Conference Schedule

How to build good quantum models: engineering simplicity biases in Fourier space -- A lot of effort in quantum machine learning is focused on designing classically intractable models. But what makes a model actually good for learning? Machine learning theory tells us that good models are flexible, but have a simplicity bias that prevents them from overfitting. In this talk I will argue why quantum computers could be suitable hardware for engineering such ”soft” simplicity biases: their ability to efficiently perform the Quantum Fourier Transform gives them access to information that can help regularise complexity. I will discuss a few examples of how this could be implemented in practice, and sketch the research required to validate this approach.

• Invited Talk

Cumulant-based Quantum Relative Rényi Functional -- Based on the connection between cumulant-generating functions and classical relative Rényi entropy, we introduce a cumulant-based quantum relative Rényi functional and investigate its basic properties. We show, through numerical counterexamples, that the functional violates the data-processing inequality under general completely positive trace-preserving channels, and therefore is not a quantum divergence in the usual sense. Nevertheless, we identify a distinguished special instance for which numerical evidence supports monotonicity under commutativity-preserving channels, namely channels that cannot generate non-commutativity from commuting input states. This suggests that the special instance captures a form of non-classicality associated specifically with non-commutativity, rather than general state distinguishability. Motivated by the structural role of non-commuting states in distinguishing quantum state space from classical probability theory, we propose this quantity as a candidate quantumness witness and as a possible starting point for a resource theory of non-commutativity. We further develop a path-integral-like representation of the functional and show that it leads to a fluctuation-theorem-like relation in quantum thermodynamic settings.

Speaker: Kasidit Srimahajariyapong, Yixian Qiu

• Oral Session

Optically Assigned Vector Magnetometry with Nitrogen-Vacancy Center Ensembles in Diamond via Polarization Anisotropy -- The nitrogen-vacancy (NV) center in diamond is a promising platform for quantum magnetometry due to its spin-dependent optical readout and ability to operate under ambient conditions. For vector magnetometry, an ensemble of NV centers aligned along four different crystallographic orientations can be used to determine both the magnitude and direction of a magnetic field via optically detected magnetic resonance (ODMR) spectra. However, due to the symmetry and equivalent crystallographic planes of the diamond lattice, full reconstruction of the magnetic field vector typically requires an external bias field to lift the degeneracy between NV orientations. Here, we demonstrate that this degeneracy can be lifted by illuminating the diamond at an oblique angle and analyzing the resulting polarization-dependent ODMR spectra. The unique dipole orientation of the NV centers enables the unambiguous assignment of ODMR transitions to specific NV axes, and we show that as few as eight measurements are sufficient to reconstruct the full magnetic field vector. Our method enables rapid vector magnetometry of arbitrary magnetic fields without the need to apply or vary a reference field, especially when all the ODMR transitions are well-resolved. This approach opens the door to more compact and portable magnetometers and may facilitate real-time magnetic field imaging in materials science and biological systems.

• QTRic

Speaker: Suppawit Ngonsamrong

• Oral Session

Speaker: Giuseppe De Riso, Caesnan Leditto, Jeremiah Rowland, Sujeet Bhalerao

• Oral Session

Resolved Sideband Cooling via Spin-1 Rotations of Single Trapped Ion -- The development of trapped-ion optical atomic clocks requires precise control of the ion’s motional state to suppress systematic frequency shifts arising from excess thermal motion. Efficient cooling of trapped ions is essential for improving clock accuracy and enabling high-fidelity quantum state manipulation. In this work, we present the Raman transitions between Zeeman sublevels to implement spin-1 rotations. The transitions are used to characterize the thermal motion of single trapped ion. The spin-1 rotational dynamics are further employed to implement resolved sideband cooling, enabling efficient reduction of the ion’s vibrational occupation number. These results establish spin-1 Raman transition as a versatile approach for motional control in trapped-ion systems and provide a pathway toward minimizing systematic effects in the optical atomic clock at National Institute of Metrology (Thailand).

• Invited Talk

Speaker: Thapawat Muensuwan, Nur Fadhillah Binti Rahimi

• Oral Session