Tuesday, October 14, 2014

Artificial Intelligence and Decision Making


In a recent discussion in my Enterprise Models class, a classmate and I discussed the limitations of Artificial Intelligence theories and human emotions. Here is my response:

From the research I have been doing over the years on AI specifically, one of the biggest challenges is how to program emotions into a computer system. I think there are two primary problems currently. One, and the main problem, is that modern computing technology processes thing in a linear fashion, every time slice of a CPU cycle is occupied by either a 1 or a 0. There is no middle ground, there is no gray area. Everything is black or white, and follows a strict logic rule set. What is currently being done with systems like Watson and Google's web crawler software is using software to simulate scenarios and have the hardware crunch the data, while another part of the software provides the processing logic through algorithmic manipulation thereby creating an intelligent system. Current intelligent systems are limited by the scope of their programming environment. Two, there isn't a programming language that yet exists that can accurately tell a computer how it needs to do what it needs to do in order to understand the logic behind a feeling. Most of the researchers I have found over the years say that technology isn't there yet, and I happen to agree. The possible solution to this quandary could be quantum computing.

With quantum computing a quibit offers a system the ability to see a data stream in two states simultaneously. Each quibit is BOTH on and off (1 and 0) in the same "time slice" of a processing cycle, leveraging the power of superposition and entanglement. This allows the system to perform many operations on the same data stream. Neural networks simulate this through software, but over hardware that still processes data in a linear fashion. What we need is the hardware to perform this, because it can perform it much faster than software could ever process the same data stream. Enter quantum computing. D-Wave Systems is the current leader in true quantum computing with their current D-Wave quantum computer, but their system is highly specialized at the moment due to a lack of programming knowledge...while the system has amazing potential, as you will see form a couple of the links below, no one really truly understands how to use it. There are other links below with details on their system and methodology.

The problem with quantum computing is it requires a completely new way of perceiving computers and also a completely new way for users to interface with computers, not to mention new hardware that performs in ways modern hardware cannot. That is what I see as the next way of technological evolution. As transistors become subatomic through the help of graphine and carbon nanotubes, and technologies like memristors look to shatter our perceptions on information storage capacities and data throughput, quantum computers will become more common place across the landscape. The ability to create a true quantum system capable of processing complex emotional patterns is very real. Once we have a true quantum processor, and a true quantum operating system, then we will not only have the power to process it in fractions of nanosecond but also the programming logic and syntax to leverage an intelligent system, and possibly create a sentient computer system, otherwise known as AI.

AI is an fascinating concept, and exactly why it will be the focus of my post grad work. Quantum computing has been a subject I have dreamed about and followed since I was a young boy, before computers were common place and technology was still considered a super luxury. Today technology is seen as a necessary commodity, but there are still concepts that have yet to be discovered or invented, and quantum computing is currently the field of interest. Once we researchers and scientists figure it out, it will change the world.


D-Wave System References:
http://recode.net/2014/09/25/d-wave-ceo-our-next-quantum-processor-will-make-computer-science-history-video/
http://www.dwavesys.com/quantum-computing
http://www.dwavesys.com/d-wave-two-system
http://time.com/4802/quantum-leap/


Quantum Computing References:
http://techland.time.com/2013/09/25/the-carbon-nanotube-breakthrough-moores-law-needs-to-survive-well-see/
http://phys.org/news387.html
http://www.physics.ox.ac.uk/nanotech/research/nanotubes/index.html
http://www.tum.de/en/about-tum/news/press-releases/short/article/30589/

Sunday, October 5, 2014

Technological Evolution - Quantum Computing, Memristors, and Nanotechnology

It is amazing how evolution of technology changes perspectives so quickly on the future. With holographic interactive screens currently in use, memristors and atomic-level transistor technologies at our fingertips, and new developments in using light as a means to interact with systems or store system data, the reality of AI and systems like Jarvis are finally able to go from drawing board concept to real life prototype. For as long as I can remember, I have been talking about quantum computing and nanotechnology and how that is the future of systems and human interactions. As a younger teen, when I first started learning about quantum mechanics and ultra microscopic hardware theories, I saw then that the future of computer systems and computer-human interactions were going to be largely logic based and function faster than the speed of human thought. By marrying the concepts of quantum mechanics and advanced computer system theory, intelligent systems and true AI are highly viable and will be here within the current generation. As advances in nanotechnology take transistors to the subatomic level, and theories in quantum computing become a reality, we are quickly going to see the industry change as the traditional system paradigm is shattered, and a new evolution in technology is ushered in - I would call it the quantum age - where Schroedinger's cat can exist in both physical states without the concern of observation tainting the true reality of the objects existence. The potential gains with quantum processors and quantum computing methods that scientists around the world are currently developing into physical models are, at the moment, limited only to manufactured hardware capacities. As physical hardware capacities become perceived as unimportant to system planning schemes - due to advances like the memristor and photonics, including the newest nano-laser (see reference) - the focus can be given to writing programs that can take advantage of this new systems paradigm. What is going to take time is the change in mindset to understand how to use a quantum system because it requires a completely new approach to hardware and software design. Modern systems process data in a linear manner, processing one bit after another based on the time slice protocol programming in to the operating systems and CPU itself. Regardless of how many processors you throw at a system, it still only processes one bit of data at any given time slice. The fastest super computer, China's Tianhe-2, can process more than 6.8 quadrillion bits per second (3.12 million processors x 2.2GHz each = 6,864x10^12 processes per second), but it still only processes one bit at a time. Quantum systems do not function in this manner, they function in a far different reality where a bit can be both a 1 and a 0 simultaneously within a single time slice, though quantum processors would not use a time slice function, it would require something else yet to be defined. As scientists gain a better understanding of how to create a truly quantum computer systems, and quantum capable operating system, we will see technology advance to arenas yet to be discovered. What we once called science fiction, is quickly becoming scientific fact.

~Geek


References:http://www.sciencealert.com.au/news/20143009-26256.html (nano-laser)

http://www.top500.org/system/177999 (Tianhe-2 details)