Discussion: View Thread

I'm curious to know whether any Consulting  section members have access to or work(-ed)  with quantum computing. 

An excellent recent paper from statisticians at UWisconsin on the role of statistics in quantum computing

https://pages.stat.wisc.edu/~yzwang/paper/quantum-ARSIA.pdf

excerpting the abstract

Quantum computing is widely considered as a frontier of interdisciplinary
research involving  elds ranging from computer science to
physics and from chemistry to engineering. The stochastic essence of
quantum physics renders the random nature of quantum computing,
thus there is an important role for statistics to play in the development
of quantum computing. On the other hand, quantum computing
has great potential to revolutionize computational statistics and data
science. This paper provides an overview on the statistical aspect of
quantum computing. We review the basic concepts of quantum computing
and introduce quantum research topics like quantum annealing
and quantum machine learning where statistics is heavily demanded.

also excerpting the paper - a claim from the media about the quantum computing field

...It is featured by many media as a story that a calculation of 3 minutes 20 seconds by a
quantum computer would take 10; 000 years for the most powerful supercomputer in the world

As I'm sure is true for all statisticians, computing, programming, numerical analysis and related were part of our training and education.

I recall my advisor in graduate school (Paul Levy) bragging about his brand new Apple computer which could run a regression and invert a matrix of a data set with as I vaguely recall, 10 variables and 500 rows of data. And Our Biostat department had purchased  an IBM Series 1 computer for something on the order of $100,000 (or other large $$$$ number) that had perhaps 500K  of memory and perhaps 3  megabytes of hard disk storage. IBM's technical history. at the time No one in the department knew how to program the Series 1. Our department chair was awarded an NIH grant and I was able to hire a programmer who could program the series 1. A portion of his programming time was figuring out how much of his program could fit into the limited computer memory and splitting up the computer program into small enough pieces to run the entire program

https://www.ibm.com/ibm/history/exhibits/vintage/vintage_4506VV4024.html

Not very long ago I recall attending a course at a JSM on the newest addition to computing toolbox-  Markov Chain Monte Carlo. The advice at the time for running MCMC, was set up the software, hit the "run button" and return 2 or 3 days possibly a week later to check for convergence and check for the results. 

As a grad student, I had the same type of issue. But, I was running on a 5.0GHz processor with 8 cores. 

When I hear claims about how Quantum Computing speeds up calculations, I have to wonder is it really the qubits or the algorithms. 

The simulation I ran on the "fast" computer was limited to using only 1 core of the 8 and using really poorly designed algorithms as well. When I upgraded from Basic R to Microsoft R Open, on that same computer, using quality algorithms and a multithreaded math library, that same calculation that took 2 weeks and usually didn't finish because the processor overheated, was done in under 45 mins. Further refinement of my code and the math library got larger calcs done faster. How much of quantum computing is done faster because of the processor? How much because of algorithms?

I'm curious to know whether any Consulting  section members have access to or work(-ed)  with quantum computing. 

An excellent recent paper from statisticians at UWisconsin on the role of statistics in quantum computing

https://pages.stat.wisc.edu/~yzwang/paper/quantum-ARSIA.pdf

excerpting the abstract

Quantum computing is widely considered as a frontier of interdisciplinary
research involving  elds ranging from computer science to
physics and from chemistry to engineering. The stochastic essence of
quantum physics renders the random nature of quantum computing,
thus there is an important role for statistics to play in the development
of quantum computing. On the other hand, quantum computing
has great potential to revolutionize computational statistics and data
science. This paper provides an overview on the statistical aspect of
quantum computing. We review the basic concepts of quantum computing
and introduce quantum research topics like quantum annealing
and quantum machine learning where statistics is heavily demanded.

also excerpting the paper - a claim from the media about the quantum computing field

...It is featured by many media as a story that a calculation of 3 minutes 20 seconds by a
quantum computer would take 10; 000 years for the most powerful supercomputer in the world

As I'm sure is true for all statisticians, computing, programming, numerical analysis and related were part of our training and education.

I recall my advisor in graduate school (Paul Levy) bragging about his brand new Apple computer which could run a regression and invert a matrix of a data set with as I vaguely recall, 10 variables and 500 rows of data. And Our Biostat department had purchased  an IBM Series 1 computer for something on the order of $100,000 (or other large $$$$ number) that had perhaps 500K  of memory and perhaps 3  megabytes of hard disk storage. IBM's technical history. at the time No one in the department knew how to program the Series 1. Our department chair was awarded an NIH grant and I was able to hire a programmer who could program the series 1. A portion of his programming time was figuring out how much of his program could fit into the limited computer memory and splitting up the computer program into small enough pieces to run the entire program

https://www.ibm.com/ibm/history/exhibits/vintage/vintage_4506VV4024.html

Not very long ago I recall attending a course at a JSM on the newest addition to computing toolbox-  Markov Chain Monte Carlo. The advice at the time for running MCMC, was set up the software, hit the "run button" and return 2 or 3 days possibly a week later to check for convergence and check for the results. 

I'm curious to know whether any Consulting  section members have access to or work(-ed)  with quantum computing. 

An excellent recent paper from statisticians at UWisconsin on the role of statistics in quantum computing

https://pages.stat.wisc.edu/~yzwang/paper/quantum-ARSIA.pdf

excerpting the abstract

Quantum computing is widely considered as a frontier of interdisciplinary
research involving  elds ranging from computer science to
physics and from chemistry to engineering. The stochastic essence of
quantum physics renders the random nature of quantum computing,
thus there is an important role for statistics to play in the development
of quantum computing. On the other hand, quantum computing
has great potential to revolutionize computational statistics and data
science. This paper provides an overview on the statistical aspect of
quantum computing. We review the basic concepts of quantum computing
and introduce quantum research topics like quantum annealing
and quantum machine learning where statistics is heavily demanded.

also excerpting the paper - a claim from the media about the quantum computing field

...It is featured by many media as a story that a calculation of 3 minutes 20 seconds by a
quantum computer would take 10; 000 years for the most powerful supercomputer in the world

As I'm sure is true for all statisticians, computing, programming, numerical analysis and related were part of our training and education.

I recall my advisor in graduate school (Paul Levy) bragging about his brand new Apple computer which could run a regression and invert a matrix of a data set with as I vaguely recall, 10 variables and 500 rows of data. And Our Biostat department had purchased  an IBM Series 1 computer for something on the order of $100,000 (or other large $$$$ number) that had perhaps 500K  of memory and perhaps 3  megabytes of hard disk storage. IBM's technical history. at the time No one in the department knew how to program the Series 1. Our department chair was awarded an NIH grant and I was able to hire a programmer who could program the series 1. A portion of his programming time was figuring out how much of his program could fit into the limited computer memory and splitting up the computer program into small enough pieces to run the entire program

https://www.ibm.com/ibm/history/exhibits/vintage/vintage_4506VV4024.html

Not very long ago I recall attending a course at a JSM on the newest addition to computing toolbox-  Markov Chain Monte Carlo. The advice at the time for running MCMC, was set up the software, hit the "run button" and return 2 or 3 days possibly a week later to check for convergence and check for the results.