General Modeling Language Research Articles

УПРОВАДЖЕННЯ СИСТЕМ УПРАВЛІННЯ НАВЧАННЯМ З ФУНКЦІЯМИ ГЕНЕРАТИВНОГО ШТУЧНОГО ІНТЕЛЕКТУ В УМОВАХ ПОСТ-ПАНДЕМІЇ

In recent years, the global education system has undergone a transformative shift, relying increasingly on online tools to prove its resilience in the face of various challenges, including epidemiological events. This paper seeks to introduce a Learning Management System (LMS) designed to elevate the delivery of educational content, fostering the training and development of students' knowledge in a dynamic and adaptive learning environment anchored to the connectivity and synchronization needs of smart urban living. The system proposes the integration of a language generation model equipped with Application Programming Interface (API) access. This choice not only streamlines instructor support but also facilitates the creation of courses using an adapted and coherent language. The system goes beyond traditional LMS functionalities by providing suggestions to enhance and diversify lessons. Emphasizing flexibility and adaptability, the proposed LMS caters to various training levels, being able to accommodate the needs of learners across different domains. To identify the appropriate solution, a comparative analysis was conducted, evaluating various platforms based on functionality, intuitive presentation, and customization options. The selected system emerged as the most favorable, offering a robust framework for the development and delivery of educational content. To demonstrate the system's effectiveness, a curriculum was crafted for a specialized field of study - Artificial Intelligence (AI), with a specific focus on the practical application of Machine Learning algorithms. This curriculum incorporates theoretical and practical application components, complemented by a suite of assessment tools and assignments tailored to the proposed subject. The lessons within this curriculum were constructed by drawing insights from various bibliographical sources. What sets this system up-to-date is its integration of generative AI features directly into the LMS, enriching the teaching-learning-evaluating experience.

Generative AI Reshaping International Trade Pattern: How Do Foreign Trade Enterprises Seize Opportunities

The year 2023 is the first year of the outbreak of generative artificial intelligence applications, and digital transformation has also become the development trend of international trade today. Digital transformation, driven by the rapid development and adoption of generative AI technologies, has also become the dominant trend of international trade in the contemporary world. Generative AI, exemplified by ChatGPT, a powerful natural language generation model, has attracted global attention and sparked heated discussions about its implications and possibilities in different fields. Generative AI is a subset of artificial intelligence, which create new content from existing data by learning the underlying patterns and structures of the data. International trade, as an important part of the global economy, is inevitably influenced by the emergence and advancement of generative artificial intelligence. This paper aims to explore the current and potential impacts of generative artificial intelligence on international trade, focusing on three main aspects: creating new value from data, data analysis and preparation. The paper argues that generative artificial intelligence offers unprecedented opportunities for the international trade industry, as it help sellers in international trade to create more diverse and customized products and services, to optimize their production and marketing strategies, and to increase their competitiveness and profitability. The foreign trade enterprises should seize the opportunity and adopt proactive and innovative approaches to leverage the benefits of generative artificial intelligence, while also being aware of and prepared for the potential risks and challenges it poses to international trade.

Large language models for biomedicine: foundations, opportunities, challenges, and best practices.

Generative large language models (LLMs) are a subset of transformers-based neural network architecture models. LLMs have successfully leveraged a combination of an increased number of parameters, improvements in computational efficiency, and large pre-training datasets to perform a wide spectrum of natural language processing (NLP) tasks. Using a few examples (few-shot) or no examples (zero-shot) for prompt-tuning has enabled LLMs to achieve state-of-the-art performance in a broad range of NLP applications. This article by the American Medical Informatics Association (AMIA) NLP Working Group characterizes the opportunities, challenges, and best practices for our community to leverage and advance the integration of LLMs in downstream NLP applications effectively. This can be accomplished through a variety of approaches, including augmented prompting, instruction prompt tuning, and reinforcement learning from human feedback (RLHF). Our focus is on making LLMs accessible to the broader biomedical informatics community, including clinicians and researchers who may be unfamiliar with NLP. Additionally, NLP practitioners may gain insight from the described best practices. We focus on 3 broad categories of NLP tasks, namely natural language understanding, natural language inferencing, and natural language generation. We review the emerging trends in prompt tuning, instruction fine-tuning, and evaluation metrics used for LLMs while drawing attention to several issues that impact biomedical NLP applications, including falsehoods in generated text (confabulation/hallucinations), toxicity, and dataset contamination leading to overfitting. We also review potential approaches to address some of these current challenges in LLMs, such as chain of thought prompting, and the phenomena of emergent capabilities observed in LLMs that can be leveraged to address complex NLP challenge in biomedical applications.

Enhancing Voice Assistants with Generative AI Language Models for Interactive Conversations

This paper presents comprehensive research for the development of a voice assistant application using HTML/CSS, JavaScript, Python, SQLite, and the Hugging Face API. The proposed methodology encompasses key steps in data collection, model training, frontend and backend development, integration, testing, and deployment, ensuring the creation of a robust and user-friendly voice assistant system. By leveraging a combination of web technologies and Al tools, the developed voice assistant application offers an interactive and intelligent user experience across its users. Key Words: Voice assistant, HTML/CSS, JavaScript, Python, SQLite, Hugging Face API, LLMs, AI.

Generative large language models are all-purpose text analytics engines: text-to-text learning is all your need.

To solve major clinical natural language processing (NLP) tasks using a unified text-to-text learning architecture based on a generative large language model (LLM) via prompt tuning. We formulated 7 key clinical NLP tasks as text-to-text learning and solved them using one unified generative clinical LLM, GatorTronGPT, developed using GPT-3 architecture and trained with up to 20 billion parameters. We adopted soft prompts (ie, trainable vectors) with frozen LLM, where the LLM parameters were not updated (ie, frozen) and only the vectors of soft prompts were updated, known as prompt tuning. We added additional soft prompts as a prefix to the input layer, which were optimized during the prompt tuning. We evaluated the proposed method using 7 clinical NLP tasks and compared them with previous task-specific solutions based on Transformer models. The proposed approach achieved state-of-the-art performance for 5 out of 7 major clinical NLP tasks using one unified generative LLM. Our approach outperformed previous task-specific transformer models by ∼3% for concept extraction and 7% for relation extraction applied to social determinants of health, 3.4% for clinical concept normalization, 3.4%-10% for clinical abbreviation disambiguation, and 5.5%-9% for natural language inference. Our approach also outperformed a previously developed prompt-based machine reading comprehension (MRC) model, GatorTron-MRC, for clinical concept and relation extraction. The proposed approach can deliver the "one model for all" promise from training to deployment using a unified generative LLM.

The impact of ChatGPT on academia: A comprehensive analysis of AI policies across UT system academic institutions

Since ChatGPT was released by OpenAI, an American company, in 2022 for the public, ChatGPT has become the talk of every town, as evident by its over 180 million users worldwide. This chatbot's ability to engage in human-like conversations, answer questions, and generate diverse content has sparked widespread debates across various fields, including education. In response to the growing rise and influence of ChatGPT, educators have contrasting opinions; some view ChatGPT as an opportunity, whereas others regard it as a challenge that needs to be addressed on time. In order to deal with the complexities caused by ChatGPT in the field of education, universities have formulated their policies on AI. Guided by the research question, "How does universities' policy on AI reflect academia's view toward ChatGPT?" this study attempts to review the AI policy of the nine academic institutions under the UT system of the United States. The primary goal is to understand the extent to which universities have adapted their policies in response to the challenges and opportunities posed by ChatGPT and how these policies reflect the broader sentiments within academia. To achieve this, this study reviews the universities' policies regarding AI using a qualitative data analysis methodology. The primary data sources include official policies, statements, and guidelines developed by the universities in response to the challenges and opportunities presented by ChatGPT. While reviewing the policies, the study determines whether ChatGPT is banned and why. Or embraced, and if so, in what ways? By examining these policies, the study aims to uncover the various approaches universities have taken to integrate or regulate the use of ChatGPT within academic environments. The thesis of this study is twofold. First, it seeks to provide a comprehensive overview of how US universities have responded to ChatGPT in the educational landscape. This involves identifying common themes, concerns, and strategies institutions employ to deal with the complexities introduced by this generative language model. Second, the study aims to contribute to existing scholarship by offering insights into how academia adapts to the influence of AI technologies like ChatGPT. This study examines the intersection of AI and education and the evolving nature of educational norms in the digital age by uncovering the diverse perspectives and approaches within university policies.

Evidence-based potential of generative artificial intelligence large language models in orthodontics: a comparative study of ChatGPT, Google Bard, and Microsoft Bing.

The increasing utilization of large language models (LLMs) in Generative Artificial Intelligence across various medical and dental fields, and specifically orthodontics, raises questions about their accuracy. This study aimed to assess and compare the answers offered by four LLMs: Google's Bard, OpenAI's ChatGPT-3.5, and ChatGPT-4, and Microsoft's Bing, in response to clinically relevant questions within the field of orthodontics. Ten open-type clinical orthodontics-related questions were posed to the LLMs. The responses provided by the LLMs were assessed on a scale ranging from 0 (minimum) to 10 (maximum) points, benchmarked against robust scientific evidence, including consensus statements and systematic reviews, using a predefined rubric. After a 4-week interval from the initial evaluation, the answers were reevaluated to gauge intra-evaluator reliability. Statistical comparisons were conducted on the scores using Friedman's and Wilcoxon's tests to identify the model providing the answers with the most comprehensiveness, scientific accuracy, clarity, and relevance. Overall, no statistically significant differences between the scores given by the two evaluators, on both scoring occasions, were detected, so an average score for every LLM was computed. The LLM answers scoring the highest, were those of Microsoft Bing Chat (average score = 7.1), followed by ChatGPT 4 (average score = 4.7), Google Bard (average score = 4.6), and finally ChatGPT 3.5 (average score 3.8). While Microsoft Bing Chat statistically outperformed ChatGPT-3.5 (P-value = 0.017) and Google Bard (P-value = 0.029), as well, and Chat GPT-4 outperformed Chat GPT-3.5 (P-value = 0.011), all models occasionally produced answers with a lack of comprehensiveness, scientific accuracy, clarity, and relevance. The questions asked were indicative and did not cover the entire field of orthodontics. Language models (LLMs) show great potential in supporting evidence-based orthodontics. However, their current limitations pose a potential risk of making incorrect healthcare decisions if utilized without careful consideration. Consequently, these tools cannot serve as a substitute for the orthodontist's essential critical thinking and comprehensive subject knowledge. For effective integration into practice, further research, clinical validation, and enhancements to the models are essential. Clinicians must be mindful of the limitations of LLMs, as their imprudent utilization could have adverse effects on patient care.

Towards the Use of Social Robot Furhat and Generative AI in Testing Cognitive Abilities

Abstract Spoken communication between social robotic devices, powered by generative AI tools such as ChatGPT, and the senior population offers great potential for researching social interaction and robot identity perceptions as well as exploring the potential opportunities and challenges when implementing this human-machine interactions in real life situations and health care. In this paper we explore people’s perceptions of the social robot Furhat when administering verbal tasks similar to those used in screening for Alzheimer’s disease. We describe the Slovak system mounted on the robot that includes a speech recognizer, the scenarios powered by generative large language model ChatGPT, and a speech synthesizer. We tested the functionality of the proposed approach with two groups of participants: attendees of a large science fair and a scientific conference. The observations from 87 collected questionnaires suggest good potential and applicability of such an approach for the given task and more positive attitudes of older subjects compared to younger ones.

CySecBERT : A Domain-Adapted Language Model for the Cybersecurity Domain

The field of cysec is evolving fast. Security professionals are in need of intelligence on past, current and —ideally — upcoming threats, because attacks are becoming more advanced and are increasingly targeting larger and more complex systems. Since the processing and analysis of such large amounts of information cannot be addressed manually, cysec experts rely on machine learning techniques. In the textual domain, pre-trained language models such as Bidirectional Encoder Representations from Transformers (BERT) have proven to be helpful as they provide a good baseline for further fine-tuning. However, due to the domain-knowledge and the many technical terms in cysec, general language models might miss the gist of textual information. For this reason, we create a high-quality dataset 1 and present a language model 2 specifically tailored to the cysec domain that can serve as a basic building block for cybersecurity systems. The model is compared on 15 tasks: Domain-dependent extrinsic tasks for measuring the performance on specific problems, intrinsic tasks for measuring the performance of the internal representations of the model, as well as general tasks from the SuperGLUE benchmark. The results of the intrinsic tasks show that our model improves the internal representation space of domain words compared with the other models. The extrinsic, domain-dependent tasks, consisting of sequence tagging and classification, show that the model performs best in cybersecurity scenarios. In addition, we pay special attention to the choice of hyperparameters against catastrophic forgetting, as pre-trained models tend to forget the original knowledge during further training.

Transformative Technologies in the Evaluation of a Vocational Education System

The increasing demand for vocational education has necessitated the presence of highly skilled teachers. This study presents a novel framework for the effective management of vocational college instructors’ professional development through the utilization of advanced technologies. The system utilizes deep learning technology to analyze many data points, including academic achievements, teaching experience, student comments, and professional activities, in order to assess the performance and potential of teachers. The system evaluates both the positive and negative aspects, offers customized training programs, and enhances the delivery of instruction through the utilization of a generative language model. The effectiveness of the system is supported by a case study, which demonstrates enhancements in talent management, professional development, teaching quality, and student happiness. This proposed solution aims to improve vocational education by empowering educators and transforming the processes of evaluation, support, and guidance throughout their professional trajectories.

A systematic evaluation of text mining methods for short texts: Mapping individuals’ internal states from online posts

Short texts generated by individuals in online environments can provide social and behavioral scientists with rich insights into these individuals’ internal states. Trained manual coders can reliably interpret expressions of such internal states in text. However, manual coding imposes restrictions on the number of texts that can be analyzed, limiting our ability to extract insights from large-scale textual data. We evaluate the performance of several automatic text analysis methods in approximating trained human coders’ evaluations across four coding tasks encompassing expressions of motives, norms, emotions, and stances. Our findings suggest that commonly used dictionaries, although performing well in identifying infrequent categories, generate false positives too frequently compared to other methods. We show that large language models trained on manually coded data yield the highest performance across all case studies. However, there are also instances where simpler methods show almost equal performance. Additionally, we evaluate the effectiveness of cutting-edge generative language models like GPT-4 in coding texts for internal states with the help of short instructions (so-called zero-shot classification). While promising, these models fall short of the performance of models trained on manually analyzed data. We discuss the strengths and weaknesses of various models and explore the trade-offs between model complexity and performance in different applications. Our work informs social and behavioral scientists of the challenges associated with text mining of large textual datasets, while providing best-practice recommendations.

The influence of informal structures on corporate strategy: An African perspective

This paper provides a detailed exploration of the Afrocentric epistemological perspective of socio-cultural constructs that drive the influence of informal structures on corporate strategy and employees. The research focuses on the African context, which has a dearth of studies despite ample research exploring informal structures in Western and Asian socio-ecological contexts. To achieve this, the study employed a systematic literature review through the aid of NVivo 14 and ChatGPT, a language generation model. NVivo 14 provided the study with a tool to manage and analyze the data, and ChatGPT offered a dialogical string that created a clear path for critical analysis. The paper adopted a systematic literature review approach to investigate the origin of the concept of informal structures and their impact on corporate strategy in an Afrocentric context. The findings indicate that informal structures can have a positive influence by facilitating the flow of information, flexibility, and fulfilling social needs and shared values. However, the negative influence of informal structures, including resistance to corporate strategy, social fragmentation, conflict, and political discord, outweighs the positive influence. The study also reveals that there is limited research on the social-ecological context of Africa and its impact on corporate strategy. The paper addresses a contextual gap in the body of knowledge through its contribution that focuses specifically on the Afrocentric context. The research also sheds light on the need for more studies on the social-ecological context of Africa to gain a better understanding of its impact on corporate strategy. Overall, the paper provides a comprehensive and detailed exploration of the Afrocentric perspective of informal structures and their impact on corporate strategy and employees.

Leveraging Generative Pre-trained Language Models for Advanced Unsupervised Neural Machine Translation

Leveraging Generative Pre-trained Language Models for Advanced Unsupervised Neural Machine Translation

Future applications of generative large language models: A data-driven case study on ChatGPT

This study delves into the evolving role of generative Large Language Models (LLMs). We develop a data-driven approach to collect and analyse tasks that users are asking to generative LLMs. Thanks to the focus on tasks this paper contributes to give a quantitative and granular understanding of the potential influence of LLMs in different business areas. Utilizing a dataset comprising over 3.8 million tweets, we identify and cluster 31,747 unique tasks, with a specific case study on ChatGPT. To reach this goal, the proposed method combines two Natural Language Processing (NLP) Techniques, Named Entity Recognition (NER) and BERTopic. The combination makes it possible to collect granular tasks of LLMs (NER) and clusters them in business areas (BERTopic). Our findings reveal a wide spectrum of applications, from programming assistance to creative content generation, highlighting LLM's versatility. The analysis highlighted six emerging areas of application for ChatGPT: human resources, programming, social media, office automation, search engines, education. The study also examines the implications of these findings for innovation management, proposing a research agenda to explore the intersection of the identified areas, with four stages of the innovation process: idea generation, screening/idea selection, development, and diffusion/sales/marketing.

Generative Language Patterns and the Phenomenon of Anti-Anthropocentrism — New Perspectives on the Linguistic Paradigm of «Posthumano» and «General/Strong» AI

The presented material introduces an authorial hypothesis on the prospects of the emergence of the linguistic paradigm «posthumano», based on an analytical interdisciplinary review and analysis of the development feature of large generative language models through the enhancement of neural network algorithms in deep machine learning. The authors put forward a theoretical and methodological foundation of the appearance of such a phenomenon in the context of the development of anti-anthropocentrism theories. The article actualizes the development of «post-anthropocentric» approaches overcoming the traditional binary opposition «man/machine», extending into the realm of linguistics. Building on achievements in the development and training of generative language models, a new concept of language as an autonomous, self-developing system, termed «posthumano» is formulated and proposed for further criticism and development. The authorial concept of the emerging a new linguistic paradigm is based both on the principles of the interdisciplinary approach and on the discourse of transdisciplinarity that developed in the second half of the 20th century. In this contribution, the authors outline some of the foundational principles and reasons that may underlie the new linguistic paradigm and offer some thoughts on the immediate prospects for the development of linguistic practices and humanitarian knowledge in the era of expectation/development of «general/strong» artificial intelligence technologies. The authors put forward a theoretical and methodological basis for the emergence of such a phenomenon in the context of the development of theories of anti-anthropocentrism. The article updates the development of «post-anthropocentric» approaches that go beyond the traditional binary opposition «man/machine», including the field of linguistics. Based on advances in the development and training of generative language models, the formulation and a new concept of language as an autonomous, self-developing «posthumano» system is proposed for further analysis, criticism, and development. The author’s concept of the formation and development of a new emerging linguistic paradigm is based both on the principles of an interdisciplinary approach and on the discourse of transdisciplinarity, which developed in the second half of the 20th century. The authors set out both some basic foundations and reasons that can form the basis of a new linguistic paradigm and offer some thoughts on the immediate prospects for the development of language practices and humanitarian knowledge in the era of anticipation/development of «general/strong» artificial intelligence technologies.

Does Artificial Intelligence Speak Our Language?: A Gadamerian Assessment of Generative Language Models

The language argument is a classic argument for human distinctiveness that, for millenia, has been used to distinguish humans from non-human animals. Generative language models (GLMs) pose a challenge to traditional language-based models of human distinctiveness precisely because they can communicate and respond in a manner resembling humanity’s linguistic capabilities. This article asks: have GLMs acquired natural language? Employing Gadamer’s theory of language, I argue that they have not. While GLMs can reliably generate linguistic content that can be interpreted as “texts,” they lack the linguistically mediated reality that language provides. Missing from these models are four key features of a linguistic construction of reality: groundedness to the world, understanding, community, and tradition. I conclude with skepticism that GLMs can ever achieve natural language because they lack these characteristics in their linguistic development.

Small Language Model Can Self-Correct

Generative Language Models (LMs) such as ChatGPT have exhibited remarkable performance across various downstream tasks. Nevertheless, one of their most prominent drawbacks is generating inaccurate or false information with a confident tone. Previous studies have devised sophisticated pipelines and prompts to induce large LMs to exhibit the capability for self-correction. However, large LMs are explicitly prompted to verify and modify their answers separately rather than completing all steps spontaneously like humans. Moreover, these complex prompts are extremely challenging for small LMs to follow. In this paper, we introduce the Intrinsic Self-Correction (ISC) in generative language models, aiming to correct the initial output of LMs in a self-triggered manner, even for those small LMs with 6 billion parameters. Specifically, we devise a pipeline for constructing self-correction data and propose Partial Answer Masking (PAM), aiming to endow the model with the capability for intrinsic self-correction through fine-tuning. We conduct experiments using LMs with parameters sizes ranging from 6 billion to 13 billion in two tasks, including commonsense reasoning and factual knowledge reasoning. Our experiments demonstrate that the outputs generated using ISC outperform those generated without self-correction. We believe that the output quality of even small LMs can be further improved by empowering them with the ability to intrinsic self-correct.

A General Search-Based Framework for Generating Textual Counterfactual Explanations

One of the prominent methods for explaining the decision of a machine-learning classifier is by a counterfactual example. Most current algorithms for generating such examples in the textual domain are based on generative language models. Generative models, however, are trained to minimize a specific loss function in order to fulfill certain requirements for the generated texts. Any change in the requirements may necessitate costly retraining, thus potentially limiting their applicability. In this paper, we present a general search-based framework for generating counterfactual explanations in the textual domain. Our framework is model-agnostic, domain-agnostic, anytime, and does not require retraining in order to adapt to changes in the user requirements. We model the task as a search problem in a space where the initial state is the classified text, and the goal state is a text in a given target class. Our framework includes domain-independent modification operators, but can also exploit domain-specific knowledge through specialized operators. The search algorithm attempts to find a text from the target class with minimal user-specified distance from the original classified object.

The Generalization and Robustness of Transformer-Based Language Models on Commonsense Reasoning

The advent of powerful transformer-based discriminative language models and, more recently, generative GPT-family models, has led to notable advancements in natural language processing (NLP), particularly in commonsense reasoning tasks. One such task is commonsense reasoning, where performance is usually evaluated through multiple-choice question-answering benchmarks. Till date, many such benchmarks have been proposed and `leaderboards' tracking state-of-the-art performance on those benchmarks suggest that transformer-based models are approaching human-like performance. However, due to documented problems such as hallucination and bias, the research focus is shifting from merely quantifying accuracy on the task to an in-depth, context-sensitive probing of LLMs' generalization and robustness. To gain deeper insight into diagnosing these models' performance in commonsense reasoning scenarios, this thesis addresses three main studies: the generalization ability of transformer-based language models on commonsense reasoning, the trend in confidence distribution of these language models confronted with ambiguous inference tasks, and a proposed risk-centric evaluation framework for both discriminative and generative language models.

SocialStigmaQA: A Benchmark to Uncover Stigma Amplification in Generative Language Models

Current datasets for unwanted social bias auditing are limited to studying protected demographic features such as race and gender. In this work, we introduce a comprehensive benchmark that is meant to capture the amplification of social bias, via stigmas, in generative language models. Taking inspiration from social science research, we start with a documented list of 93 US-centric stigmas and curate a question-answering (QA) dataset which involves simple social situations. Our benchmark, SocialStigmaQA, contains roughly 10K prompts, with a variety of prompt styles, carefully constructed to systematically test for both social bias and model robustness. We present results for SocialStigmaQA with two open source generative language models and we find that the proportion of socially biased output ranges from 45% to 59% across a variety of decoding strategies and prompting styles. We demonstrate that the deliberate design of the templates in our benchmark (e.g., adding biasing text to the prompt or using different verbs that change the answer that indicates bias) impacts the model tendencies to generate socially biased output. Additionally, through manual evaluation, we discover problematic patterns in the generated chain-of-thought output that range from subtle bias to lack of reasoning. Warning: This paper contains examples of text which are toxic, biased, and potentially harmful.