Transformative Conversational AI: Sentiment Recognition in Chatbots via Transformers DOI

Sadam Hussain Noorani,

Sheharyar Khan, Awais Mahmood

et al.

Published: Nov. 17, 2023

Recent years have seen a dramatic increase in the use of conversational artificial intelligence (CAI) for both academic and commercial applications, primarily context chatbots AI virtual assistants. The user's engagement produces human like responses. However, capacity to discern sentiments respond adequately is one major difficulties faced by conversation systems. In present study, we propose transformer-based framework sentiment-aware chatbot. suggested transformer neural network architecture that highly parallelizable solely dependent on self-attention mechanism. A model controls variable-sized input using stacks layers rather than deep networks or CNNs. this manner, language creation carried out cutting-edge pre-trained CTRL, which can easily adapt various models without needing architectural adaptations. Our was trained DailyDialogues dataset evaluated automated metrics. Findings from experiments confirm that, terms content quality emotion perception, our technique works better baselines.

Language: Английский

Chatbots Put to the Test in Math and Logic Problems: A Comparison and Assessment of ChatGPT-3.5, ChatGPT-4, and Google Bard DOI Creative Commons
Vagelis Plevris, George Papazafeiropoulos, Alejandro Jiménez Ríos

et al.

AI, Journal Year: 2023, Volume and Issue: 4(4), P. 949 - 969

Published: Oct. 24, 2023

In an age where artificial intelligence is reshaping the landscape of education and problem solving, our study unveils secrets behind three digital wizards, ChatGPT-3.5, ChatGPT-4, Google Bard, as they engage in a thrilling showdown mathematical logical prowess. We assess ability chatbots to understand given problem, employ appropriate algorithms or methods solve it, generate coherent responses with correct answers. conducted using set 30 questions. These questions were carefully crafted be clear, unambiguous, fully described plain text only. Each question has unique well-defined answer. The divided into two sets 15: Set A consists “Original” problems that cannot found online, while B includes “Published” are readily available often their solutions. was presented each chatbot times May 2023. recorded analyzed responses, highlighting strengths weaknesses. Our findings indicate can provide accurate solutions for straightforward arithmetic, algebraic expressions, basic logic puzzles, although may not consistently every attempt. However, more complex advanced tasks, chatbots’ answers, appear convincing, reliable. Furthermore, consistency concern conflicting answers when same multiple times. To evaluate compare performance chatbots, we quantitative analysis by scoring final based on correctness. results show ChatGPT-4 performs better than ChatGPT-3.5 both Bard ranks third original A, trailing other chatbots. achieves best performance, taking first place published B. This likely due Bard’s direct access internet, unlike ChatGPT which, designs, do have external communication capabilities.

Language: Английский

Citations

43

AI-Powered Mental Health Virtual Assistants Acceptance: An Empirical Study on Influencing Factors Among Generations X, Y, and Z DOI Open Access
Turki Alanzi,

Abdullah A Alsalem,

Hessah Alzahrani

et al.

Cureus, Journal Year: 2023, Volume and Issue: unknown

Published: Nov. 27, 2023

Study purpose: This study aims to analyze various influencing factors among generations X (Gen X), Y Y), and Z Z) of artificial intelligence (AI)-powered mental health virtual assistants. Methods: A cross-sectional survey design was adopted in this study. The sample consisted outpatients diagnosed with illnesses, such as anxiety, depression, schizophrenia, behavioral disorders. questionnaire designed based on the (performance expectancy, effort social influence, facilitating conditions, behavioural intention) identified from unified theory acceptance use technology model. Ethical approval received Ethics Committee at Imam Abdulrahman Bin Faisal University, Saudi Arabia. Results: total 506 patients participated study, over 80% having moderate high experience using AI ANOVA results for performance expectancy (PE), (EE), influence (SI), conditions (FC), intentions (BI) indicate that there are statistically significant differences (p < 0.05) between Gen X, Y, participants. Conclusion: findings underscore significance considering generational attitudes perceptions, demonstrating more positive stronger assistants, while appears be cautious.

Language: Английский

Citations

9

Overviewing Biases in Generative AI-Powered Models in the Arabic Language DOI
Mussa Saidi Abubakari

Advances in computational intelligence and robotics book series, Journal Year: 2025, Volume and Issue: unknown, P. 361 - 390

Published: Feb. 28, 2025

Natural Language Processing (NLP) is an emerging field often integrated into Artificial Intelligence (AI) technologies. NLP has significantly advanced, leading to the widespread use of generative AI-powered (Gen-AI) models across various domains. However, while Gen-AI systems have been successfully implemented in several languages, AI-based language still face considerable challenges and shortcomings, including generating biases sensitive languages like Arabic. Therefore, primary objective this chapter provide overview Gen-AI-powered context Arabic language, exploring sources these biases, their implications, potential strategies for mitigation. The underscore need ongoing research development create more equitable accurate AI systems. By understanding origins implications implementing effective mitigation strategies, we can work towards that better serve diverse linguistic communities.

Language: Английский

Citations

0

Understanding question-answering systems: Evolution, applications, trends, and challenges DOI Creative Commons

Amer Farea,

Frank Emmert‐Streib

Engineering Applications of Artificial Intelligence, Journal Year: 2025, Volume and Issue: 156, P. 110997 - 110997

Published: May 22, 2025

Language: Английский

Citations

0

An inquiry smart chatbot system for Al-Zaytoonah University of Jordan DOI Open Access
Nagham A. Al-Madi, Khulood Abu Maria, Mohammad Al-Madi

et al.

Bulletin of Electrical Engineering and Informatics, Journal Year: 2024, Volume and Issue: 13(4), P. 2758 - 2773

Published: June 1, 2024

Chatbots are important in artificial intelligence (AI) and natural language processing (NLP). The development of the chatbot is viewed as a continuous issue field. This suitable for Arabic chatbots that not widely available. study aims to fill gap by creating an system university admissions. uses deep neural network model manually constructed dataset conversation pairings, utilizing Jordanian dialect from Al-Zaytoonah University Jordan’s (ZUJ) website. efficiently answers most user queries, improving counseling experience reducing workload admissions department. adoption this also minimizes website traffic congestion. contributes improvement technology learning-based optimized admissions, demonstrating its potential impact Arabic-speaking context. Future research can further enhance system’s capabilities applicability other disciplines.

Language: Английский

Citations

1

The Evolution of Transformers in Education: A Literature Review DOI

Chaimaa Bouafoud,

Khalid Zine-Dine,

Abdellah Madani

et al.

Published: June 28, 2024

Language: Английский

Citations

1

Deep Learning in Written Arabic Linguistic Studies: A Comprehensive Survey DOI Creative Commons
Manar Almanea

IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 172196 - 172233

Published: Jan. 1, 2024

Language: Английский

Citations

1

PixieGPT: Design and Implementation of a Generative Pre-Trained Transformer for Universities of Bangladesh DOI Open Access
Hasan Islam, Mehedi Hasan,

Sumiaya Ahmed

et al.

Published: Feb. 20, 2024

In a densely populated country like Bangladesh, universities grapple with the challenge of efficiently addressing myriad queries from large student body, leading to heightened workload for university stakeholders. To tackle these challenges, we introduce PixieGPT, tailor-made Generative Pre-Trained Transformer Bangladeshi universities. PixieGPT significantly mitigates by adeptly handling common university-related queries, thereby enhancing user experience. The hierarchical structure plays crucial role in managing diverse thousands students about system. solution introduces modular knowledge base (KB) simpler complexities, intricacies volumes queries. is designed way so that also adaptable implementation other worldwide based on requirements particular administrative nature facilitates easy adaptation minor changes specific requirements, ensuring seamless integration process. This paper delves into PixieGPT&#039;s design, emphasizing its pivotal mitigating challenges stakeholders Bangladesh. incorporation BERT Natural Language Understanding(NLU) and GPT models Generation(NLG) enhances capabilities, contributing scalability efficiency presented use case underscores practical benefits positioning it as promising globally similar operational frameworks.

Language: Английский

Citations

0

Pixiegpt: Design and Implementation of a Generative Pre-Trained Transformer for Universities of Bangladesh DOI
Hasan Islam, Mehedi Hasan,

Sumiaya Ahmed

et al.

Published: Jan. 1, 2024

Language: Английский

Citations

0

Transformative Conversational AI: Sentiment Recognition in Chatbots via Transformers DOI

Sadam Hussain Noorani,

Sheharyar Khan, Awais Mahmood

et al.

Published: Nov. 17, 2023

Recent years have seen a dramatic increase in the use of conversational artificial intelligence (CAI) for both academic and commercial applications, primarily context chatbots AI virtual assistants. The user's engagement produces human like responses. However, capacity to discern sentiments respond adequately is one major difficulties faced by conversation systems. In present study, we propose transformer-based framework sentiment-aware chatbot. suggested transformer neural network architecture that highly parallelizable solely dependent on self-attention mechanism. A model controls variable-sized input using stacks layers rather than deep networks or CNNs. this manner, language creation carried out cutting-edge pre-trained CTRL, which can easily adapt various models without needing architectural adaptations. Our was trained DailyDialogues dataset evaluated automated metrics. Findings from experiments confirm that, terms content quality emotion perception, our technique works better baselines.

Language: Английский

Citations

0