The Role of Interactive Visualization in Fostering Trust in AI

Visualization for Trust in Machine Learning Revisited: The State of the Field in 2023

Visualization for explainable and trustworthy machine learning remains one of the most important and heavily researched fields within information visualization and visual analytics with various application domains, such as medicine, finance, and bioinformatics. After our 2020 state-of-the-art report comprising 200 techniques, we have persistently collected peer-reviewed articles describing visualization techniques, categorized them based on the previously established categorization schema consisting of 119 categories, and provided the resulting collection of 542 techniques in an online survey browser. In this survey article, we present the updated findings of new analyses of this dataset as of fall 2023 and discuss trends, insights, and eight open challenges for using visualizations in machine learning. Our results corroborate the rapidly growing trend of visualization techniques for increasing trust in machine learning models in the past three years, with visualization found to help improve popular model explainability methods and check new deep learning architectures, for instance.

Determining a Trustworthy Application for Medical Data Visualizations through a Knowledge-Based Fuzzy Expert System

Medical data, such as electronic health records, are a repository for a patient's medical records for use in the diagnosis of different diseases. Using medical data for individual patient care raises a number of concerns, including trustworthiness in data management, privacy, and patient data security. The introduction of visual analytics, a computing system that integrates analytics approaches with interactive visualizations, can potentially deal with information overload concerns in medical data. The practice of assessing the trustworthiness of visual analytics tools or applications using factors that affect medical data analysis is known as trustworthiness evaluation for medical data. It has a variety of major issues, such as a lack of important evaluation of medical data, the need to process much of medical data for diagnosis, the need to make trustworthy relationships clear, and the expectation that it will be automated. Decision-making strategies have been utilized in this evaluation process to avoid these concerns and intelligently and automatically analyze the trustworthiness of the visual analytics tool. The literature study found no hybrid decision support system for visual analytics tool trustworthiness in medical data diagnosis. Thus, this research develops a hybrid decision support system to assess and improve the trustworthiness of medical data for visual analytics tools using fuzzy decision systems. This study examined the trustworthiness of decision systems using visual analytics tools for medical data for the diagnosis of diseases. The hybrid multi-criteria decision-making-based decision support model, based on the analytic hierarchy process and sorting preferences by similarity to ideal solutions in a fuzzy environment, was employed in this study. The results were compared to highly correlated accuracy tests. In conclusion, we highlight the benefits of our proposed study, which includes performing a comparison analysis on the recommended models and some existing models in order to demonstrate the applicability of an optimal decision in real-world environments. In addition, we present a graphical interpretation of the proposed endeavor in order to demonstrate the coherence and effectiveness of our methodology. This research will also help medical experts select, evaluate, and rank the best visual analytics tools for medical data.