Maldonado Huayaney, Frank Lucio: VLSI implementation of a calcium-based plasticity learning model. 2018
Inhalt
- CERTIFICATE OF ORIGINALITY
- Abstract
- 1 Introduction
- 1.1 Microprocessor Evolution and Technology Challenges
- 1.2 Learning in Autonomous Systems
- 1.3 The Neuromorphic Approach
- 1.4 Outline of this Thesis
- 1.5 Acknowledgement to the Contributors
- 2 Models of Synaptic Plasticity
- 2.1 The Spike Response Model
- 2.2 Synaptic Plasticity
- 2.3 The Simplified Calcium-based Learning Model
- 2.4 Discussion
- 3 Neuromorphic Circuits Blocks
- 3.1 CMOS Operation in Inversion Region
- 3.2 MOSFET Characterization
- 3.3 Mismatch
- 3.4 The Diff-Pair Integrator Circuit (DPI)
- 3.5 The Operational Transconductance Amplifier (OTA)
- 3.6 The Winner-take-all Circuit
- 3.7 Discussion
- 4 First Synapse Circuit Implementation
- 5 Second Synapse Circuit Implementation
- 5.1 The Calcium Circuit
- 5.2 The Synapse Core and The Bistability Circuits
- 5.3 The Linearizer
- 5.4 The Configurable Bias Current Generator
- 5.5 The Neural Network Block
- 5.6 Simulation Results
- 5.6.1 Bistability
- 5.6.2 Potentiation and Depression
- 5.6.3 STDP Waveform
- 5.6.4 Configurable Bias Circuit
- 5.7 Hardware measurement results
- 5.8 Discussion
- 6 Network Operation
- 7 Mismatch Characterization
- 8 Conclusions
- Bibliography
- Publications