Skip to content

Defense of a Ph.D. Dissertation- Gangotree Chakma

Gangotree Chakma,
Candidate for Doctor of Philosophy

Defense of Ph.D. Dissertation

Faculty Advisor: Dr. Garrett Rose

When: Friday, June 21, 2019, 11:00am-1:00pm

Where: Min H. Kao room 435

Design of a CMOS-Memristive Mixed-Signal Neuromorphic System with Energy and Area Efficiency in System Level Applications

The von Neumann architecture has been the backbone of modern computers for several years. This computational framework is popular because it defines an easier, simpler and cheaper design of the processing unit and the memory. Unfortunately, this architecture faces a huge bottleneck in going forward since complexity in computations now demands huge parallelism and this architecture is not efficient in parallel processing. Moreover, post-Moore's law era involves the constant demand of energy-efficient computing with fewer resources and less area. Hence, researchers are interested in establishing other alternatives to von Neumann architecture and Neuromorphic Computing is one of the few aspiring computing architectures that contributes to this research effectively. Initially, neuromorphic computing attracted the researchers because of the parallelism in the bio-inspired networks and they were interested in leveraging this advantage on a single chip. Moreover, the need of speed in real time performance also escalated the popularity of neuromorphic computing and different research groups started working on hardware implementation of neural networks utilizing the advantages of neuromorphic computing. Also, neuroscience is consistently helping in understanding the detailed connection of bio-inspired networks and has a huge contribution in bridging the gap between the biological neuronal activities and artificial neural networks. As a consequence, the idea behind the neuromorphic computing grew eventually. In this research, a memristive neuromorphic system for improved power and area efficiency has been presented. This particular implementation introduces mixed-signal platform to implement neural networks in a synchronous way. In addition to mixed-signal design, a nano-scale memristive device has been introduced to provide power and area efficiency for the system. The system also includes synchronous digital long term plasticity (DLTP), an online learning methodology that helps the system train the neural networks during the operation phase and improves the efficiency in learning considering the power consumption and area overhead. This research also proposes a stochastic neuron design which sigmoidal firing rate. The design introduces variability of membrane capacitance to reach different membrane potential leading variable firing rate.  

Friday, June 21, 2019 at 11:00am to 1:00pm

Min H. Kao Electrical Engineering and Computer Science, 435
1520 Middle Drive, Knoxville, TN 37996

Recent Activity