BP's Quantum Computing event

My first brush with Quantum computing was yesterday thanks to an invitation by Claus Nehmzow to the BP event on Quantum Computing (QC). The event consisted of 3 presentations, followed by a panel discussion.

Before the event my friend Avishek Ghosh gave a brief primer on what Quantum computing is. The key takeaways from that discussion was:

1.     QC is a new way of computing that relies on probabilistic determination and massive parallel computing. 

2.     It moves away from the Binary states of the bits in the classical computers to the quantum states of qubits. (This is something that I have not completely understood. The challenges of the cat being alive and the dead at the same time is still very perplexing for me :D)

3.     While the QC approach is a very new way to represent and compute data, the physical environment is very challenging and often require near absolute zero temperatures. Slight changes in the physical environments can impact the data and the current generation of the Quantum systems is extremely error prone. 

4.     Error checking is a big consideration for this emerging technology and IBM has done some work to make this technology accessible to the general public.

Presentation 1: Overview CQT (Centre for Quantum Technology)

The first talk was by Chune Yang Lum, the Advisor, Industry Relations, Centre for Quantum Technologies. CQT was established as Singapore's inaugural Research Centre of Excellence (RCE) in 2007 with a funding of almost $150 million. Based out of NUS, this 180-man organization has people of 30+ nationalities with the bulk coming from NUS, NTU and SUTD. 

My general impression about research bodies is that they are very publication focused. This is also true for CQT but my biggest takeaway from the session was that they were making a serious attempt to bridge the gap between the industry and the academia. Chuen Yang also made the interesting point of how they were trying to change their language from technical to more business like. I found the approach sincere but a lot would be needed before the work can make a significant business impact. 

CQT has also done some interesting projects like: 

·      Testing the Singtel fiber network for Quantum considerations, 

·      Cutting down the fabrication time of Quantum chips from 6 months to under a month for DSO national laboratories. 

·      Design of a 9 qubit processor in a collaboration with Google to study chemistry problems. 

·      MOU with the Infocomm Media Development Authority to promote Quantum technologies. 

In addition CQT alumni have gone on to startup more than 6 organizations in the Quantum space. This emphasis on application was impressive. 

Presentation 2: Quantum Computing - Entropica Labs

Ewan Munro, the founder of Entropica Labs gave a beautiful overview of the Quantum Landscape. His clean slides were instructive even for some like myself whose knowledge of the field is limited to the jargons. 

He highlighted the challenges around errors and predicted that in the near term (less than 5 years) you would have Noisy Intermediate Scale Quantum Computing (NISQ) with about 1000 noisy qubits. There could be interesting possibilities for the Quantum Processing Units (QPUs) to act as co-processors in a hybrid quantum-classical approach to problem solving. 

He predicted the arrival of large-scale fault tolerant quantum computing with a processing power of about 50-100+ clean qubits in about a decade’s time when those Quantum processors could be used for end-to-end computations.

Describing the application of quantum computing, he highlighted the study of chemical reaction mechanisms. He went on to highlight how ammonia production for fertilizers consumes about 2% of the world's total energy consumption. The Haber process discovered in the early 1900s uses significant temperature and pressure conditions. However, bacteria at ambient temperature and pressure conditions can achieve the same results. Simulation of this mechanism is impossible on classical computers. However a quantum computer with around 200 clean qubits should be able to solve this problem in about a day!! In the nearer future Munro predicted the use of NISQ systems for small-molecule chemistry simulations that could remove the need for approximation methods that introduces significant errors. Apparently IBM has made some interesting progress in this field. He also talked about how Quantum approaches could enhance optimizations and machine learning outcomes.

He also highlighted how large companies with long outlooks like Boeing have identified complex computational challenges and put it out in the open to crowd source solutions. Some of the problems in Airbus quantum computing challenge included:

·      Aircraft climb optimization.

·      Aircraft loading optimization.

·      Wingbox design optimization.

·      Computational fluid dynamics.

·      Quantum neural networks for solving partial differential equations.

In summary he made two strong recommendations:

1.     NISQ regime (2+ years): Identify the computational pain-points that classical solutions most struggle with, e.g. in runtime or solution quality.

2.     Fault-tolerant QC regime (10+ years): Identify game-changing problems thought to be computationally impossible and revisit the industry-wide assumptions. 

With an understanding of these problems an engagement with the quantum companies and research labs could help determine the feasibility and requirement of quantum resources for a foray into this field. He also talked about IBM's Quantum tools that abstracts away the Quantum and makes the capability available for real world problems. 

Presentation 3: Quantum Sensing: Atomionics 

I met Sahil Tapiawala, the founder of Atomionics at an Entrepreneur’s First event sometime back and I was impressed by his evolution from then. The presentation was about Quantum sensing, and Sahil exuded confidence and spoke very clearly about what his company Atomionics was attempting to do. The talk evoked the largest number of questions and I was impressed by the poise with which some of those tough questions were handled. 

Sahil claimed that this company was working on Quantum Sensing devices that enables measurements of fundamental parameters at a precisions that would many orders more precise that the solutions available today. Eg he claimed that the device could measure gravity 1000X, acceleration 3500X and rotation 3000X more precisely. However one of the audience members later had a question about what this precision meant in terms of accuracy and sensitivity and the discussion around Accelerometer and Gravimeter became a bit technical, though the sense I got is that the solution isn't something conceptually radical and it’s the application of the Quantum approaches that is causing the improvements in precision.

The device could potentially disrupt multiple markets and he predicted that the Global Positioning Systems could in future benefit from a gravity based navigation and positioning system. This could potentially be useful underwater and in places with weak signals.

Atomionics's primary focus is on Hydrocarbon and Mineral Exploration and he claimed that a device pilot launch is aimed for Q4 of 2019. The current resource exploration is a trial and error exercise and sees nearly a 75% failure rate. A more accurate 3D density profile of the Earth's Crust, made possible by Atomionics device would enable exactly pinpointing oil and gas reserves with greater precision. 

In one of the slides Sahil explained that "Cold Atom Interferometry"was at the core of Atomionics Quantum based approach. For tech illiterates like myself this means that the atoms are cooled to a near absolute zero temperature, which almost brings them to a stop. The theory is to then exploit the quantum mechanical properties of an atom and measure the forces acting on the atom very precisely. 

Panelists for the Panel discussion

·      Dr. Yvonne Yuan Gao,Scientist, A*Star (Moderator)

·      Ewan Munro,CTO, Entropica Labs

·      Sahil Tapiawala,Co-founder and CEO, Atomionics Pte. Ltd.

·      Dr. Joe Fitzsimons,Founder, Horizon Quantum Computing and Principal Investigator, Centre for Quantum Technologies

·      Assoc Prof. Rainer Dumke,Centre for Quantum Technologies, NUS, NTU and InnovatusQ

·      Christian Raetzsch,Vice President, CTO and Quantum Ambassador IBM Systems, Asia Pacific

Where will Quantum be used and industry application issues

The panel felt that the Chemistry would see applications of Quantum Computing esp wrt new materials and drug research. In addition any problem that required mathematical optimization, AI and Machine Language could also potentially use Quantum Computing. An interesting insight was that Quantum Approaches could lead to new insights for classical computing mechanisms and devices. Further Quantum sensing could see applications in Oil & Gas and Autonomous Vehicular systems. 

Joe highlighted that tools are available for applications in computational fluid dynamics, but a key challenge was the availability of cross-functional expertise with Quantum and the respective domain knowledge. 

On a question on how to engage the industry, Christian made the oft-repeated point that it’s crucial to understand how QC is different. Further there is a need to retune existing computers to incorporate the QC approaches and doing new experiments that were considered computationally impossible. Christian highlighted how large companies like Boeing, Daimler and Samsung were thinking of using QC approaches to look at problems that aren't possible otherwise. 

Joe made a very interesting point that there are only about 200 people on the planet who have the skills to work with a Quantum Computer. However its possible to makes tools available that could abstract away the quantumness and make the quantum capabilities more broadly available to domain experts who could use it to solves problems of their respective industries. 

Dumke made some interesting points on the Hardware side of things but his accent made it challenging to grasp the full details. He did mention that some of the QC systems would be based on Superconductors and others on Atomic systems.

A point repeated throughout the session was how there was a need to keep the technology practical and the need to reduce the error rates. While we are early in the practicality question, postponing adoption may lead to a situation where we may miss the bus. Classical computers can never be able to simulate nature and there are many other problems that classical approaches cannot to do or do with significant challenges. While fault tolerant QC may be some distance away, Intermediate Quantum Computing could see wider application. 

Regulatory concerns

Regarding the regulatory concerns it was pointed that Shor's Algorithm can theoretically crack the existing encryption systems. However these algorithms are currently very sensitive to noise and perhaps sometimes in the future a fault tolerant computer could make it possible to crack open our current encryption setups. 

Joe mentioned that with advances in Genomics, it might be possible to match medicines with the needs of an individual. In this scenario it would be imperative to protect the genomic data of that individual for the lifetime and beyond of that individual. Classical Algorithm has been studied from quantum perspective only for a short period and many aspects with regulatory implications are not yet fully understood.

Munro commented that there might be implications from the territorial regulations that place geographical limitations around sensitive data. He speculated that Singapore then, might not be able to do any QC analysis on its data till it gets a QC capability within the boundaries of Singapore. To this Joe responded that there are protocols to apply QC approaches on encrypted data. 

Quantum Wars

The panel acknowledged the existence of the major quantum powers of EU, US, the Chinese and now the Australians. They expressed their concerns on how these developments could impact international collaboration. However the group also highlighted that these concerns are relevant to many technologies beyond Quantum technologies.

Conclusion

Quantum approaches are more than theory and its real. Its not merely a faster or a more powerful approach of computing but a ground-breaking new way of looking at things that works closer to the way nature works. There is a need to democratize and make this technology more accessible and engagement with the public. There is also a need to temper the expectations, as to what QC can and cannot do, for eg. NP complete problems will still not be solvable by the QC approaches and many of the day's big data problems could be solved by existing classical computing approaches.