, and even code generation.

Pashto poetry generation: deep learning with pre-trained transformers for low-resource languages.

Generating poetry using machine and deep learning techniques has been a challenging and exciting topic of research in recent years. It has significance in natural language processing and computational linguistics. This study introduces an innovative approach to generate high-quality Pashto poetry by leveraging two pre-trained transformer models, LaMini-Cerebras-590M and bloomz-560m. The models were trained on an extensive new and quality Pashto poetry dataset to learn the underlying complex patterns and structures. The trained models are then used to generate new Pashto poetry by providing them with a seed text or prompt. To evaluate the quality of the generated poetry, we conducted both subjective and objective evaluations, including human evaluation. The experimental results demonstrate that the proposed approach can generate Pashto poetry that is comparable in quality to human-generated poetry. The study provides a valuable contribution to the field of Pashto language and poetry generation and has potential applications in natural language processing and computational linguistics.

Analyzing the Integration Interpretation of Civics and Family Poetry Culture in University Language Programs Based on Deep Learning Models

Abstract In this paper, firstly, we start from the imagery in the poem, excavate the resources of the ideology and politics in the university language poetry, and put forward the feasible path of the integration of the ideology and politics of the university language course and the poetry culture of the family training. The research on poetry culture in university language is conducted using a deep learning model. Based on the deep learning model, keywords in the poetry text are extracted and extended, taking into account the issue of word splitting and lexical annotation. The sequence coding of word vectors is needed so as to construct a poetry generation model based on RNNPG, and then the research and analysis on the concept of ideology and politics of the poetry culture in the university language are carried out. The results show that in terms of model performance, the RNNPG model is 1.17 overall, and the overall generation effect is more ideal compared to other single discriminator generation models. The acceptance level of sophomores is generally above 0.29, which is better than that of freshmen, and the teaching of ancient poems under the perspective of “curriculum politics” is especially necessary to improve their cultural taste and aesthetic interest. This study is conducive to the personalized development of students and the formation of correct concepts.

인공지능 기반 한국어 시 생성 시스템 개발 연구

인공지능 기반 한국어 시 생성 시스템 개발 연구

Poetry4painting: Diversified poetry generation for large-size ancient paintings based on data augmentation

Chinese painting poetry is an extraordinary art form, which not only describes the painting contexts but also grasps the sentiment of the painters. In this paper, we propose an automatic poetry generation method Poetry4painting, which enhances the poetry diversity for large-size ancient paintings. The basic framework is based on multiple modern sentences, that are first captioned from the ancient painting and then used to generate a poem using CNN and LSTM. To solve the repeatability issue of this framework, four kinds of data augmentation are employed during online processing, including quantity, shape, surrounding, and object augmentation. In offline training, data augmentation is also used to create an image caption dataset with over 1500 painting images and 7500 captions. Through ablation studies, evaluations of poetry qualities and diversities, and comparisons with other methods, we demonstrate the validity of the proposed method.

TPoet: Topic-Enhanced Chinese Poetry Generation

Chinese poetry generation has been a challenging part of natural language processing due to the unique literariness and aesthetics of poetry. In most cases, the content of poetry is topic related. In other words, specific thoughts or emotions are usually expressed regarding given topics. However, topic information is rarely taken into consideration in current studies about poetry generation models. In this article, we propose a topic-enhanced Chinese poetry generation model called TPoet in which the topic model is integrated into the Transformer-based auto-regressive text generation model. By feeding topic information to the input layer and heterogeneous attention mechanism, TPoet can implicitly learn the latent information of topic distribution. In addition, by setting multiple identifiers such as segment, rhyme, and tone, the model can explicitly learn the constraints of generated poems. Extensive experimental results show that the quality of TPoet-generated poems outperforms the current advanced models or systems, and the topic consistency and diversity in generated poems have been significantly improved as well.

Classical Chinese poetry generation from vernacular Chinese: a word-enhanced supervised approach

Classical Chinese poetry generation from vernacular Chinese: a word-enhanced supervised approach

Poet and Psychologist: A Conversation

ABSTRACT I consider poetry composition from both the “inside” view of a poet and the “outside” view of a cognitive psychologist. From the perspective of a psychologist, I review behavioral and neural studies of the reception and generation of poetry, with emphasis on metaphor and symbolism. Taking the perspective of a poet, I discuss how the seeds for a poem may arise. Finally, I consider the prospects for future developments in a field of computational neurocognitive poetics.

Poetry Generation for Indonesian Pantun Using SeqGAN and GPT-2

Pantun is a traditional Malay poem consisting of four lines: two lines of deliverance and two lines of messages. Each ending-line word in pantun forms an ABAB rhyme pattern. In this work, we automatically generated Indonesian pantun by applying two existing generative models: Sequential GAN (SeqGAN) and Generative Pre-trained Transformer 2 (GPT-2). We also created a 13K Indonesian pantun dataset by collecting pantun from various sources. We evaluated how well each model produced pantun by its formedness. Measured by two aspects: structure and rhyme. GPT-2 performs better with a margin of 27.57% than SeqGAN in forming the structure and 22.79% better in making rhyming patterns.

The Print Media and the Evolution of Third-Generation Nigerian Poetry

Previous literary scholarship and critical allusions to the evolution of Nigerian literature have unanimously acknowledged the emergence of a third generation of Nigerian poetry. This development which gained accentuation between 1980 and 2000, threw up an avalanche of successors to the second-generation poets, notably: Odia Ofeimun, Tanure Ojaide, Niyi Osundare, Okinba Launko, Catherine Acholonu and Omolara Ogundipe-Leslie. Despite the visible numerical strength of this new generation of writers; their aggregate prolificity and their gross contribution to our national creative output, adequate seminal attention has not been paid to the role of the print media in the catalysation of this milieu. This study, therefore, situates the third generation of Nigerian poets and their works, within the context that they are bonafide products of the print media. They sprouted and flourished in an era where creative activity, triggered by national socio-economic realities, bloomed without collateral and commensurate outlets for the expression of this creative outburst. They have continued to sustain the richness and robustness of Nigeria’s poetic traditions. Their continuing commitment to socio-political engagements has reinforced the enterprise of preceding generations. This study will constitute a veritable reference material on the emerging corpus of critical discourse on the evolution of Nigerian poetry, deriving from the pages of Nigerian newspapers, en route to the 20th century.

"Loquela gravis iuvat": Gower's O deus immense and the Place of Poetry, 1398–1400

Abstract: John Gower's medium-length Latin poem O deus immense , little thought of even by most Gowerians, brings his career into focus. O deus immense synthesizes strands of Gower's self-presentation pursued far more diffusely in his titanic trilogy Mirour de l'omme – Vox clamantis – Confessio Amantis . The place Gower clears for poetry in 1398, 1399, or 1400—the date of O deus immense is uncertain, a point addressed here at length—is characterized by its public, monitory, prophetic, and enigmatic dimensions. Focusing on the historically remote genre of political prophecy, this essay compares O deus immense with other Gowerian and non-Gowerian English political verse in English and Latin datable to 1398–1400, including Bede's Prophecy , an anonymous rhyming English poem of 1400 inedited until recently and therefore, like O deus immense , not factoring into previous critical assessments of the poetry of the Lancastrian coup. Attending to the complex relationship between poetics and politics at the turn of the fifteenth century in England, the essay positions O deus immense as pivotal in Gower's career and essential for an evaluation of what he contributed to the first generation of Lancastrian poetry. Gower's "most significant role" in the field of English political poetry, 1398–1400, was as its leading theorist and most skillful advocate, a role he plays most assiduously in O deus immense ; but it is both a credit to his depth of ambition and an explanation of the often violent contortions of his late Latin style that Gower's arguments for a poetry of moral clarity, social urgency, and political muscle transcend the very ideological commitments that transparently motivate them.

DeepCreativity: measuring creativity with deep learning techniques

Measuring machine creativity is one of the most fascinating challenges in Artificial Intelligence. This paper explores the possibility of using generative learning techniques for automatic assessment of creativity. The proposed solution does not involve human judgement, it is modular and of general applicability. We introduce a new measure, namely DeepCreativity, based on Margaret Boden’s definition of creativity as composed by value, novelty and surprise. We evaluate our methodology (and related measure) considering a case study, i.e., the generation of 19th century American poetry, showing its effectiveness and expressiveness.