UPDATE ON QUANTUM COMPUTING AND DATA & ANALYTICS



So what is Quantum Computing (QC) and, more importantly, why should we care when it comes to Data & Analytics? 


What Quantum Computing is


Traditional computers use 0's and 1's, the bits, and they take one set of inputs and generate one set of outputs. In terms of analytics, we have applied in-memory computing and parallelism (think Spark), and we can leverage lots of hardware to get things going fast. Still, we are using the same primary computer construct, just smaller and faster. 

I won't try to explain the physics behind QuBits (superposition and entanglement, and how we can build a quantum computer with them). Suffice it to say that they can represent both 0 and 1 at the same time; this is what gives a quantum computer POWER. Like, a lot of power. If we have 10 QuBits in such a computer, we can conduct calculations on up to 2xp10, or 1024, input sets at once.


Let's Natural Language Processing since it is an application area of AI today. Imagine all the books in the world digitized somewhere. Whereas a traditional computer will read all the books sequentially to find a term, the quantum computer will read the books SIMULTANEOUSLY. Again, I will spare you how, but the concept is significant for us working with data and trying to get value out of it.


We have come a long way in computing power already, but we are primarily still processing data as we did in the 1960's - sequentially - just faster. When NASA put a man on the moon, they did so with a computer that was 35,000 times slower than my junky iPhone 6s (I gave it to my kid, a $1000 iPod). QC will change how a computer works with data in a fundamental way.


We can now solve problems that were just impossible to address with a 1969 computer - it was just too long to get the answer to be useful, but it did help put a man on the moon. The same leap will occur when QC starts delivering real business in the 2025 -2028 range, according to McKinsey. Still, we should start seeing POCs this year or next that will exhibit just how transformational this will be.  


Technical Obstacles


If you were trying to leverage snowflakes for computing, imagine the difficulties and obstacles you would have to deal with. QuBITs a little like that at the moment; they are a bit unstable. 


For a regular computer to work, bits are either 0 or 1, that's it. A lot of work goes into making sure that there is no confusion as to their states and that they don't affect adjacent bits. By contrast, QuBITs can be both 0 and 1, and entanglement means their interaction with each other is part of how they do their magic. Making sure that happens correctly and not because of instability is a difficult undertaking.


How will Quantum Computing be Used


One way in which quantum computers will differ is the way they will give their answers. We are used to having very discrete results with traditional computers. The nature of simultaneous calculations at once makes QC different in the sense that the results will be more like a range of possible answers. 


For one, this means that the traditional computer will not be going away. Second, QC will likely be used for specific and very complex use cases, where a lot of computations are required and more to eliminate ranges of possibilities. 


One application area will be simulations. Imagine applying simulations in investments like this to eliminate money-losing trades, or to optimize investment portfolios. What about in the development of drugs where the need to connect genomes and proteins create impossible problems because regular computers can't adequately simulate molecules, let alone the interactions between them. 


Many long-key encryption algorithms today that protect data cannot be solved with today's computer, hence providing adequate protection. QC will be able to address that in a pinch, so one area that will be disrupted is squarely data security. 


In short, QC will have the capability to transform industries by solving problems that are elusive today.


The Next Ten Years


Quantum computers will be expensive, and only a few players will start putting them to work, such as Google and IBM. McKinsey estimates that "by 2030, only 2,000 to 5,000 quantum computers will be operational." We know that the cloud providers will get into the fray, and that will be a crucial factor that will accelerate adoption. 

Here is a bar chart with the distribution of use cases where we can expect that QC can deliver value.


How to Prepare for QC


It turns out that the advent of quantum AI is farther off, and we shouldn't expect quantum computers to be powerful enough for prime time in that space until the very late 2020s at the earliest. Still, based on the industry, organizations should start thinking about having a Quantum Strategy, so they don't get left behind. 


And it is not just about how to reap benefits either. There is a defensive aspect to it. If we think of companies that have trade secrets or rely on years of market data to eke out an advantage over new entrants, they should be concerned about how all this will play out in terms of cyber-security disruption. If nothing else, quantum cryptography should be on the mind of Chief Data Officers.

0 views

©2020 by Modern Data Analytics