Quantum Realism: A Realistic Future for Quantum Computing [CALL FOR PAPERS]
摘要截稿:
全文截稿: 2018-12-01
影响因子: 4.419
期刊难度:
CCF分类: 无
中科院JCR分区:
• 大类 : 计算机科学 - 4区
• 小类 : 计算机:硬件 - 4区
• 小类 : 计算机:软件工程 - 4区
Overview
Quantum computing is on a path of accelerating public interest due to the hope that new principles in quantum information will impact society as much as the information revolution and Moore’s law. While there is reasonable basis for quantum computing being important, current public expectations have moved far beyond reality.
In this special issue of Computer, the guest editor seeks articles that identify and explain the balance between the possibilities and the reality, as well as the limitations of various aspects of quantum computing. Articles should explain technical issues for Computer’s broad readership, and how these issues apply to the arguments of advocates and skeptics.
The authors can express their opinions, but the objective of this special issue is to allow readers to learn about the arguments from all sides, so they can choose among reasonable alternative views or have a basis from which to form their own opinion. Some general potential topics are described below, with authors invited to cover specific issues within an area:
Better solutions to current problems. What problems will quantum computers solve better than classical computers, and how much will it matter to society? There is a well-studied algorithm for factoring numbers that has implications to encryption and there are quantum computer algorithms for optimization that could find solutions closer to the global optimum than any algorithm on a classical computer. Articles can address the degree to which society may be changed by the relative improvement offered by shifting from a classical to quantum computer. For example, could a stockbroker using a quantum computer become more successful than one just using a classical one by pricing stocks more accurately?
New applications. Quantum computers are believed capable of solving some problems that are intractable for today’s computers. Some such problems are not currently considered important, thus it seems obvious that there will be no big profitable companies selling solutions to that problem. However, if a quantum computer can solve a known but previously intractable problem, could companies emerge that sell new, useful products and then grow to be big and profitable?
Manufacturing. Quantum computers have been demonstrated to the level of 50–2,000 qubits, creating a tantalizing parallel with early electronic computers with 50–2,000 relays or vacuum tubes. In this parallel, Moore’s law and other effects boosted the number of active components by a billion-fold in less than a century. However, there is only one example of such a large improvement factor, and the improvement didn’t come for “free”—there was heavy investment in the development of current semiconductors. Assuming quantum computer physics is sound, what would be required to move research demonstrations to commercial products?
Quantum computer software engineering. Early languages for both classical and quantum computers merely provide bookkeeping assistance for controlling the underlying hardware, such as assembly language for classical computers and gate sequences for quantum computers. Programming languages for classical computers evolved to embrace higher programmer productivity through, for example, object orientation, domain specificity, and group programming methods. To achieve quantum speedup, quantum algorithms must invoke one of several uniquely quantum features, such as qubit phase, interference, entanglement, and so forth. How will quantum computer software impact toolchains that improve programmer productivity for the uniquely quantum features needed to achieve quantum speedup?
Quantum machine learning and artificial intelligence. There is both a theoretical basis and experimental evidence that quantum computers can learn training sets more efficiently than classical computers. Taken at face value, this would allow quantum computers to learn from self-driving car training sets more quickly than is currently possible with a GPU. However, faster learning will not address the issue that self-driving cars get into accidents that seem due to the limitation of their “intelligence” vis-a-vis pattern recognition. What are the lessons a quantum computer might learn—that enables progress in machine learning and AI—that a classical computer cannot? In one example of this, quantum computers can learn by tunneling through barriers in an optimization space, as opposed to just moving downhill, but what types of real-world lessons are more easily learned by tunneling as opposed to descent?