Skip to Main Content
Fuzzy systems have been successfully applied to solve many engineering problems. However, traditional fuzzy systems are often manually crafted, and their structures (knowledge rule-bases) are static and cannot be trained or tuned to improve the system performance. This subsequently leads to an intense research on the autonomous construction and tuning of a fuzzy system directly from the observed training data to address the knowledge acquisition bottleneck. However, the complex and dynamic nature of real-world problems demanded that fuzzy systems be able to adapt their structures, parameters and ultimately evolve their intelligence to continuously address the non-stationary characteristics of their operating environments. This paper presents the evolving fuzzy semantic memory (eFSM) model, a neuro-fuzzy architecture with a continuously adaptive structure (rule-base). The computational principles responsible for the online identification of the proposed eFSM model and its evolving capability are based on the functional mechanisms of the human hippocampus, a brain construct that plays a significant role in the acquisition of the long-term human declarative memories.
Date of Conference: 1-6 June 2008