By Scott Hamilton
Senior Expert Emerging Technologies
I had a revelation a few months ago while preparing for a talk at an Artificial Intelligence and Machine Learning conference that is coming up in a few weeks. I was reading a document on the inner workings of quantum computers and realized something I found to be both shocking and amazing. I am not sure if many of you reading the column are old enough to remember the first computers ever invented, but they were based on analog circuits.
First I would like to explain the difference between analog and digital signals, as that will help you to understand the difference between digital and analog computers. When you talk to someone or listen to some speak, you are hearing an analog signal in the form of a sound wave. The natural analog signals have no real upper or lower limit of signal strength; in electronics there are limits to the signal strength depending on the level of power available to the device. Analog signals operate across the range of their defined limits with exact precision. For example a transistor, which is the key electronic device in both analog and digital signals, normally operates in a voltage range between zero and five volts. An analog signal can be any value between zero and five volts, for example 3.73 or 1.98. A transistor operating on analog signals is working in transition mode, meaning it is somewhere in transition between being fully turn on and fully turned off.
Digital signals are kind of the opposite of analog signals in that a digital signal can only be valid for values of fully on or fully off. The transistor is in a stable state when operating on digital signals and is either fully on, five volts, or nearly off, less than 0.3 volts. These are referred to as the logical on and off in digital circuits.
The really crazy part of computer research is that the very first computers used vacuum tubes, which were like transistors in how they operated, but much larger and required a lot more power. They used these tubes to add, multiply, subtract and divide analog signals to solve mathematical formulas. These tubes were not manufactured for precise mathematical functions and as a result, were unreliable because they created noise in the signals, resulting in a high error rate in their computations.
Over a period of about 25 years computer researchers began to realize that they needed to eliminate the noise if they were to get accurate results out of the computers. This resulted in the development of the binary number system and forced these tubes into a saturated state, meaning fully on or off, to represent numbers in the binary system. This brought about the birth of modern digital computers.
We have now come full circle in the computer industry, with the latest research devices being quantum computers that are once again operating in the analog space. They are facing the same issues seen in the early days of computing. The noise introduced due to the imperfections in the manufacturing techniques cause high error rates in quantum computing devices. Unlike the early analog computers, which were fairly easy to understand, quantum computers seem much more complex on the surface, but as you begin to dig deeper into the science behind them, they are really analog systems operating in three dimensions; three signals operating on three axis, controlling a single quantum bit. The early analog and modern digital computers operated on a single dimension, one signal on one circuit path. Also the three signals in quantum circuits are actually just representations of the modeled quantum processes. It leaves me with one question; when we cleaned up the signal of the analog computers we got the current two-state digital computers. What will the cleanup signal of a quantum computer look like? If we can figure that out, we will have the first successful quantum processor. If you have ideas regarding this, I would love to hear them.
Until next week, stay safe and learn something new.
Scott Hamilton is a Senior Expert in Emerging Technologies at ATOS and can be reached with questions and comments via email to firstname.lastname@example.org or through his website at http://www.techshepherd.org.