Cited by Leveraging Artificial Intelligence for personalised insomnia-sleep calibration via the Big Five Personality Traits

Measuring the Interpretability and Explainability of Model Decisions of Five Large Language Models DOI

Kaito Fujiwara,

Miyu Sasaki,

Akira Nakamura

et al.

Published: March 20, 2024

This study conducts a comprehensive analysis of the interpretability and explainability five leading Large Language Models (LLMs): TripoSR by Stability AI, Gemma-7b Google, Mistral 7B Llama-2-7b Meta, GemMoE-Beta-1 CrystalCare AI. Through methodical evaluation encompassing both qualitative quantitative benchmarks, we assess these models' capacity to make their decision-making processes understandable humans. Our findings reveal significant variability in ability provide transparent reasoning accurate, contextually relevant explanations across different contexts. Notably, demonstrated superior transparency, while excelled accuracy explanations. However, challenges maintaining consistent varying inputs need for enhanced adaptability feedback highlight areas future improvement. research underscores importance fostering trust reliability LLM applications, advocating continued advancement achieve more transparent, accountable, user-centric AI systems. Directions include development standardized methodologies interdisciplinary approaches enhance model transparency user understanding.

Language: Английский

Citations

Perceived support and AI literacy: the mediating role of psychological needs satisfaction DOI

Yanyan Shen,

Wencheng Cui

Frontiers in Psychology, Journal Year: 2024, Volume and Issue: 15

Published: June 14, 2024

Artificial Intelligence (AI) exerts significant influence on both professional and personal spheres, underscoring the necessity for college students to have a fundamental understanding of AI. Guided by self-determination theory (SDT), this study explores psychological needs satisfaction AI literacy among university students. A cross-sectional survey involving 445 from diverse academic backgrounds was conducted. The assessed mediation effect students’ need between two types support—technical teacher—and literacy. results indicate that support positively influenced fulfillment autonomy competence needs, which subsequently acted as mediators in enhancing However, relatedness did not mediate relationship Unexpectedly, no direct association found forms levels findings suggest although technical teacher contribute fulfilling specific only are predictive lack impact underscores importance addressing through educational interventions. It is recommended educators provide tailored education (AIEd) institutions develop specialized courses enhance

Language: Английский

Citations

VIRTSI: A novel trust dynamics model enhancing Artificial Intelligence collaboration with human users – Insights from a ChatGPT evaluation study DOI

Maria Virvou, George A. Tsihrintzis,

Evangelia-Aikaterini Tsichrintzi

et al.

Information Sciences, Journal Year: 2024, Volume and Issue: 675, P. 120759 - 120759

Published: May 21, 2024

The rapid integration of intelligent processes and methods into information systems in the Artificial Intelligence (AI) era has led to a substantial shift towards autonomous software decision-making. This evolution necessitates robust human oversight, especially critical domains like Healthcare, Education, Energy. Human trust AI plays vital role influencing decision-making users interacting with AI. paper presents VIRTSI (Variability Impact Reciprocal Trust States Intelligent systems), novel rigorous computational model for human-AI Interaction. simulates states, spanning from overtrust distrust, through user modelling. It comprises: 1. A dynamics representational based on Deterministic Finite State Automata (DFAs), illustrating transitions among cognitive states response AI-generated replies. 2. evaluation Confusion Matrices, originating machine learning Accuracy Metrics, providing quantitative framework analysing dynamics. As result, this is first time that have been thoroughly traced method developed assess impact possibly harmful distrust. An empirical study recently launched Large Language Model generative AI, ChatGPT (version 3.5), provides radical underexplored platform evaluating interaction VIRTSI. involved 1200 interactions real as well experts together two very different evaluation, namely engineering poetry. traces emerging interaction, concrete examples synergies research reveals maintaining normal optimal both need further steps goal. real-world implications can guide creation interfaces incorporation functionalities development chatbots terms by new DFA corresponding perspective confusion matrix dynamics' efficiency dialogues.

Language: Английский

Citations

Visualization obesity risk prediction system based on machine learning DOI

Jinsong Du, Sijia Yang,

Yijun Zeng

et al.

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Sept. 28, 2024

Language: Английский

Citations

Introduction to Advances in Artificial Intelligence-Empowered Decision Support Systems DOI

George A. Tsihrintzis, Maria Virvou, Haris Doukas

et al.

Learning and analytics in intelligent systems, Journal Year: 2024, Volume and Issue: unknown, P. 1 - 11

Published: Jan. 1, 2024

Language: Английский

Citations

EPATHLO Educational Game Suite and Fuzzy Reasoning DOI