General Language Research Articles

Determinants of intention to use ChatGPT for obtaining shopping information

ABSTRACT Drawing on the “Technology Acceptance Model” (TAM) and “Elaboration Likelihood Model,” this study explored the determinants of intention to use ChatGPT to obtain shopping information. We extended TAM by incorporating perceived enjoyment. Data were collected from 412 Malaysians and analyzed using the “Partial Least Squares” approach. Findings confirmed the significance of TAM and perceived enjoyment, revealing that information accuracy, completeness, and diagnosticity greatly influence perceived usefulness. These results enhance the understanding of ChatGPT’s use in shopping information gathering, extend TAM, and highlight how informational cues impact perceived usefulness, guiding developers and marketers to encourage customer use of text generators.

A longitudinal study of breastmilk feeding duration, EEG power and early academic skills

BackgroundThe cognitive benefits of breastfeeding are widely recognized; however, its effects on brain development and later academic skills require further examination. This study aimed to examine the longitudinal relations between breastmilk feeding, neurophysiological changes, and early academic skills. MethodsIn the Growing Up in Singapore Towards healthy Outcomes (GUSTO) birth cohort, breastmilk feeding practices were collected every 3 months from 3 weeks to 18 months postpartum. Resting electroencephalography (EEG) was recorded at 18 months and power spectral density was derived. The outcomes were a set of early academic assessments administered at age 4 (n = 810). Structural equation modelling was used to investigate EEG power as a mediator between breastmilk duration and early academic skills. ResultsBreastmilk feeding for ≥12 months was associated with better general knowledge, numeracy, and language at age 4 compared to shorter durations of breastmilk feeding (Cohen's d: 1.53–17.44). Linear regression showed that breastmilk duration was negatively and positively associated with low- (i.e., delta, theta) and high-frequency power (i.e., gamma), respectively (Cohen's f2: 0.03–0.09). After adjusting for demographic and child baseline covariates, a decrease in absolute and relative delta, as well as relative theta was associated with better general knowledge and numeracy (Cohen's f2: 0.16–0.25). Relative delta power provided an indirect path between breastmilk duration and early academic skills (x2: 18.390, p = 0.010; CFI: 0.978; TLI: 0.954; RMSEA: 0.040). ConclusionsExtended breastmilk feeding is associated with reduced low-frequency power and better early academic skills, suggesting benefits to brain development. Additional research to confirm this finding is warranted.

Ghosts and the machine: testing the use of Artificial Intelligence to deliver historical life course biographies from big data

This article presents the findings of an experiment in the use of Artificial Intelligence text generation processes to convert historical ‘big data’ into narrative text. Using an extensive collection of records pertaining to the Australian colonial settlement of Norfolk Island in the South Pacific (1788–1814), we investigate Generative Large Language Model technology for converting tabulated data from the site into short pieces of novel text, describing the lives of transported convicts and free individuals recorded in our databases. These personalized stories are assessed for fluency and factual correctness. Using this process, we uncover some instructive problems and caveats. We detect AI’s inherent tendency toward bias and uncritical perspectives, including potentially offensive stereotypes. We also discover an unwelcome tendency to summarize data. So, whilst the outputs are for the most part effective and functional, we find that the best results still require artful human intervention to fully capture the most human aspects of history and heritage research.

Text-guided image-to-sketch diffusion models

Recently, with the continuous advancement of deep learning techniques, research on sketch synthesis has been progressing. However, existing methods still face challenges in generating human-like freehand sketches from real-world natural images at both object and scene levels. To address this, we propose SketchDiffusion, a text-guided freehand sketch synthesis method based on conditional stable diffusion. In SketchDiffusion, we design a novel image enhancing module to efficiently extract high-quality image features. Moreover, we utilize additional guidance from global and local features extracted by a U-shaped diffusion guidance network to control the noise addition and denoising process of the diffusion model, thereby significantly improving controllability and performance in freehand sketch synthesis. Beyond the model architecture, we leverage the designed BLIP-based text generation method to create 70,280 text prompts for foreground, background, and panorama sketch synthesis in the extensive SketchyCOCO dataset, thereby improving the overall effectiveness of model training. Compared to the state-of-the-art methods, our proposed SketchDiffusion has shown an average improvement of over 16.4%, 16.75%, and 12.8% on three quantitative metrics (sketch recognition, sketch-based retrieval, and user perceptual study), respectively. Furthermore, our approach not only excels in synthesizing freehand sketches containing multiple abstract objects but also has multiple applications in supporting human–computer interaction.

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.

Current Status of ChatGPT Use in Medical Education: Potentials, Challenges, and Strategies.

ChatGPT, a generative pretrained transformer, has garnered global attention and sparked discussions since its introduction on November 30, 2022. However, it has generated controversy within the realms of medical education and scientific research. This paper examines the potential applications, limitations, and strategies for using ChatGPT. ChatGPT offers personalized learning support to medical students through its robust natural language generation capabilities, enabling it to furnish answers. Moreover, it has demonstrated significant use in simulating clinical scenarios, facilitating teaching and learning processes, and revitalizing medical education. Nonetheless, numerous challenges accompany these advancements. In the context of education, it is of paramount importance to prevent excessive reliance on ChatGPT and combat academic plagiarism. Likewise, in the field of medicine, it is vital to guarantee the timeliness, accuracy, and reliability of content generated by ChatGPT. Concurrently, ethical challenges and concerns regarding information security arise. In light of these challenges, this paper proposes targeted strategies for addressing them. First, the risk of overreliance on ChatGPT and academic plagiarism must be mitigated through ideological education, fostering comprehensive competencies, and implementing diverse evaluation criteria. The integration of contemporary pedagogical methodologies in conjunction with the use of ChatGPT serves to enhance the overall quality of medical education. To enhance the professionalism and reliability of the generated content, it is recommended to implement measures to optimize ChatGPT's training data professionally and enhance the transparency of the generation process. This ensures that the generated content is aligned with the most recent standards of medical practice. Moreover, the enhancement of value alignment and the establishment of pertinent legislation or codes of practice address ethical concerns, including those pertaining to algorithmic discrimination, the allocation of medical responsibility, privacy, and security. In conclusion, while ChatGPT presents significant potential in medical education, it also encounters various challenges. Through comprehensive research and the implementation of suitable strategies, it is anticipated that ChatGPT's positive impact on medical education will be harnessed, laying the groundwork for advancing the discipline and fostering the development of high-caliber medical professionals.

Evaluating Named Entity Recognition: A comparative analysis of mono- and multilingual transformer models on a novel Brazilian corporate earnings call transcripts dataset

Since 2018, when the Transformer architecture was introduced, Natural Language Processing has gained significant momentum with pre-trained Transformer-based models that can be fine-tuned for various tasks. Most models are pre-trained on large English corpora, making them less applicable to other languages, such as Brazilian Portuguese. In our research, we identified two models pre-trained in Brazilian Portuguese (BERTimbau and PTT5) and two multilingual models (mBERT and mT5). BERTimbau and mBERT use only the Encoder module, while PTT5 and mT5 use both the Encoder and Decoder. Our study aimed to evaluate their performance on a financial Named Entity Recognition (NER) task and determine the computational requirements for fine-tuning and inference. To this end, we developed the Brazilian Financial NER (BraFiNER) dataset, comprising sentences from Brazilian banks’ earnings calls transcripts annotated using a weakly supervised approach. Additionally, we introduced a novel approach that reframes the token classification task as a text generation problem. After fine-tuning the models, we evaluated them using performance and error metrics. Our findings reveal that BERT-based models consistently outperform T5-based models. While the multilingual models exhibit comparable macro F1-scores, BERTimbau demonstrates superior performance over PTT5. In terms of error metrics, BERTimbau outperforms the other models. We also observed that PTT5 and mT5 generated sentences with changes in monetary and percentage values, highlighting the importance of accuracy and consistency in the financial domain. Our findings provide insights into the differing performance of BERT- and T5-based models for the NER task.

Unmasking large language models by means of OpenAI GPT-4 and Google AI: A deep instruction-based analysis

Large Language Models (LLMs) have become a hot topic in AI due to their ability to mimic human conversation. This study compares the open artificial intelligence generative pretrained transformer-4 (GPT-4) model, based on the (GPT), and Google's artificial intelligence (AI), which is based on the Bidirectional Encoder Representations from Transformers (BERT) framework in terms of the defined capabilities and the built-in architecture. Both LLMs are prominent in AI applications. First, eight different capabilities were identified to evaluate these models, i.e. translation accuracy, text generation, factuality, creativity, intellect, deception avoidance, sentiment classification, and sarcasm detection. Next, each capability was assessed using instructions. Additionally, a categorized LLM evaluation system has been developed by means of using ten research questions per category based on this paper's main contributions from a prompt engineering perspective. It should be highlighted that GPT-4 and Google AI successfully answered 85 % and 68,7 % of the study prompts, respectively. It has been noted that GPT-4 better understands prompts than Google AI, even with verbal flaws, and tolerates grammatical errors. Moreover, the GPT-4 based approach was more precise, accurate, and succinct than Google AI, which was sometimes verbose and less realistic. While GPT-4 beats Google AI in terms of translation accuracy, text generation, factuality, intellectuality, creativity, and deception avoidance, Google AI outperforms the former when considering sarcasm detection. Both sentiment classification models did work properly. More importantly, a human panel of judges was used to assess and evaluate the model comparisons. Statistical analysis of the judges' ratings revealed more robust results based on examining the specific uses, limitations, and expectations of both GPT-4 and Google AI-based approaches. Finally, the two approaches' transformers, parameter sizes, and attention mechanisms have been examined.

From Optimism to Concern: Unveiling Sentiments and Perceptions Surrounding ChatGPT on Twitter

Artificial intelligence (AI) technologies, as a product of processes aimed at imitating human intelligence through computers and software, affect our daily lives in various aspects, including cultural, technological, and economic. The advent of ChatGPT, developed by OpenAI, signifies a pivotal AI advancement in human language comprehension and generation, heralding profound implications across societal, economic, and cultural dimensions. However, a notable gap exists in scholarly literature concerning the examination of perceptions and discussions surrounding ChatGPT. This study aims to address this gap by analyzing Twitter conversations, comprising 1.1 million tweets containing “ChatGPT” or “#ChatGPT” between December 1 2022 and June 1 2023, with geolocation data. Employing a text as data approach encompassing text, sentiment, and semantic network analyses, sentiment polarity and lexical patterns were explained, while semantic network analysis revealed central expressions and prominent themes in Twitter discussions. The study highlights the social effects of AI technologies from the perspective of Twitter users and reveals the sentimental tendency and themes in tweets with geographical and economic dimensions. The findings reflect the prevalence of positive sentiment and hype toward ChatGPT while there are also concerns regarding privacy, cybersecurity, and misuse of AI tools. More importantly, content generation, technical aspects, business applications, educational use, and competitors are among the main themes of ChatGPT-related discussions on Twitter.

The Dilemma and Countermeasures of Rural Teachers Professional Development in the Smart Era

The emergence of artificial intelligence has caused a great impact on the field of education, and the powerful natural language generation and contextual dialog capabilities have brought various opportunities and challenges to the field of education. As one of the elements in education and teaching, teachers professional development will certainly be affected by the wave of artificial intelligence. Based on the problems in the process of rural teachers professional development in the era of artificial intelligence, the paper sorted out the action logic of artificial intelligence to boost the professional development of rural teachers and explored the strategy of artificial intelligence to boost the professional development of rural teachers.

Language Choice and the Problematics of Ideology in the Pre- and Post-Independence Ghanaian Press: A Historical and Cultural Analysis

This study adds to the existing literature on the history of the Ghanaian press from pre-colonial times to 1992, focusing on language: its choice and ideological, socio-cultural, and political ramifications. While the history of the press has received massive scholarly attention, the same cannot be said of language and its use in historical accounts. Thus, from a historical research perspective, employing an analytical and interpretive study of secondary data sources, and underpinned by cultural theories such as linguistic imperialism and hegemony, this study analyses how language was featured in the press during the research period and the implications thereof. This analysis shows that various institutions and individuals exploited language to foster narrow socio-cultural, ideological, and political agendas in the Gold Coast, later Ghana. This study also revealed that language use and its ideological forces during the pre- and immediate post-independence periods contributed immensely to Ghanaian media’s current state and Ghanaians’ general language attitudes. This study’s significance lies in the realisation that language, as a cultural artifact, and its choice, use, and consumption could have far-reaching consequences for the self-realisation, actualisation, and general progress of a society. Thus, African societies should be mindful of the ideological implications of language choices not just in the press but also in other societal contexts. This study, therefore, invites further studies on language and its use in the media, especially from 1992, for a more comprehensive appreciation of the issues raised in this study.

MoTion: A new declarative object matching approach in Pharo

Pattern matching is an expressive way of matching data and extracting pieces of information from it. The recent inclusion of pattern matching in the Java and Python languages highlights that such a facility is more and more adopted by developers for everyday development. Other main stream programming languages also offer pattern matching capabilities as part of the language (Rust, Scala, Haskell, and OCaml), with different degrees of expressivity in what can be matched. In the meantime, in graphs, pattern matching takes a slightly different turn; it enhances the expressivity of the patterns that can be defined. Smalltalk currently offers little pattern matching capability to find specific objects inside a large graph of objects using a declarative pattern. In Pharo, the closest library to classical pattern matching that exists is the RBParseTreeSearcher, which allows to express specialized patterns over a Pharo Abstract Syntax Tree to find some inner node. The question arises of what features a flexible pattern matching language should have. In this paper, we review the features found in different existing pattern matching languages, both in General Purpose Languages (like Java) and in declarative graph pattern matching languages. We then describe MoTion, a new pattern matching engine for Pharo smalltalk, combining all these features. We discuss some aspects of MoTion’s implementation and illustrate its use with real case examples.

A multifaceted approach to detect gender biases in Natural Language Generation

Recent advances in generative models have skyrocketed the popularity of conversational chatbots and have revolutionized the way we interact with artificial intelligence. At the same time, research has shown that machine learning models can unconsciously reflect and amplify human biases. This is particularly dangerous for generative models given the huge popularity of such technologies. Specifically, a fundamental source of bias of such technologies is the resources on which the models are trained. To address this issue, this paper proposes a methodology to analyze intrinsic gender bias in Natural Language Generation (NLG). Some works already propose metrics and approaches to measure bias in the Natural Language processing field. However, there is a lack of standard methodology to measure gender bias in NLG. Therefore, adapting the Bias Score approach, our proposal involves three sequential stages applied to individual texts to detect intrinsic gender bias on NLG effectively. Those steps are as follows: (i) word scoring; (ii) word filtering; and (iii) generative-word analysis. This methodology is applied to recent datasets and pre-trained models widely used for the generation of text with common sense. In particular, this paper analyzes the potential gender bias in CommonGen and C2Gen datasets and the SimpleNLG and T5 models. The results show the ability of the proposed methodology to detect gender bias in word distributions, presenting a strong correlation with the words typically associated with a specific gender. Results indicate that both tested datasets are intrinsically gender-biased, and therefore, tested models fine-tuned with those datasets also are.

Combined location online weather data: easy-to-use targeted weather analysis for agriculture

The continuing effects of climate change require farmers and growers to have greater understanding of how these changes affect crop production. However, while climatic data is generally available to help provide much of that understanding, it can often be in a form not easy to digest. The proposed Combined Location Online Weather Data (CLOWD) framework is an easy-to-use online platform for analysing recent and historical weather data of any location within Australia at the click of a map. CLOWD requires no programming skills and operates in any HTML5 web browser on PC and mobile devices. It enables comparison between current and previous growing seasons over a range of environmental parameters, and can create a plain-English PDF report for offline use, using natural language generation (NLG). This paper details the platform, the design decisions taken and outlines how farmers and growers can use CLOWD to better understand current growing conditions. Prototypes of CLOWD are now online for PCs and smartphones.

AMGPT: A large language model for contextual querying in additive manufacturing

Generalized large language models (LLMs) such as GPT-4 may not provide specific answers to queries formulated by materials science researchers. These models may produce a high-level outline but lack the capacity to return detailed instructions on manufacturing and material properties of novel alloys. We introduce “AMGPT”, a specialized LLM text generator designed for metal AM queries. The goal of AMGPT is to assist researchers and users in navigating a curated corpus of literature. Instead of training from scratch, we employ a pre-trained Llama2-7B model from Hugging Face in a Retrieval-Augmented Generation (RAG) setup, utilizing it to dynamically incorporate information from ∼50 AM papers and textbooks in PDF format. Mathpix is used to convert these PDF documents into TeX format, facilitating their integration into the RAG pipeline managed by LlamaIndex. A query retrieval function has also been added, enabling the system to fetch relevant literature from Elsevier journals based on the context of the query. Expert evaluations of this project highlight that specific embeddings from the RAG setup accelerate response times and maintain coherence in the generated text.

DKA-RG: Disease-Knowledge-Enhanced Fine-Grained Image–Text Alignment for Automatic Radiology Report Generation

Automatic radiology report generation is a task that combines artificial intelligence and medical information processing, and it fully relies on computer vision and natural language processing techniques. Nowadays, automatic radiology report generation is still a very challenging task because it requires semantically adequate alignment of data from two modalities: radiology images and text. Existing approaches tend to focus on coarse-grained alignment at the global level and do not take into account the disease characteristics of radiology images at fine-grained semantics, which results in the generated reports potentially omitting key disease diagnostic descriptions. In this work, we propose a new approach, disease-knowledge-enhanced fine-grained image–text alignment for automatic radiology report generation (DKA-RG). The method combines global and disease-level alignment, thus facilitating the extraction of fine-grained disease features by the model. Our approach also introduces a knowledge graph to inject medical domain expertise into the model. Our proposed DKA-RG consists of two training steps: the image–report alignment stage and the image-to-report generation stage. In the alignment stage, we use global contrastive learning to align images and texts from a high level and also augment disease contrastive learning with medical knowledge to enhance the disease detection capability. In the report generation stage, the report text generated from the images is more accurate in describing the disease information thanks to sufficient alignment. Through extensive quantitative and qualitative experiments on two widely used datasets, we validate the effectiveness of our DKA-RG on the task of radiology report generation. Our DKA-RG achieves superior performance on multiple types of metrics (natural language generation and clinical efficacy metrics) compared to existing methods, demonstrating that the method can improve the reliability and accuracy of automatic radiology report generation systems.

Enhancing AI Responses in Chemistry: Integrating Text Generation, Image Creation, and Image Interpretation through Different Levels of Prompts

Enhancing AI Responses in Chemistry: Integrating Text Generation, Image Creation, and Image Interpretation through Different Levels of Prompts

Even laypeople use legalese

Whereas principles of communicative efficiency and legal doctrine dictate that laws be comprehensible to the common world, empirical evidence suggests legal documents are largely incomprehensible to lawyers and laypeople alike. Here, a corpus analysis (n = 59) million words) first replicated and extended prior work revealing laws to contain strikingly higher rates of complex syntactic structures relative to six baseline genres of English. Next, two preregistered text generation experiments (n = 286) tested two leading hypotheses regarding how these complex structures enter into legal documents in the first place. In line with the magic spell hypothesis, we found people tasked with writing official laws wrote in a more convoluted manner than when tasked with writing unofficial legal texts of equivalent conceptual complexity. Contrary to the copy-and-edit hypothesis, we did not find evidence that people editing a legal document wrote in a more convoluted manner than when writing the same document from scratch. From a cognitive perspective, these results suggest law to be a rare exception to the general tendency in human language toward communicative efficiency. In particular, these findings indicate law's complexity to be derived from its performativity, whereby low-frequency structures may be inserted to signal law's authoritative, world-state-altering nature, at the cost of increased processing demands on readers. From a law and policy perspective, these results suggest that the tension between the ubiquity and impenetrability of the law is not an inherent one, and that laws can be simplified without a loss or distortion of communicative content.

Large language models (LLMs): survey, technical frameworks, and future challenges

Artificial intelligence (AI) has significantly impacted various fields. Large language models (LLMs) like GPT-4, BARD, PaLM, Megatron-Turing NLG, Jurassic-1 Jumbo etc., have contributed to our understanding and application of AI in these domains, along with natural language processing (NLP) techniques. This work provides a comprehensive overview of LLMs in the context of language modeling, word embeddings, and deep learning. It examines the application of LLMs in diverse fields including text generation, vision-language models, personalized learning, biomedicine, and code generation. The paper offers a detailed introduction and background on LLMs, facilitating a clear understanding of their fundamental ideas and concepts. Key language modeling architectures are also discussed, alongside a survey of recent works employing LLM methods for various downstream tasks across different domains. Additionally, it assesses the limitations of current approaches and highlights the need for new methodologies and potential directions for significant advancements in this field.

Navigating the Ethical Landscape of ChatGPT Integration in Scientific Research: Review of Challenges and Recommendations

This study aims to investigate the transformative potential of incorporating ChatGPT into academic research while addressing consequential challenges and ethical concerns. The chatbot, as a large language model, allows users to interact with it, providing advantages in the academic research context. The benefits of real-time interaction, information synthesis, and language generation have the potential to modernize how researchers engage with and extract insights from vast amounts of data. Additionally, ChatGPT demonstrates a strong command of programming languages, which is highly relevant in modern research related to machine learning and artificial intelligence. However, it still has certain limitations, such as the potential for generating fake citations and issues related to plagiarism. Users need to be aware of these limitations. Therefore, researchers must be careful regarding the potential and limitations of ChatGPT, as it can improve efficiency but may decrease the quality of work if misused. This study recommends guidelines in academic research for using LLMs like ChatGPT. Overall, it’s not ChatGPT that downgrades the quality of academic work; rather, it’s the irresponsible use of ChatGPT that does it.