Design and Comparison of Memristor Implementation for Different Machine Learning Algorithms

Please enable javascript in your browser.

Haghighat, Bahar | 2012

1117 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 43085 (05)
University: Sharif University of Technology
Department: Electrical Engineering
Advisor(s): Bagheri Shouraki, Saeed
Abstract:
The first physical realization of the missing fourth fundamental element of electrical circuits, namely memristor, in 2008 by HP labs triggered an immense amount of research on the capabilities of this element in implementing artificial neurons and artificial brain. In this project we will propose several reinforcement learning-based algorithms that are implemented on a specific memristor-based structure, the memristor crossbar structure. Hence we provide a learning paradigm that resembles the human learning paradigm not only because of the the algorithmic core, which is based on learning from sparse and delayed rewards and penalties, but also because of the hardware over which the algorithms are run. As a matter of fact, such achievement implies that we have been able to encode the inherent similarities between the biological brain inspired algorithms and also the hardware. Since the resulting paradigm is able to conquer several challenges in both domains, we can claim that we have actually found a missing ring linking the domains of biologically inspired hardware and software
Keywords:
Machine Learning ; Reinforcement Learning ; Fuzzy Modeling ; Active Learning ; Memristor