Concept of validation and its tool for intelligent systems

One problem in applying intelligent methods to real applications is to validate their expertise. Turing test approaches are useful in validating intelligent systems. Nevertheless, validation burdens experts with a heavy task. A concept called the “bi-directional, many-sided explanation typed multi-step validation method”, which enables the sharing of validation tasks among experts, knowledge engineers and computers, is proposed to diminish this load. The principal characteristics of the validation tool are as follows: 1) effective and anonymous bidirectional interactive functions by way of a virtual anonymous ID, which enables the protection of the experts´ social position; 2) a response accelerator that greatly improves the responsiveness of many-sided explanations by providing an experts´ validation server to fetch and store large volumes of information in advance; and 3) a validation process management function that detects delays in busy expert validation processes