Date: Thursday 7 and Friday 8 December 2023, from 9:30am
Venue: Nagoya University, Higashiyama Campus, Humanities and Social Sciences Building (文系総合館), 7th floor
Coordinates: 35.153838053176955, 136.96406005462265 (copy and paste these in your favourite map application)
Joonwoo Bae (KAIST), Francesco Buscemi (Nagoya U), Woonsang Choi (KAIST), Michele Dall’Arno (Toyohashi Institute of Tech), Kieran Flatt (KAIST), Haiyu Guan (Nagoya U), Donghoon Ha (Kyung Hee University), Jinhyuk Heo (Kyung Hee University), Hyunseong Jang (KAIST), Kohtaro Kato (Nagoya Uni), Spiros Kechrimparis (KIAS), Jaemin Kim (KAIST), Jeong San Kim (Kyung Hee University), Kodai Kobayashi (Nagoya U), Jiaxi Kuang (Nagoya U), Se Kang Kwon (Kyung Hee University), Hyeokjea Kwon (KAIST), Yonghae Lee (Kangwon National Univ.), Hanwool Lee (KAIST), Hojun Lee (KAIST), Shintaro Minagawa (Nagoya Uni), Karthik Mohan (KAIST), Mio Murao (Tokyo University), Teruaki Nagasawa (Nagoya Uni), Reiji Okada (Nagoya U), Jeonghoon Park (Kangwon National Univ.), Jongjun Park (Kyung Hee University), Ashutosh Rai (KAIST), Daisei Sakugawa (Nagoya U), Jiheon Seong (KAIST), Seungchan Seo (KAIST), Soyoung Shim (KAIST), Chikako Uchiyama (Yamanashi University), Eyuri Wakakuwa (Nagoya Uni), Jiyoung Yun (KAIST), Sung Won Yun (KAIST), Sang Woon Yun (SKKU)
Organizers: Joonwoo Bae (KAIST) and Francesco Buscemi (Nagoya U)
| time | Thursday 12/7 | Friday 12/8 |
|---|---|---|
| 9:30~10:15 | Jeong San Kim | Kohtaro Kato |
| 10:15~10:45 | Se Kang Kwon | Jiheon Seong |
| 10:45~11:15 | coffee break and discussion | coffee break and discussion |
| 11:15~12:00 | Michele Dall'Arno | Ashutosh Rai |
| 12:00~12:30 | Spiros Kechrimparis | Hanwool Lee |
| 12:30~14:00 | lunch break and discussion | lunch break and discussion |
| 14:00~14:30 | Jae Min Kim | Teruaki Nagasawa |
| 14:30~15:00 | Kieran Flatt | Shintaro Minagawa |
| 15:00~15:30 | Sung Won Yun | Karthik Mohan |
| 15:30~16:15 | coffee break and discussion | coffee break and discussion |
| 16:15~17:00 | Mio Murao | Yonghae Lee |
| 17:00~17:45 | Chikako Uchiyama | Eyuri Wakakuwa |
The social dinner is on the 7th, at the restaurant 鮮魚とおばんざい 我屋, starting from 6:30 p.m.
Title: Nonlocal quantum state ensembles and quantum data hiding
Abstract: We consider the discrimination of bipartite quantum states and establish a relation between nonlocal quantum state ensemble and quantum data hiding processing. Using a bound on optimal local discrimination of bipartite quantum states, we provide a sufficient condition for a bipartite quantum state ensemble to be used to construct a quantum data-hiding scheme. Our results are illustrated by examples in multidimensional bipartite quantum systems.
Title: Simulation of a general measurement on a circuit-based quantum computer
Abstract: We propose a protocol for simulating a general measurement on a circuit-based quantum computer. Based on Neumark’s theorem as well as the fact that every positive operator-valued measure (POVM) can be simulated by applying 2-outcome POVMs sequentially, we provide a measurement-operator decomposition algorithm to implement on a quantum computer. Moreover, we exhibit implementation results of a 3-outcome qutrit measurement on an IBMQ machine and IonQ Harmony.
Title: Bayesian inference of quantum devices
Abstract: We consider the scenario in which a black box with buttons and light bulbs is given so that, if any button is pressed a certain number of times, the corresponding probability distribution on the light bulbs lighting up can be observed. We model the black box as a prepare-and-measure setup, that is, an unspecified state is prepared upon the pressure of any button, and an unspecified measurement is performed on such a state. We consider the problem of the Bayesian inference of, say, the measurement (but, of course, we could consider the dual problem of inferring the states), that is, we aim at finding the measurement that maximizes the Bayesian posterior probability density, given the observations, for any given prior probability density on states and measurements. Our main result is to characterize such optimal measurements in the informationally complete (IC) case when uniform probability densities (i.e. maximal ignorance) are assumed on states and measurements, as it is natural in the first iteration of the Bayesian inference. In particular, we prove that any measurement that produces the observations upon the input of a 2-design set of states is optimal, thus settling in closed-form the case of non-overcomplete measurements, for which the only 2-design is the symmetric, informationally complete (SIC) set of states. Being data-driven, the inferential setup we consider offers a solution to the chicken-or-egg problem of usual quantum tomography, that is, the fact that the tomography of a measurement requires the knowledge of the input states, whereas the tomography of the states requires the knowledge of the measurement, in a never ending loop.
Slides: PDF
Title: Causal Asymmetry of Classical and Quantum Agents
Abstract: Why is it easier to degrade a clock than to make it more accurate? Here, we formalize this phenomenon by introducing process causal asymmetry - a fundamental difference in the amount of past information an agent must track to transform one stochastic process to another over an agent that transforms in the opposite direction. We then illustrate that this asymmetry can paradoxically be reversed when agents possess a quantum memory. Thus, the spontaneous direction in which processes get 'simpler' may be different, depending on whether quantum information processing is allowed or not.
Title: “Advanced” fully-quantum learning: Higher-order quantum algorithm for inversion of an unknown unitary operation
Abstract: Quantum learning is learning properties of unknown quantum objects (states and operations) by connecting the objects to a quantum computer. Fully-quantum learning is directly calculating a specific property of a quantum object by a quantum computer without extracting unnecessary classical information about the quantum objects and formalized by a higher-order quantum transformation (a quantum supermap). We consider an “advanced” version of fully-quantum learning, which “learns” an output of a function of an unknown quantum operation inside a quantum computer and applies the output on an arbitrary input state. We present a higher-order quantum algorithm to implement such advanced fully-quantum learning for inverting an arbitrary qubit unitary operation deterministically and exactly with four calls of the qubit unitary operation. This algorithm is catalytic, using one of the calls saved in a quantum state, which can be fully retrieved at the end of the algorithm. With one extra call, we can make the algorithm a “clean” protocol that does not leak information about the input unitary operation to the environment. This clean protocol allows unitary inversion of a two-qubit controlled unitary operation.
Reference: S. Yoshida, A. Soeda and M. Murao, Phys. Rev. Lett. 131, 120602 (2023)
Title: Environmental engineering for quantum transport
Abstract: The quick development of quantum technology has shortened the scale of measurements, approaching the correlation time of the environmental variables surrounding the system of interest. With the development, new research avenues have been opened for environmental engineering. We study the possibility of improving the efficiency of quantum transport[1] and work extraction of a quantum heat engine[2] by utilizing characteristic features of the environment, such as spatial-time correlation and spectral density. We also show a formulation to describe the reduced dynamics of quantum interacting systems that enables us to extend the applicabilities on the parameter setting as Bohr frequencies of subsystems, to preserve positivity in the short time region, and to describe the dynamics approaching the thermodynamically consistent stationary state, even if only the terminal site interacts with its environment[3].
References:
[1]C. Uchiyama, W. J. Munro, and K. Nemoto, npj Quantum Information 4,33 (2018).
[2]Y. Shirai, K. Hashimoto, R. Tezuka, C. Uchiyama, and N. Hatano, Phys. Rev. Res. 3, 023078 (2021).
[3]C. Uchiyama, Phys. Rev. A 108, 042212 (2023).
Title: Detecting Entanglement by State Preparation and a Fixed Measurement
Abstract: It is shown that a fixed measurement setting, e.g., a measurement in the computational basis, can detect all entangled states by preparing multipartite quantum states, called network states. We present network states for both cases to construct decomposable entanglement witnesses (EWs) equivalent to the partial transpose criteria and also non-decomposable EWs that detect undistillable entangled states beyond the partial transpose criteria. Entanglement detection by state preparation can be extended to multipartite states such as graph states, a resource for measurement-based quantum computing. Our results readily apply to a realistic scenario, for instance, an array of superconducting qubits. neutral atoms, or photons, in which the preparation of a multipartite state and a fixed measurement are experimentally feasible.
Slides: PDF
Title: Contextual advantages of maximum confidence measurements
Abstract: Contextuality is a property of quantum theories which is not present in classical ones. This allows noncontextuality to be used as a notion of classicality, and thus establishes advantages for quantum theory in certain information processing tasks. Recently, a number of works have constructed 'noncontextual inequalities' for various protocols, demonstrating that quantum advantages can be observed even for measurements on single qubits. These protocols have often involved state discrimination. In this talk, I will discuss our recent works building witnesses for contextuality based upon maximum confidence measurements. The latter are a class of state discrimination tasks more amenable to realistic performance than other techniques. It is hoped that our results will allow for experimental demonstrations of quantum contextuality.
Slides: PDF
Title: Parametrized versus Exact Quantum Circuits for Optimal Measurements
Abstract: In this work, we consider the construction of a parameterized quantum circuit that can be used to realize measurements for an ensemble of qubit states and show its variational training for optimal discrimination and maximal detection. In particular, an optimal quantum circuit is sought for a measurement that maximizes confidence, called maximum confidence measurements, which take detected events only into account. Variational training with a parameterized quantum circuit also implements a measurement that maximizes detection events. Our results are readily applied to practical and photon-based quantum information applications where measurement is noisy and lossy.
Slides: PDF
Title: Exact and Local Compression of Quantum Bipartite States
Abstract: Quantum data compression is one of the most fundamental quantum information processing. We study an exact local compression of a quantum bipartite state; that is, exact and noiseless one-shot quantum data compression of general mixed state sources without side information or entanglement assistance. We provide a formula for computing the minimal achievable compression dimensions, provided as a minimization of the Schmidt rank of a particular pure state constructed from that state. We will then discuss a possible application to tensor-network states.
Title: Detecting Entanglement-Generating Circuits in Cloud-Based Quantum Computing
Abstract: Entanglement, which is a direct consequence of elementary quantum gates such as controlled-NOT and Toffoli, is a key resource that provides quantum advantages. In this work, we establish a framework for certifying entanglement generation in cloud-based quantum computing services and present the construction of quantum circuits that certify entanglement generation in a circuit-based quantum computing model. The framework relaxes the assumption of qubit allocation, which, in cloud services, relates the physical qubits in hardware to a circuit proposed by a user. Consequently, certification is valid in cloud computing regardless of the success or trustworthiness of qubit allocation. The certification of entanglement generation was demonstrated on 2 and 3 qubits in the IBMQ and IonQ services. Remarkably, entanglement generation was successfully certified in the IonQ service, which does not offer manual qubit allocation. The capabilities of entanglement generation in IBMQ and IonQ circuits were also quantified. We envisage the application of the proposed framework in cloud-based quantum computing services for practical computation and information tasks, with the results determining whether it is possible to achieve quantum advantages.
Slides: PDF
Title: Nonlocal and Quantum advantages in network coding for multiple access channels
Abstract: We consider two-sender and single receiver discrete memoryless multiple access channels. We assume there is no feedback from the receiver to the senders. The input and output of the channel are classical information. Senders are space-like separated (i.e., there is no communication link between senders). Suppose senders share some entangled quantum state, then we study enhancement in reliable rates of transmission due to entanglement assistance to the considered channels. We show that entangled states shared between the senders can lead to better block codes to multiple copies of a channel and achievable rates of transmission can be improved by sharing entangled states between the senders. Considered channels are modeled on nonlocal games where output from the channel depends on whether inputs to the channel satisfy winning conditions in such games.
Slides: PDF
Title: Maximum confidence measurement for qubit states and its certification
Abstract: In quantum state discrimination, one aims to guess a quantum state in a given ensemble. Confidence, which is defined as a conditional probability that a given measurement outcome correctly concludes a state preparation, is a figure of merit that unifies different state discrimination strategies such as minimum-error discrimination and unambiguous discrimination. Maximum confidence measurements (MCMs) maximize confidence for all measurement outcomes, and they provide the highest guessing probability per detection event. In this research, we approach the problem of MCMs with a technique in convex optimization. We first find the optimality conditions of an MCM and show that MCMs for qubit states can be obtained by analyzing the geometry of the ensemble. We apply this method to several ensembles, such as geometrically uniform states and noisy SIC states. We then consider MCMs in a semi-device-independent scenario where the measurement device is completely untrusted and includes undetected events, while the preparation device is well-characterized. We show that MCMs can be certified from the outcome statistics only. Our results can be used to certify measurement devices in a realistic state discrimination scenario.
Slides: PDF
Title: Macroscopic state and entropy
Abstract: To show the irreversibility of the measurement process, von Neumann derived the microscopic entropy (von Neumann entropy) and showed that it does not decrease with measurement (entropy increasing law). However, such an entropy increasing law is quite different from the one by the classical physics. To fill this gap, von Neumann introduced macroscopic entropy (observational entropy) and showed that it changes with time evolution (consistent with classical physics). We study the relationship between macroscopic and microscopic entropy. We have derived new conditions for quantum states (macroscopic states) and measurements such that macroscopic and microscopic entropies coincide.
Title: Thermodynamics of quantum information processing: measurement, feedback control and erasure
Abstract: Maxwell’s demon is a problem that feedback control performed based on information obtained by measurement appears to violate the second law of thermodynamics. Since control techniques for microscopic systems such as quantum systems have improved and are being used for information processing, this problem is now beyond the foundations of statistical thermodynamics and is connected to a large direction in modern physics: the relationship between information and physics. The topic of this presentation is the thermodynamics of a protocol consisting of quantum information processing such as measurement, feedback control, and erasure, as a model of Maxwell’s demon. Based on our recent results [1], we derive universally applicable work formulas for this protocol and provide consistency conditions between the protocol and the second law of thermodynamics.
Reference:
[1] S. Minagawa, M. H. Mohammady, K. Sakai, K. Kohtaro and F. Buscemi, Universal validity of the second law of information thermodynamics, arXiv: 2308.15558 (2023)
Title: Distributing Entanglement over Separable Quantum Networks
Abstract: We consider distributing entanglement over a separable quantum network, i.e., multipartite states prepared by local operations and classical communication without quantum communication. We investigate the distribution of entanglement over a separable network with an intermediate system called a carrier, such that the carrier always shares no entanglement with a network during the distribution protocol. Relations between a separable quantum network and a permanently separable carrier for creating an entangled network are characterized and applied to networks at various structures, such as linear and circular chains. The results can be used to establish a secure entangled network.
Slides: PDF
Title: Modified phase estimation algorithm and its application
Abstract: In this talk, we first introduce the quantum phase estimation algorithm (QPEA). From the QPEA, one can estimate bit information on the eigenvalues of a given Hermitian operator, i.e., one can obtain the first n bits of the eigenvalues. Then, we generalize the QPEA by adding some controlled gates into its original circuit. We call this variation a modified phase estimation algorithm (MPEA). For a given i, the MPEA allows us to obtain from the (i)-th bit to the (i+n)-th bit of the eigenvalues. The main idea of the MPEA is to apply the left shift operation (or concept) for bit strings to the eigenvalue bits. As an application of this result, we suggest a circuit implementation method to improve the performance of the HHL algorithm under the NISQ era.
Slides: PDF
Title: Exact Exponent for Atypicality of Random Quantum States
Abstract: We study properties of the random quantum states induced from the uniformly random pure states on a bipartite quantum system by taking the partial trace over the larger subsystem. Most of the previous approaches have focused on the behavior of the states close to the average, as known under the name of measure concentration. In contrast, we investigate the large deviation regime, where the states may be far from the average.
We prove the following results: First, the probability that the induced random state is within a given set obeys the large deviation principle, i.e., it decreases no slower or faster than exponential in the dimension of the system traced out. Second, the exponent is equal to the quantum relative entropy of the maximally mixed state and the given set, multiplied by the dimension of the remaining system. Third, the total probability of a given set strongly concentrates around the element closest to the maximally mixed state, a property that we call conditional concentration. Along the same line, we also investigate an asymptotic behavior of coherence of random pure states in a single system with large dimension.