Agents that Argue and Explain Classifications of Retinal Conditions

Abstract

Expertise for auditing AI systems in medical domain is only now being accumulated. Conformity assessment procedures will require AI systems: 1) to be transparent, 2) not to rely decisions solely on algorithms, or iii) to include safety assurance cases in the documentation to facilitate technical audit. We are interested here in obtaining transparency in the case of machine learning (ML) applied to classification of retina conditions. High performance metrics achieved using ML has become common practice. However, in the medical domain, algorithmic decisions need to be sustained by explanations. We aim at building a support tool for ophthalmologists able to: i) explain algorithmic decision to the human agent by automatically extracting rules from the ML learned models; ii) include the ophthalmologist in the loop by formalising expert rules and including the expert knowledge in the argumentation machinery; iii) build safety cases by creating assurance argument patterns for each diagnosis. Methods: For the learning task, we used a dataset consisting of 699 OCT images: 126 Normal class, 210 with Diabetic Retinopathy (DR) and 363 with Age Related Macular Degeneration (AMD). The dataset contains patients from the Ophthalmology Department of the County Emergency Hospital of Cluj-Napoca. All ethical norms and procedures, including anonymisation, have been performed. We applied three machine learning algorithms: decision tree A.

Inteligencia Artificial y Auditoría: Tendencias de la literatura científica

“…Se observa una relación una relación significativa entre las palabras claves con mayor coocurrencia como: "Expert Systems-Machine Learning", "Learning system -Finance", "Decision Making -Audit". En el caso de "Expert Systems" y "Machine Learning", se identificó el estudio en el que se describe el desarrollo y evaluación de sistemas de agentes de IA que utilizan diferentes técnicas para clasificar y explicar diferentes tipos de enfermedades oculares (Groza et al, 2021). Con respecto a "Learning System-Finance", uno de los estudios identificados aborda el análisis de contenido de video empleado en la entidades financieras y de seguridad para contribuir en la eficacia de la auditoría de cumplimiento y la detección de actividades fraudulentas e inconsistentes (Fan et al, 2020).…”

Information-seeking dialogue for explainable artificial intelligence: Modelling and analytics

“…Calegari et al [10] implement a mechanism of reasoning over defeasible preferences using elements of abstract and structured argumentation. Groza et al [26] model explanatory dialogues combining rulebased arguments extracted from both ML classifiers and expert knowledge in favour or against a given classification of retinal disorder. Subsequently, the arguments are used to persuade the other parties in multi-agent system settings.…”

Interleaving Automatic Segmentation and Expert Opinion for Retinal Conditions

“…Optical coherence tomography (OCT) has become the leading diagnostic tool in modern ophthalmology [ 1 , 2 , 3 ], providing various OCT-based biomarkers, including retinal thickness, intraretinal cystoid fluid, subretinal fluid, alterations of outer retinal layers, and hyperreflective foci.…”