arXiv subjects

Yi-Ting Lee

Publications and source records attributed to Yi-Ting Lee.

7 recordsLinked to original sources

arXiv2026

HSMLog: Small Language Model-Assisted Hardware Security Module Log Anomaly Detection with Behavioral Analysis

Hardware Security Module (HSM) logs capture security-critical behavior, but anomalies emerge from relationships across event sequences, keys, object states, sessions, and temporal patterns rather than isolated events. Existing methods separate detection from HSM-specific evidence validation and reporting. In this paper, we present HSMLog, a two-stage framework for HSM log anomaly detection with retrieval-grounded behavioral analysis. In Stage 1, a small language model (SLM) identifies candidate alerts from sliding windows of structured HSM events and performs policy-guided assessment using HSM-specific operational rules. In Stage 2, retrieved policies and historical suspicious-key records strictly predating the alert window, together with candidate-related log context, support conservative candidate review and incident analysis. Evaluated on real industrial HSM background logs augmented with anomaly scenarios co-defined with industrial partners, HSMLog achieves 98.97% precision, 96.00% recall, 98.66% anomalous-event coverage, and a 97.46% F1 score, demonstrating effective anomaly alerting and incident triage in the studied setting.

cs.CR↗

arXiv2026

Benchmarking quantum simulation with neutron-scattering experiments

Realistic simulation of quantum materials is a central goal of quantum computation. Although quantum processors have advanced rapidly in scale and fidelity, it has remained unclear whether pre-fault-tolerant devices can perform quantitatively reliable material simulations. We demonstrate that a superconducting quantum processor operating on up to 50 qubits can already produce meaningful, quantitative comparisons with inelastic neutron-scattering measurements of KCuF$_3$, a canonical realization of a gapless Luttinger liquid system with a strongly correlated ground state and a spectrum of emergent spinons. The quantum simulation is enabled by a quantum-classical workflow for computing dynamical structure factors (DSFs). The resulting spectra are benchmarked against experimental measurements using multiple metrics, highlighting the impact of circuit depth and circuit fidelity on simulation accuracy. Finally, we extend our simulations to a 1D XXZ Heisenberg model with next-nearest-neighbor (NNN) interactions and a strong anisotropy, producing a gapped excitation spectrum, which could be used to describe the CsCoX$_3$ compounds above the Néel temperature. Our results establish a framework for computing DSFs for quantum materials in classically challenging regimes of strong entanglement and long-range interactions, enabling quantum simulations that are directly testable against laboratory measurements.

quant-ph↗

arXiv2026

Quantum Computing and Error Mitigation with Deep Learning for Frenkel Excitons

Quantum computers, currently in the noisy intermediate-scale quantum (NISQ) era, have started to provide scientists with a novel tool to explore quantum physics and chemistry. While several electronic systems have been extensively studied, Frenkel excitons, as prototypical optical excitations, remain among the less-explored applications. Here, we first use variational quantum deflation to calculate the eigenstates of the Frenkel Hamiltonian and evaluate the observables based on the oscillator strength for each eigenstate. Furthermore, using NISQ quantum computers requires performing error mitigation techniques alongside simulations. To deal with noisy qubits, we developed a deep-learning-based framework combined with a post-selection technique to learn the noise pattern and mitigate the error. Our mitigation methods work well and outperform the conventional post-selection and remain valid on real hardware.

quant-ph↗

arXiv2025

Digital Quantum Simulation of Spin Transport

Understanding transport phenomena in quantum spin systems has long intrigued physicists due to their potential applications in spintronic devices and spin qubits. Here, using a superconducting-qubit-based transmon device, we show that pre-fault-tolerant digital quantum simulation is reliable for studying transport phenomena via spin-current autocorrelation function (ACF). While quantum simulations of the spin-spin ACF have been used to probe spin transport, methods based on the spin-current ACF have yet to be demonstrated due to their high gate cost, despite offering more direct information relevant to the transport properties. Overcoming the resource constraints set by indirect measurement schemes like the Hadamard test, we showcase a direct measurement scheme that utilizes non-unitary operations, in particular mid-circuit measurements, to investigate spin transport for the 40-site 1D XXZ Heisenberg model in the near-ballistic, superdiffusive, and diffusive regimes. We successfully reproduce the expected power-law behavior in the superdiffusive regime and vanishing of the Drude weight in the diffusive regime.

quant-ph↗

arXiv2025

MD-Syn: Synergistic drug combination prediction based on the multidimensional feature fusion method and attention mechanisms

Drug combination therapies have shown promising therapeutic efficacy in complex diseases and have demonstrated the potential to reduce drug resistance. However, the huge number of possible drug combinations makes it difficult to screen them all in traditional experiments. In this study, we proposed MD-Syn, a computational framework, which is based on the multidimensional feature fusion method and multi-head attention mechanisms. Given drug pair-cell line triplets, MD-Syn considers one-dimensional and two-dimensional feature spaces simultaneously. It consists of a one-dimensional feature embedding module (1D-FEM), a two-dimensional feature embedding module (2D-FEM), and a deep neural network-based classifier for synergistic drug combination prediction. MD-Syn achieved the AUROC of 0.919 in 5-fold cross-validation, outperforming the state-of-the-art methods. Further, MD-Syn showed comparable results over two independent datasets. In addition, the multi-head attention mechanisms not only learn embeddings from different feature aspects but also focus on essential interactive feature elements, improving the interpretability of MD-Syn. In summary, MD-Syn is an interpretable framework to prioritize synergistic drug combination pairs with chemicals and cancer cell line gene expression profiles. To facilitate broader community access to this model, we have developed a web portal (https://labyeh104-2.life.nthu.edu.tw/) that enables customized predictions of drug combination synergy effects based on user-specified compounds.

cs.LG↗

arXiv2025

Quantum simulations of defects near the (0001) surface of $α$-Al$_2$O$_3$

Defects in materials are ubiquitous and one of their adverse effects in $α$-Al$_2$O$_3$ is the initiation of corrosion. While this process starts near the surface, the defects involved and their electronic structure need to be elucidated with high accuracy. Since point defects are confined to a small spatial region, defect embedding theory allows the definition of an active space, comprising of the defect electronic states, that is coupled to the environment of the host material. The active space Hamiltonian is of small rank, enabling access to its electronic properties using a high-level or even exact quantum theory. In this paper we use these techniques and first-principles simulations to compute the structural and electronic properties of near-surface vacancies for the (0001) surface of $α$-Al$_2$O$_3$, and investigate the influence of defects and hydration on the initiation and propagation of corrosion. We report the defect electronic structure for strongly localized ground and excited states of the surface O vacancy and compare results obtained using full configuration interaction and a variational quantum eigensolver on a quantum computer. Error mitigation techniques are explored and shown to reduce the error due to the hardware noise to the point where the quantum result agrees with the exact solution within chemical accuracy.

cond-mat.mtrl-sci↗

arXiv2023

Exposing the Functionalities of Neurons for Gated Recurrent Unit Based Sequence-to-Sequence Model

The goal of this paper is to report certain scientific discoveries about a Seq2Seq model. It is known that analyzing the behavior of RNN-based models at the neuron level is considered a more challenging task than analyzing a DNN or CNN models due to their recursive mechanism in nature. This paper aims to provide neuron-level analysis to explain why a vanilla GRU-based Seq2Seq model without attention can achieve token-positioning. We found four different types of neurons: storing, counting, triggering, and outputting and further uncover the mechanism for these neurons to work together in order to produce the right token in the right position.

cs.NE↗