Page 152 - AIH-2-3
P. 152

Artificial Intelligence in Health                                     Explainable solutions from AI for HSSs




































                       Figure 2. Development process by the Intelligent Adaptive Clinical Platform as a Service editor of the knowledge base


                                                                  explain the results. Using software engineering
                                                                  techniques, designers and programmers then declare
                                                                  and implement solver components (agents). Notably,
                                                                  creating  a  knowledge  base  and  creating  a  problem
                                                                  solver (particularly assembling program blocks from
                                                                  the library) are parallel processes. The ontological
                                                                  structure ensures the compatibility of the components
                                                                  within the portals created.
            Figure 3. The example of the Intelligent Adaptive Clinical Platform as a   5.3. Ensuring clinical KESs’ viability
            Service medical document ontology
                                                               The properties of KESs related to viability are provided
               consideration, e.g., for risk assessment or prevention,   by the development environments implemented on the
               the knowledge engineer uses the ontology editor   IACPaaS platform. Consider the example of the knowledge
               to describe a set of relevant concepts, relationships,   updatability property and support for knowledge updating.
               and  constraints  (creating  a knowledge  ontology  for   When an update to the knowledge base is required
               the problem or task). Often, the engineer will work   in  connection  with  the  acquisition  of  new  knowledge
               with the expert to establish a set of decision rules   (statements), it is possible to modify it manually. This
               (ontological agreements). A new task may require an   development (production) environment has knowledge
               extension of the thesaurus                      base editors.
            (ii)  Formation of the knowledge base by medical experts.   Then, it is important to evaluate the consistency with
               The IACPaaS knowledge base editor (for example,   the available facts.  It must check the non-decrease of a
               based on the extended knowledge ontology) may be   set of correctly solved tasks when replacing the version
               required. It is possible to formulate clinical guidelines   of the knowledge base (according to the importance of
               without obstacles. Using the tools for translating text   monotonous improvement of knowledge bases ). The
                                                                                                       29
               documents into a given structure is advisable. As a   procedure for checking the non-deterioration of knowledge
               rule, creating the knowledge base is a collective process  is as follows: from each reference task (precedent), enter
            (iii) If the creation of a solver is required, then the designer   input conditions into the solver integrated with the new
               uses the ontology editor to create an ontology to   version of the knowledge base, and obtain the explanation


            Volume 2 Issue 3 (2025)                        146                               doi: 10.36922/aih.5736
   147   148   149   150   151   152   153   154   155   156   157