General purpose representation and association machine: Part 4: Improve learning for three-states and multi-tasks
Huihui Li, Lei Wei · 2013
In this paper, we continue our journey on how to use simple LDPC codes to build a GPRAM prototype. We examine a switch function in the system which will allow us to perform revision in code structures in the GPRAM prototype in future. In this study we focus on how to improve learning for threestate systems; how to learn using perfect codewords; and finally how to perform progressive learning methods to perform multi-tasks. We start to see that many interesting features begin to merge and some of them may have bio-implications which need to be explored further.