1 May 2019 14:00—16:30
EWI lecture hall Chip
Electromagnetic Modeling for Solid State Quantum Computers
A present view of a quantum computer is that it is a large assembly of extremely high quality factor microwave resonators, of varying degrees of anharmonicity (large anharmonicity=qubit), coupled into a two dimensional network. We can imagine a detailed high frequency simulation of the network -- but what do we extract from these simulations (difficult due to the high Q). I will review some approaches that we have laid out for doing these simulations, along with some handy formulas for relating the results of simulations to quantum computer properties.
David DiVincenzo (Philadelphia, VS, 1959) has a secondary appointment as professor at the EEMCS Department at the TU Delft and a staff member at QuTech since 2017. His primary appointment is as Director, Theoretical Nanoelectronics, Forschungszentrum Jülich (Germany). He received his doctorate at the University of Pennsylvania, Philadelphia, USA in 1983; was a postdoc at Cornell University, Ithaca, USA; then Research Staff Member at IBM Watson Research Center, Yorktown Heights, New York, USA (1985-2011). Having been granted an Alexander von Humboldt Professorship in 2010, he became a professor at the Institute of Theoretical Quantum Information at RWTH Aachen University and director of the Peter Grünberg Institute, where he serves to the present. He is a Fellow of the American Physical Society (1999), and Associate Editor of the Reviews of Modern Physics (2011-present).
David DiVincenzo was one of the first physicists to engage in quantum information research and is considered an authority on quantum information processing. In particular, his name is associated with the development of criteria for the quantum computer, known as the DiVincenzo Criteria, and also with the Loss-DiVincenzo approach to solid-state spin-based qubits.
M. Fernando Gonzalez Galba
A dynamic random-access architecture based on FDSOI technology for radio-frequency readout of quantum devices
Quantum computing is maturing at a relentless pace, yet individual quantum bits are wired one by one. As quantum processors become more complex, they will require efficient interfaces to deliver signals for control and readout while keeping the number of inputs manageable. Digital electronics could offer solutions to the scaling challenge by leveraging established industrial infrastructure to interface silicon-based quantum devices with conventional CMOS circuits. Here, we combine both technologies at milikelvin temperatures and demonstrate the building blocks of a dynamic random-access architecture for efficient readout of complex quantum circuits. Our results demonstrate a path to reducing the number of input lines per qubit and enable addressing of large arrays of devices.
Fernando is Senior Research Scientist and Head of Quantum Information at the Hitachi Cambridge Laboratory, the R&D centre of Hitachi Europe Ltd. on fundamental device physics. His research is focused on new computing paradigms and more recently on the development of a silicon-based quantum computer.
PhD from the University of Cambridge with a thesis on Single-atom Electronics in 2013, he received the R&D Technology Award from Hitachi's Centre for Social Innovation for the development of the aforementioned technology in 2016. He was awarded the Young Scientist Award by the Spanish Royal Society of Physics in 2017 and became a Royal Society Industry Fellow in 2019.
His research is funded by the European Commission’s H2020 program, EPSRC and the Winton Programme for the Physics of Sustainability. He has published more than 30 peer reviewed articles, has 3 granted patents and often participates in outreach scientific events for all ages.
Simulation based design of qubit control – and future quantum computers?
For classical electronics, a tool chain for simulation and design is an essential part of the ecosystem. In quantum computing research, this role is currently filled by home-grown insular solutions and the exchange between theory and experiment. Is it time to start developing a similar tool chain for the development of quantum computers?
I will explore this vision based on our experimental and simulation work on the control of two-electron spin qubits in GaAs quantum dots. Starting from a numerical optimization of control pulses  followed by fine-tuning based on characterization measurements, we have realized single-qubit gate fidelities of 99.5 %, the current state of the art for this system . The measured values are in good agreement with the simulations. We predict similar fidelities for two-qubit gates based on the same models and optimization procedures. Furthermore, we have carried out a detailed analysis of the impact of residual coupling on single-qubit gates in a two-qubit device, thus designing a complete two-qubit gate-set.
I propose that such realistic simulations will be useful for developing and evaluating architectures for future quantum computers.
Hendrik Bluhm is a Professor in Experimental Physics and Institute leader at RWTH Aachen University and Forschungszentrum Jülich. His main research interest is the advancement of semiconductor spin qubits. He won an ERC Starting Grant and the Alfried Krupp Prize for Young Professors.
Deep Reinforcement Learning for Coherent Transport of Spin Qubits
The diffusion of deep learning algorithms has boosted the research in several fields. The paradigm shift from knowledge-based to representation-based artificial intelligence has opened the chance to apply novel methods to physics. I review quantum computer architectures  and I show how to improve quantum computers by exploiting deep reinforcement learning . I present two practical examples of how to steer a qubit by exploiting deep reinforcement learning, namely in the case of spatial coherent transport by adiabatic passage (CTAP)  of quantum states  across an array of quantum dots  and in the field of quantum compiling, managed by using A2C and TRPO deep reinforcement learning algorithms. By reverse engineering the network, it is possible to achieve better understanding of the physical process itself by identifying those physical quantities more contributing to the process.
 D. Rotta, F. Sebastiano, E. Charbon, and E. Prati, "Quantum information density scaling and qubit operation time constraints of cmos silicon-based quantum computer architectures," npj Quantum Information, vol. 3, no. 1, p. 26, 2017.
 A. Bonarini, C. Caccia, A. Lazaric, and M. Restelli, "Batch reinforcement learning for controlling a mobile wheeled pendulum robot," in IFIP International Conference on Artificial Intelligence in Theory and Practice, pp. 151-160, Springer, 2008.
 E. Ferraro, M. D. Michielis, M. Fanciulli, and E. Prati, "Coherent tunneling by adiabatic passage of an exchange-only spin qubit in a double quantum dot chain," Phys. Rev. B, vol. 91, p. 075435., 2015.
 R. Porotti, D. Tamascelli, M. Restelli, and E. Prati, Coherent Transport of Quantum States by Deep Reinforcement Learning, arXiv:submit/2546122 [quant-ph] 20 Jan 2019
 E. Prati, M. Hori, F. Guagliardo, G. Ferrari, and T. Shinada, Anderson-Mott transition in arrays of a few dopant atoms in a silicon transistor, Nature Nanotechnology 7, 443-447 (2012)
Enrico Prati (1974) is research scientist of Italian National Research Council (CNR) at IFN in Milano and Adjunct Professor at Politecnico di Milano for the course "Quantum Artificial Intelligence". He has been working on silicon quantum devices and single atom nanoelectronics, architectures of silicon quantum computers, and he is currently working on deep learning methods applied to quantum information processing. He is PI of the project QUASIX on single photon emission in silicon funded by the Italian Space Agency and in charge of machine learning applied to quantum processes in the H2020 Project Narciso. He published more than 80 papers and he is regularly invited speaker of main international workshops. He has been Keynote speaker at IEDM in 2014, TEDx speaker in 2016 and he is author of the dissemination book Mente Artificiale (2017, in italian - transl. Artificial Mind).