FCMAC-Yager: A Novel Yager-Inference-Scheme-Based Fuzzy CMAC

Abstract

The cerebellum is a brain region important for a number of motor and cognitive functions. It is able to generate error correction signals to drive learning and for the acquisition of memory skills. The cerebellar model articulation controller (CMAC) is a neural network inspired by the neurophysiologic theory of the cerebellum and is recognized for its localized generalization and rapid algorithmic computation capabilities. The main deficiencies in the basic CMAC structure are: (1) it is difficult to interpret the internal operations of the CMAC network and (2) the resolution (quantization) problem arising from the partitioning of the input training space. These limitations lead to the synthesis of a fuzzy quantization technique and the mapping of a fuzzy inference scheme onto the CMAC structure. The discrete incremental clustering (DIC) technique is employed to alleviate the quantization problem in the CMAC structure, resulting in the fuzzy CMAC (FCMAC) network. The Yager inference scheme (Yager), which possesses firm fuzzy logic foundation and maps closely to the logical implication operations in the classical (binary) logic framework, is subsequently mapped onto the FCMAC structure. This results in a novel fuzzy neural architecture known as thefuzzy cerebellar model articulation controller-Yager (FCMAC-Yager) system. The proposed FCMAC-Yager network exhibits learning and memory capabilities of the cerebellum through the CMAC structure while emulating the human way of reasoning through the Yager. The new FCMAC-Yager network employs a two-phase training algorithm consisting of structural learning based on the DIC technique and parameter learning using hebbian learning (associative long-term potentiation). The proposed FCMAC-Yager architecture is evaluated using an extensive suite of real-life applications such as highway traffic-trend modeling and prediction and performing as an early warning system for bank failure classification and medical diagnosis of breast cancer. The experimental results are encouraging.

A nature inspired Ying–Yang approach for intelligent decision support in bank solvency analysis

“…These are global learning structures. Furthermore, efforts to develop localized semantic learning memory structures that functionally mimic the neocortical semantic association include (Ng, Quek, & Han, 2007;Nhut, Shi, & Quek, 2006;Sim, Tung, & Quek, 2006) as well as neural associative memories (Quek & Ang, 2000;Zhou et al, 1996). Advanced knowledge reduction techniques are also developed to improve the interpretability of the semantic memory structures (Ang & Quek, 2005;Liu, Quek, & Ng, 2007;Quah & Quek, 2007).…”

A brain-inspired fuzzy semantic memory model for learning and reasoning with uncertainty

“…Fuzzy and rough sets methodology that map formal fuzzy logical structures onto neural systems to perform the fuzzy set derivation and rule identification/ reduction processes are investigated [40,41]. These research endeavors culminated with the developments of the human hippocampus-inspired learning memory systems such as GenSoFNN [10,28,35], Pseudo Adaptive Complementary Learning networks [42,43] and POPFNN [44][45][46], as well as cerebellar-based computational models [47,48] for the modeling of complex, dynamic and non-linear problem domains. The application of these brain-inspired learning memory systems is actively pursued, and they have been successfully applied to automated driving [49], signature forgery detection [50], gear control for continuous-variable-transmission in automobile [51], fingerprint verification [52], medical decision-support [28,53] and computational finance [35,46,54].…”

Hierarchically Clustered Adaptive Quantization CMAC and Its Learning Convergence

“…Other CMAC variants such as the kernel CMACs [48]- [50] and the fuzzy CMACs [51], [52] do not employ quantizationbased addressing schemes. In kernel CMAC, kernel functions such as B-spline kernels are used to derive the receptive fields of the network.…”

“…Fuzzy CMACs, on the other hand, employ input fuzzification methods to remove the sharp quantization boundary of the original CMAC network. The use of fuzzy inference schemes such as in [52] also enables the extraction of fuzzy rules from the trained network. While these variants offered the computational interpretability missing in the black-box CMAC model, there is a significant increase in the computational complexity of the hybrid network without a comparable performance gain.…”