Bose-Einstein condensate (BEC) is a versatile experimental platform on which fundamental tests of quantum physics are performed and applications to quantum informatics are realized. Among others, characterization of nonclassical properties particularly quantum non-Gaussianity (QNG) & quantum entanglement (QE) is of crucial importance but still posing major challenges. In this talk, we briefly discuss the recent attempts of demonstrating quantum nature of gravity using BEC that has a bearing on  QNG & QE and show how the superselection rule can be useful to derive an entanglement criterion for BEC

양자검색 알고리즘은 양자중첩 원리를 사용하여 질의복잡도(query complexity) 기준 이차감소 양자이득을 보인다. 하지만, 양자검색 알고리즘에서 양자중첩 형태의 데이터에 액세스하고 이를 처리하는 방법론 자체에 대한 아키텍처는 현재까지 구체적으로 연구된 바가 거의 없다. 실제로, 기존 양자검색 알고리즘은 블랙박스 연산—오라클 연산—에 의해 준비되고 액세스된다고 “가정”되어 있을 뿐이다. 하지만, 이러한 아키텍쳐가 적절히 설계되지 않으면 양자검색 알고리즘의 양자이득은 단순히 계산과학적 이점에 불과하며 현실화 가능하지 않을 수 있다. 이에, 본 연구에서는 효율적 양자데이터 액세스 프로세스인, 소위 “양자데이터 액세스 머신(Quantum Data-access Machine; QDAM)”을 정의/소개하고 양자검색 알고리즘을 위한 일반적인 아키텍처를 제시한다. 본 연구에서 제안한 알고리즘의 런타임은 결함허용 양자컴퓨팅 관점에서 분석한다. 보다 구체적으로, T-게이트 깊이를 분석한다. 분석을 통해, N개의 데이터 검색 문제에 대해 O(logN) 스케일의 T-게이트 깊이 복잡도의 QDAM 모델 구성이 가능함을 보이고, 그 결과 본 연구에서 제안한 QDAM 형태를 포함한 양자검색은 총 O(sqrt(N) x logN) 런타임 비용을 나타냄을 증명한다.

참고논문: Quant. Sci. & Tech. 9, 015011 (2023)

Quantum metrology can achieve enhanced sensitivity for estimating unknown parameters beyond the standard quantum limit. Recently, multiple-phase estimation exploiting quantum resources has attracted intensive interest for its applications in quantum imaging and sensor networks. In this seminar, I will present my recent results on simultaneous estimation of multiple-phases using entangled photons [1-3] and distributed quantum sensing of multiple-phases using a fewer number of photons [4]. In addition, I will introduce our ongoing research topics along quantum-enhanced multiple-phase estimation [5] and discuss our perspectives on this topic.

  A fundamental computational problem is to find a find a shortest non-zero vector in Euclidean lattices, a problem known as the Shortest Vector Problem (SVP). This problem is believed to be hard even on quantum computers and thus plays a pivotal role in post-quantum cryptography. In this talk, I will introduce some recent result on NISQ (Noisy Intermediate Scale Quantum) devices which may be used to solve SVP.

  Rydberg atom array is one of the most promising platforms in quantum information processing due to its scalability and long coherence time. Due to the resemblance between the Rydberg Hamiltonian and the maximum independent set problem’s cost function, the Rydberg atom arrays are emerging as new solvers for this paradigmatic NP-hard problem. In this talk, we suggest a strategy to approximate maximum independent sets on Rydberg atom arrays by adjusting local detunings according to vertex support, which quantifies connectivity between vertices. Our strategy reduces error rate three times for square lattice graphs with defects with sufficient adiabaticity. Also, our strategy achieves success for the random graphs, whether the adiabaticity is sufficient or not. Furthermore, we find that using vertex support as the measure of connectivity helps to prepare 2D cat state.

  Standard teleportation (ST) is a protocol for transmitting an unknown quantum state to a spatially separated receiver without the need to physically transmit qubits. ST requires a single pair of maximally entangled qubits, a classical communication channel and a quantum operation of the receiver. As a variant of quantum teleportation, port-based teleportation (PBT), has been introduced that does not require quantum correction by the receiver. The protocol is possible as extending to multi pairs of the entangled qubits. Accordingly, the joint POVM implemented by the sender must have a larger dimension measurement than the Bell measurement used in ST. The square-root measurement (SRM) makes PBT optimal, and perfect teleportation fidelity is asymptotically achieved for an infinite number of pairs or ports. We generalize the PBT protocol by increasing possible measurement outcomes from each port, while maintaining the SRM that produces optimal teleportation. We connect the relationship between ST and PBT through port-based generalized measurement teleportation (PBGMT) and present a new class of teleportation protocols that exist between them. We anticipate PBGMT as a fresh approach to discovering implementable teleportation protocols.

  In this talk, we consider the ranks of the marginal states of a tripartite pure state. By comparing the ranks, we present the relations between ppt marginal states and separable marginal states. In other words, we show that if Alice, Bob, and Charlie share a tripartite pure state |ψ_ABC⟩, and if two of its three marginal states ρ_AB, ρ_AC, and ρ_BC are ppt states, then their ranks are the same, and hence both are separable. We also show that if the ranks of the two marginal states are the same, then either at least two of the three ones are separable or GHZ-distillable.

  Estimating fundamental properties of quantum infomation theory such as von Neumann, Renyi, tsallis, quantum relative entropy, trace distance, fidelity have gained significant interest. Many algorithms has been proposed to estimate each fundamental properties, but there are no unified way to estimate all the fundamental properties. Our work propose a unified way to estimate fundamental properties by using layerwise quantum convolutional neural networks(LQCNN). Recently, some works have showed parameterized quantum circuit can be used to estimate fundamental properties, but it suffer barren plateau and due to high complexity it can't be used in large qubit states. We argue that our work avoid barren plateau and it's a feasible way for large qubit states. Our first result states the mathematical proof that LQCNN structure preserves the fundamental properties. Our second result analyze the complexity of our algorithm and show that it avoids barren plateau due to it's local cost function. Finally, we perform numerical simulations to verify our analytic results.

  벨 부등식은 양자 얽힘을 검증하기 위한 가장 쉬우면서도 효율적인 방법이다. 본 발표에서는 기존의 벨 부등식을 기반으로 한 양자 얽힘 검증이 가지는 한계점을 살펴보고, 이를 극복하기 위한 새로운 방법으로 비파괴적 양자상태 구별을 제시한다. 이 방식은 기존의 방법에 비해 보다 큰 고전/양자 대비를 가지며, 고전적인 통신이 허용된 상태에서도 양자 얽힘을 검증할 수 있다는 장점을 지닌다.

  Quantum illumination is on the purpose of discriminating the presence or absence of a low-reflectivity target using entangled states. Here I introduce detection bounds for quantum illumination with Gaussian states.

  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.

  The smooth quintic del Pezzo variety Y is well-known to be obtained as a linear sections of the Grassmannian variety Gr(2,5) under the Plücker embedding into P^9. Through a local computation, we show the Hilbert scheme of conics in Y for dim Y ≥3 can be obtained from a certain Grassmannian bundle by a single blowing up/down transformation.

  Quantum coherence and entanglement are fundamental quantum resources. It has been shown that these two properties of quantum states are closely related. Recent results from dynamic resource theory show that the properties of quantum channels, which affect the two properties of quantum states, also have an analogous relationship. Here we show that the properties of quantum channels that affect the quantum resources of quantum measurements also have a similar relationship.

  Quantum computers that are currently under development mostly have cloud computing environments, which is why quantum homomorphic encryption (QHE) could be a useful application in the future. As quantum error correction is essential part of quantum computation, it is also important to devise the method to simultaneously exploit both QHE and quantum error correction codes (QECCs). We present an improved QHE scheme which provides higher security with fewer resources and has lower computational overhead in comparison to the previous work adopting a concatenated code. In addition, using additional resources, both security and error correction capabilities can be improved simultaneously.

  In this work, we show that any noisy quantum channel on qubit systems can make the zero-error classical capacity additive. However, we present that a noiseless qubit channel can generate nonadditivity even if the other quantum channel has no channel capacity. We call this phenomenon the activation of the channel capacity. Moreover, we focus on properties when the zero-error classical capacity is activated.

  We discuss metrics on the space of quantum states that are monotone under CPTP maps. Such metrics naturally arise from infinitesimal values of various types of relative entropies. We focus on these metrics especially on the space of Gaussian states. As Gaussian states are uniquely determined through their first- and second-order moments, we show how these monotone metrics are also determined through the first- and second-order moments.

  2014년 Farhi, Goldstone, Gutmann이 제안한 Quantum Approximate Optimization Algorithm(QAOA)을 간단히 소개합니다. QAOA는 양자 컴퓨터와 고전 컴퓨터의 연산으로 approximate solution을 구할 수 있는 알고리즘입니다. 양자 컴퓨터에서 문제의 objective function을 양자 상태에 대응하여 solution을 찾을 수 있도록 연산하고, 고전 컴퓨터에서 parameter optimization을 통해 양자 컴퓨터에서 solution을 양자 상태로 잘 측정할 수 있도록 합니다. QAOA는 양자 회로에서 연산 단계가 존재합니다. QAOA의 단계가 높아지면 solution의 측정 확률이 증가합니다. 그런데 QAOA는 NISQ 시대의 양자컴퓨터에서 사용하기에 적합한 알고리즘이지만 QAOA의 단계가 높아지면 복잡도가 증가하는 단점이 있습니다. 이러한 단점을 보완하고자 QAOA의 양자 회로를 변형하여 낮은 양자 회로 복잡도에서 높은 성능을 갖는 알고리즘을 소개합니다.

  In the era of NISQ computing, quantum machine learning is a promising field. We focus on quantum generative adversarial networks, which have applications in fields like image generation, finance, and probability distribution modeling. However, like (Q)GANs [1], these networks require solutions to inherent challenges such as mode collapse. Our study introduces InfoQGAN [2], a novel approach that uses Mutual Information Neural Estimator (MINE) within quantum generative adversarial networks to address mode collapse [3]. Furthermore, similar to InfoGAN, we demonstrated in InfoQGAN the ability to manipulate the generated distribution by adjusting the input latent codes. We also demonstrate its practicality in financial scenarios, specifically for generating portfolio return distributions through dynamic asset allocation. This highlights the potential real-world applications of InfoQGAN. Our goal is to demonstrate that by utilizing InfoQGAN, the modeling performance is significantly enhanced compared to using QGAN alone.