This document summarizes a research paper that proposes a new neural network algorithm called C-Mantec. C-Mantec adds competition between neurons using a thermal perceptron learning rule. This allows existing neurons to continue learning even as new neurons are added. The algorithm was tested on a diabetes dataset and shown to generate compact neural network architectures with good generalization capabilities. It was implemented in an FPGA using Xilinx ISE for synthesis and ModelSim for simulation. The output was obtained over three clock cycles, first loading inputs/weights, then multiplying/accumulating weights, and finally checking the output against a threshold. The study showed C-Mantec can effectively model glucose level fluctuations to determine if a patient has diabetes.