Registration

Opening

Tensor Networks at the Interface of Machine Learning and Quantum Computing

• Synge Todo (University of Tokyo)

Tensor networks connect quantum many-body physics, machine learning, and quantum computing through compact representations of states, probability distributions, and learning models. In this talk, we explore how these representations support learning and sampling, and how they can mediate between classical computation and quantum circuits. We discuss tensor-network approaches to generative modeling, including Born machines and sampling from matrix product states, before turning to the embedding of classically trained models into quantum circuits. Particular attention is given to quantum state preparation, the overhead of postselection, and adiabatic encoding strategies for constructing quantum classifiers. Learning from quantum data is illustrated through tensor-network tomography and Bayesian inference of noise models from error-correction syndromes. We also introduce tensor-network Monte Carlo as an approach to combining compression with stochastic sampling for error control. These examples motivate a discussion of when quantum processing can add value to classical tensor-network methods. We conclude with prospects for workflows that integrate classical learning, quantum state preparation, and quantum processing, with attention to computational cost, trainability, and accuracy.

Efficient characterization and processing of quantum states in linear optical networks

• Jun-yi Wu (Tamkang University,)

Multiphoton interference in linear optical networks offers a versatile platform for generating high-dimensional entanglement and processing quantum information. Realizing the potential of these entangled states requires efficient methods for both their preparation and characterization. In this talk, we will discuss the generation of high-dimensional entanglement in linear optical networks and present a two-detector method for reconstructing multiphoton states with reduced detection hardware. Building on these capabilities, we will explore how photonic entanglement can serve as a resource for distributed quantum computing, facilitating quantum operations across separate photonic processors. The talk will highlight the connections between entanglement generation, efficient state characterization, and distributed quantum information processing.

Normalizing Flow-Based Quantum State Estimation

• Hsien-Yi Hsieh (NTHU)

We develop a normalizing flow-based Bayesian parameter estimation (BPE) framework to characterize noisy quantum states and uncover correlations among physical parameters. By incorporating physically interpretable parameters into quantum state ansatzes, our approach infers their posterior distributions directly from experimental data, enabling uncertainty quantification and correlation analysis. We demonstrate the framework using two physical ansatzes for optical cat-state experiments, where noisy single-photon states are added to impure squeezed states. By examining correlations among Wigner function negativity, photon-addition fraction, noise contributions, and pump powers, we gain insights into the experimental mechanisms underlying the generation of noisy non-Gaussian states. Combining Bayesian inference with physical and experimental constraints, our framework provides interpretable parameter estimates and practical guidance for optimizing desirable quantum features.

Coherent control of undoped Ge/SiGe quantum dot spin arrays

• Tzu-kan Hsiao (NTHU)

Electrostatically-defined quantum dot arrays are a natural and versatile platform for quantum computation and quantum simulations. In particular, spin qubit arrays based on undoped Ge/SiGe quantum wells have attracted tremendous research interest due to their long coherence time and strong spin-orbital coupling. In this talk, I will present our progress in the fabrication and measurement of Ge quantum dot spin qubit devices. First, we present a new design of Ge quantum dot devices with a global accumulation gate and depletion-style fine gates. The device fabrication consists of only three metal layers, two oxide layers, and needs only micrometer alignment accuracy. We measured these devices and successfully observed Coulomb blockade and Coulomb diamond both in DC transport and RF reflectometry. In addition, we used the Coulomb peaks to characterize the charge noise and obtained a noise power spectrum density of 1.1μeV/√Hz at 1Hz, confirming the quality of the material and device. Next, we measure the energy spectrum of two spins in a Ge double dot system. By initializing a singlet state and exciting it into triplet states using microwave burst, we map out the singlet-triplet energy spectrum as a function of external magnetic field. We also achieve coherent spin control with T_2^Rabi~10 μs and T_2^*~2 μs.

A geometric view of dynamics in quantum phase space

• Ole Steuernagle

Classical phase space dynamics is governed by a continuity equation. The same is true for quantum dynamics in phase space: Wigner's quantum phase space current, J, governs the evolution of Wigner's phase space distribution W. Unlike the classical case, W is typically negative in some regions. These negative regions represent quantum coherences. Also, the current's velocity field is ill-defined. It is singular when W = 0. I show that this implies that there are no trajectories in quantum phase space. With these velocity singularities Liouvillian phase space volumes feature singular changes too. I will show that this is necessary in order for quantum dynamics to create coherences. This is worth knowing particularly for numerical investigations. Stationary points of J are important for both, classical and quantum dynamics, maybe even more so in the quantum case, since the dynamics can move and split or merge these stagnation points. But the existence of stagnation points is constrained by a topological conservation law which quantum mechanics has to obey (see Figure). While velocity fields and trajectories are ill-defined in quantum phase space, I show that the current J is always well behaved. It can be studied, providing new insights, and it can be numerically integrated to study the dynamics.

Monitoring Beam Splitter-Assisted Entanglement using Quantumness

• Hua Li Chen

We generated two-mode entangled states by superposing two squeezed vacuum states from degenerate OPOs via a 50:50 beam splitter (90° phase shift), achieving ∆2(p1 - p2) = -6.14 dB and ∆2(x1 - x2) = -6.17 dB. From the measured covariance matrix, both PPT and EPR criteria were evaluated.However, these only allows us to witness the entanglement, but not to determine the deleterious effects of experimental imperfections due to the beam splitter mixing and the associated mode-mismatch and detection imperfections.We therefore introduce an alternative framework suitable for continuous variable systems using the states’ quantumness, Ξ. This allows us to benchmark the experiment’s performance by comparing the states’ quantumness Ξ′ after the beam splitter mixing with Ξ before. That is a feature unachievable by any standard entanglement criteria.

Free Discussion

Registration

Coherent erbium ions on photonic and superconducting chips

• Steven Touzard

Erbium ions in crystals combine an optical transition in the telecom band, where optical fibres have their lowest loss, with an electron-spin transition at gigahertz frequencies, where superconducting quantum circuits operate. This makes them a natural candidate for linking superconducting processors to optical quantum networks. In integrated devices, however, erbium coherence is usually limited by the host material. I will present an approach that avoids this limitation: bonding a bulk, coherence-optimised Er:CaWO₄ crystal directly onto a chip. On the optical side, bonding to an electro-optically tuneable thin-film lithium niobate microring resonator yields a collective cooperativity of 6.7 while preserving millisecond optical coherence. On the microwave side, bonding the same crystal to high-quality-factor niobium resonators allows microwave pulses to be stored and retrieved with about 70% efficiency and added noise close to the vacuum level. Along the way, I will introduce the physics of spin ensembles collectively coupled to resonators for a non-specialist audience.

Quantum technologies with quantum linear optics

• Juan Carlos Garcia Escartin (Universidad de Valladolid)

Optical systems give a flexible way to explore quantum mechanics. The quantum states of light are fundamental in quantum communications and, in quantum computation, they provide an alternative way to obtain quantum advantage without a full universal quantum computer. This talk will introduce the main applications of linear optical devices to quantum information explaining their capabilities and limitations. From a description in terms of Lie groups, we can study what can and cannot be done with linear systems and give implementation and optimization methods and invariants to determine the feasibility of certain transformations. With these techniques, we will review the use of linear systems for state preparation and in different primitives in quantum information like state comparison and symmetry tests.

Observation and Manipulation of Optical Parametric Down-Conversion in the Langevin Regime

• Chun-Yuan Cheng

We experimentally realized optical parametric down-conversion in the Langevin regime on a chip. By precisely controlling the loss inherently tied to fluctuations, we also observed the asymmetric Hong-Ou-Mandel dip—a hallmark of quantum fluctuation in the fluctuation-driven parametric down-conversion— and the fluctuation-induced compression of single photons. Our findings pave the way for manipulating quantum fluctuations, quantum states, and system-reservoir interactions.

Nonparametric Learning of Non-Gaussian Quantum States of Continuous-Variable Systems

• Liubov Markovich (Leiden University)

Continuous-variable quantum systems play an important role in quantum computation, communication, and sensing. Characterizing their quantum states from experimental measurements, however, remains a challenging task, particularly for non-Gaussian states, which often exhibit complex structures that cannot be described by a small number of parameters. In this talk, I will introduce a statistical approach to quantum state estimation based on tomographic probability distributions, which describe the outcomes of experimentally accessible measurements. Unlike conventional approaches that rely on assumptions about the underlying quantum state, we develop a nonparametric kernel quantum state estimation (KQSE) framework that allows us to reconstruct unknown quantum states directly from noisy measurement data. The proposed method provides estimates of the density matrix in different representations, as well as important physical quantities such as purity, higher moments, overlap, and trace distance. Moreover, it achieves a near-optimal convergence rate of (\tilde{O}(T^{-1})), where (T) denotes the total number of measurements. I will illustrate how this approach can be applied to non-Gaussian quantum states, including states with complex, multimodal probability distributions, and discuss its potential for quantum state characterization without prior knowledge of the state.

Free Discussion / Poster session

Registration

Atom as an Agent: Emergent Order by Collective Predictive Coding

• Koji Hashimoto (Kyoto University)

What is the intrinsic reason of emergence of macroscopic order parameter out of microscopic systems? In this article, we apply to physics the collective predictive coding (CPC), the theory of language/symbol emergence. As our language emerges out of our human communications, physical macroscopic order parameter emerges out of communications among atoms --- each atom is regarded as an agent who predicts each other's behavior through inference. We present three examples of classical spin systems: two-dimensional ferromagnetic and antiferromagnetic Ising models, a two-dimensional spin model with competing stripe orders, and the fully connected Sherrington--Kirkpatrick spin-glass model. In all of these examples, proper macroscopic order parameters including the replica overlap emerge automatically through communication informatics of CPC, only from observed spin configurations. We call the learned macroscopic coordinate the collective predictive order. Furthermore, for the ferromagnetic and antiferromagnetic Ising models, the equilibrium distribution of the collective predictive order yields an emergent potential with the symmetry and thermal evolution characteristic of a Landau potential. Notably, the emergence of the collective predictive order does not refer to any structure of microscopic Hamiltonians/Lagrangians. Thus our ``Atom-as-an-Agent" (AaaA) principle offers a completely data-driven method for the emergence of macroscopic order, that is, machine learning physics. This work is in collaboration with Tadahiro Taniguchi, and the preprint paper is available at https://arxiv.org/abs/2610.36673

Readout optimization of superconducting transmon qubits

• Chung-Ting Ke

Qubit readout is essential and fundamental to qubit fidelity. Currently, qubit readout still suffers from thermal excitation, transition, and control errors. Although active reset could ease part of the issues, it increases the operation overhead. To mitigate these issues, we explore the effect of package and readout measurement-induced state transitions (MIST) based on our superconducting transmon qubits. We first investigate the effect of the package on cooling our superconducting qubits. We developed a three-layer infrared (IR) shielding configuration to reduce the effect of thermal radiation and demonstrate that improved shielding suppresses the quasiparticle-induced parity-switching rate. The effective qubit temperature follows the phonon bath down to 17 mK, enabling initialization errors of ∼ 0.01% for 3 GHz qubits without any reset metrics. In the second work, we systematically investigate ng-dependent MIST dynamics governed by the diabaticity and symmetry of pulse shaping of a charge-sensitive transmon architecture. We engineer fast-load and fast-clear pulses that selectively suppress resonator photon overshoots, thereby demonstrating a practical strategy to mitigate MIST without requiring complex waveforms or real-time feedback. Utilizing active gate-voltage control and rapid feedback, the MIST probability is precisely mapped against ng and the steady-state resonator photon number, demonstrating a concrete agreement with numerical Floquet branch analysis. Lastly, we evaluate the ng-averaged total error probabilities for both readout and post-readout stages, verifying that a straightforward three-step pulse scheme consistently minimizes overall readout errors. These results show a great improvement in superconducting transmon qubit readout. Allowing us to start with a high readout fidelity as a great starting point in qubit operation. References: 1. W.-E. Lin, L.-C. Hsiao, C.-H. Ma, E.-H. Yeh, W.-L. Peng, H.-S. Goan, C.-S. Wu, Y.-N. Chen, Y.-F. Chen#, C.-T. Ke#, C.-D. Chen, “Mitigation of Measurement-Induced State Transitions via a Fast-Load and Fast-Clear Readout” arXiv:2607.23681v1 (2026). 3. W.-E. Lin, C.-H. Ma, E.-H. Yeh, W.-L. Peng, Y.-S. Wei, H.-S. Goan, C.-S. Wu, C.-T. Ke#, Y.-F. Chen#, C.-D. Chen, “Suppression of Quasiparticle Poisoning to 10−11 Levels in Superconducting Qubits via Infrared Shielding” arXiv:2606.07339 (2026).

Support Problems for Free Convolutions and Applications to Quantum Information Theory

• Hao-Wei Huang (NTHU)

In this talk, I will discuss support problems for free additive convolutions of probability measures on the real line, with emphasis on applications to quantum information theory. A central question is how the left endpoint of the support changes under free convolution powers. Using analytic subordination, global inversion, and boundary behavior of reciprocal Cauchy transforms, we obtain an explicit formula for the left endpoint of the support of a free convolution power. The formula is expressed in terms of the Nevanlinna representation and a critical integral condition associated with the representing measure. I will also explain why such support formulas are relevant to quantum information theory. In particular, they provide tools for studying positivity and $(k)$-positivity questions arising from random matrix models and positive maps. The support of the limiting spectral distribution often determines whether certain transformed random matrices are asymptotically positive, and hence whether the corresponding maps satisfy positivity-type properties. This connection gives a bridge between analytic free probability, random matrix theory, and structural questions in quantum information theory.

Unlocking the Potential of Fluxonium Qubits for Quantum Information Science

• Yen-Hsiang Lin (NTHU)

Superconducting qubits have emerged as one of the most promising platforms for quantum computing. Among them, the fluxonium qubit stands out for its long coherence times, large anharmonicity, and reduced sensitivity to noise, making it a promising candidate for scalable quantum processors and quantum-optical applications. In this talk, I will present our recent efforts to advance fluxonium-based quantum technologies, spanning materials, device fabrication, automated characterization, and quantum optics. I will first discuss our development of high-quality, non-aging superconducting materials and surface-passivation techniques that preserve the quality factors of superconducting microwave resonators for up to 14 months. By incorporating these techniques into qubit fabrication, we have achieved fluxonium energy-relaxation times, T1, as long as 3.3 ms. Toward scaling up device characterization and control, I will introduce our AI-assisted, FPGA-based measurement platform for fast and automated characterization of superconducting qubits. Finally, I will present our recent observation of electromagnetically induced transparency (EIT) in a single fluxonium artificial atom coupled to a microwave waveguide. These results illustrate how fluxonium can serve not only as a high-coherence computational qubit, but also as a versatile light–matter interface for manipulating microwave photons, opening opportunities for quantum interconnects and quantum networks.

Closing Remark