A5067558566- Advanced Memory and Neural Computing
- CCD and CMOS Imaging Sensors
- Ferroelectric and Negative Capacitance Devices
- Photoreceptor and optogenetics research
- Advanced X-ray and CT Imaging
- Neural dynamics and brain function
- Reservoir Engineering and Simulation Methods
- Neural Networks and Reservoir Computing
- Machine Learning in Materials Science
- Neuroscience and Neural Engineering
- Advanced Neural Network Applications
- Autonomous Vehicle Technology and Safety
- Topic Modeling
- Semiconductor materials and devices
- Scientific Computing and Data Management
- Advanced Optical Sensing Technologies
- EEG and Brain-Computer Interfaces
- Computational Physics and Python Applications
University of Tennessee at Knoxville
2020-2024
Knoxville College
2023
Memristors are a suitable candidate to design synapse circuits and neuromorphic systems. Due device voltage variability, operating memristive with reliability is big challenge. To enhance the of synapse, RESET failure needs be considered. In this work, we focused on modeling variation. Here, defined as hard due high being applied. The proposed Verilog-A model derived based experimental data collected from 1T1R devices, which fabricated 65 nm CMOS process. system-level simulation, will...
Memristors provide a tempting solution for weighted synapse connections in neuromorphic computing due to their size and non-volatile nature. However, memristors are unreliable the commonly used voltage-pulse-based programming approaches require precisely shaped pulses avoid failure. In this paper, we demonstrate current-limiting-based that provides more predictable analog memory behavior when reading writing memristive synapses. With our proposed design READ current can be optimized by about...
The synapse is a key element of neuromorphic computing in terms efficiency and accuracy. In this paper, an optimized current-controlled memristive circuit proposed. Our proposed demonstrates reliability the face process variation inherent stochastic behavior memristors. Up to 82% energy optimization can be seen during SET operation over prior work. addition, READ shows up 54% savings. approach also provides more reliable programming traditional methods. This design demonstrated with 4-bit...
This paper presents a new scalable 8 × single photon avalanche diode (SPAD) based vision sensor with integrated spiking neuromorphic system on chip. The proposed sensing adopts the benefits of SPAD's high quantum efficiency and energy memristive processing. SPAD includes biologically inspired address event representation (AER) readout to generate asynchronous digital events at output reducing computation making it suitable process directly on-chip in faster more efficient way. A novel...
Neuromorphic computing systems have emerged as powerful computation tools in the field of object recognition and control systems. However, training these systems, which are usually characterized by recurrent connectivity, requires abundant computational resources: memory, computation, data, time. Reservoir (RC) framework reduces this high cost focusing effort on only a small subset connections thus allowing to be amenable hardware implementation. Using memristors construct reservoir...
The idea of a self-driving car is one which actively studied and tested for use on the road. In addition, machine learning tools required to create such vehicle has become more available public as time goes on. With number different libraries softwares free download design train neural networks with affordable but powerful miniature computers market, can explore possibility creating vehicle. goal our project was construct small scale using parts software that are accessible anyone an budget...
In this paper, we present a novel on-chip interface for event based sensor and spiking neuromorphic processing. More specifically, design an converting the output events of single photon avalanche diode (SPAD) into Temporally Coded Spikes (TCS) enabling processing with integrated system. The has been implemented in 65 nm CMOS process. Results show great potential proposed to reduce gap between real time processor. enables complete sensing systems processors chip resulting compact system high...
This paper presents some of the current challenges in designing deep learning artificial intelligence (AI) and integrating it with traditional high-performance computing (HPC) simulations. We evaluate existing packages for their ability to run models applications on large-scale HPC systems efficiently, identify challenges, propose new asynchronous parallelization optimization techniques heterogeneous upcoming exascale systems. These developments, along AI software capabilities, have been...
The synapse is a key element of neuromorphic computing in terms efficiency and accuracy. In this paper, an optimized current-controlled memristive circuit proposed. Our proposed demonstrates reliability the face process variation inherent stochastic behavior memristors. Up to 82% energy optimization can be seen during SET operation over prior work. addition, READ shows up 54% savings. approach also provides more reliable programming traditional methods. This design demonstrated with 4-bit...
Memristors provide a tempting solution for weighted synapse connections in neuromorphic computing due to their size and non-volatile nature. However, memristors are unreliable the commonly used voltage-pulse-based programming approaches require precisely shaped pulses avoid failure. In this paper, we demonstrate current-limiting-based that provides more predictable analog memory behavior when reading writing memristive synapses. With our proposed design READ current can be optimized by about...