Nick N. Oza
Tools: Calibration

Personal Background

Born and raised in the Silicon Valley, I grew up in the heart of technological innovation with engineering running through my blood. At the end of my high school experience, with any of many paths in front of me, I chose to become an engineer because my father, also an engineer by training, pointed out that no other field of study teaches the students how to think about and solve problems. As he put it, “No matter what you choose to do in the future, you will need to solve problems. By becoming an engineer, you will be taught the tools you need to efficiently, effectively, and accurately analyze and solve problems.” Using this backdrop at the start of my career, I set out to learn how to solve problems.

I first learned to solve problems as a Computer Engineering major from UC Santa Barbara where I graduated with my Bachelor’s Degree. My undergraduate research and academic interests were very varied, ranging from Optical Communications to Computer Networking and Distributed Systems. I eventually settled on digital design and computer architecture because I thought it would be fascinating to design new, faster, more power-efficient computers.


Having enjoyed my research experience at UCSB, I attended Stanford University to pursue a Master's degree in Electrical Engineering. While at Stanford, I took a course in neuroengineering taught by Krishna Shenoy. During his class, I had a moment of epiphany, where I realized that neuroengineering was a truly fascinating field – the ideal intersection of interdisciplinary work, learning, and engineering with a direct benefit to society.


While studying how to be a better engineer, I also saw a program at Stanford that offered the opportunity to learn how to take cool engineering projects to market, while creating a successful company. As such, I also earned an MS in Management Sciences and Engineering. After a few years in industry at a start-up company, shoring up my PCB design and testing skills, a Staff Position as a Hardware Engineer at Brains in Silicon offered me the opportunity to return to neuroengineering.

Project Goals

Growing up, I remember having a book of facts that stated that the world’s most powerful computer was the human brain. This little tidbit fascinated me. How is it that our brain can perform so many advanced, real-time tasks using only 10 watts? Only recently, with all of the technological progress that we have seen over the past 30 years, can a computer, finally, drive a car automatically. Even then, it requires hundreds of watts to do so. As such, we still face the problem of building more powerful, but efficient, computers. The human brain offers some insight into possible, viable ways to engineer a solution.


I have been heavily involved with understanding, characterizing, and calibrating our current circuit board, Neurogrid. Neurogrid is composed of 16 chips called Neurocores. Each Neurocore emulates 65k neurons using a mixed-signal design. Each neuron is designed leveraging the analog characteristics of transistors, while the communication between neurons is implemented digitally. On each Neurocore, setting current biases controls the neuronal parameters. Calibration is the process that ensures that a certain adjustment to the current biases results in the expected neuronal behavior. As an analogy, this is like turning a knob that controls your home’s thermostat. With a poorly calibrated thermostat, when you turn your knob to set it at 70°F, it may set the house’s temperature to 60°F – something vastly different. A properly calibrated thermostat would set it at the value that you desired. Similarly, we need to ensure the proper calibration of Neurogrid; otherwise, as a simulation platform, it is useless to neuroscientists.


Looking forward, I am eagerly anticipating the opportunity to help design our next chip and the surrounding circuitry necessary to interface with the chip. The next chip will leverage much of the research done by Alex and Sam to implement an efficient hardware backend for the Neural Engineering Framework (NEF), developed by Chris Eliasmith and Charles Anderson. This chip will perform advanced tasks, such as robot control, in a power-efficient, biologically plausible manner. Developing this chip will be an exciting endeavor into the world of nanoscale design.


Publications

ID Article Full Text
J44
A Neckar, S Fok, B Benjamin, T C Stewart, N N Oza, A R Voelker, C Eliasmith, R Manohar, K Boahen, Braindrop: A Mixed-Signal Neuromorphic Architecture With a Dynamical Systems-Based Programming Model, Proceedings of the IEEE, vol 107, no 1, pp 144-64, 2019.

Full Text

 

Miscellaneous Facts

  • I have a twin brother who is getting his PhD in Electrical Engineering from Northwestern University.
  • When I am not working, you can find me playing any of many sports, including but not limited to volleyball, ultimate frisbee, and soccer.
  • I love attending Stanford athletic events, especially football, soccer, and volleyball.
  • I spend many of my Sunday mornings teaching South Asian styles of dance.
  • I have a deep passion for good communication, and towards that end, I help tutor and enlighten others in the art and science of public speaking.