To maximize the efficiency of knowledge learning, it is essential that the knowledge system itself be well structured. Well designed knowledge systems make easy to access for knowledge acquisition and extraction. Expert knowledge plays a role controlling. We propose a Hierarchical modular system with an expert-knowledge gating mechanism that consists of mechanisms for acquiring knowledge, constructing associative memory and enabling knowledge inference and extraction based on expert-knowledge gating. We applied this to medical diagnostics for classifying Viruses (coxackie, echovirus and cold virus), Rhinitis (Nonallergic and allergic) and tested using symptom data.