arXiv subjects

Tse-Yu Lin

Publications and source records attributed to Tse-Yu Lin.

5 recordsLinked to original sources

arXiv2025

Fundamental limit on the heralded single photons' spectral brightness

Heralded single photons (HSPs) are the versatile flying qubits in quantum communication and networks due to their ability to remove the randomness of arrival time and enhance the transmission reliability. As the generation rate of HSPs increases or their linewidth narrows, both of which are desirable for quantum information processing, the fundamental limit of spectral brightness (SB), defined as the generation rate per unit linewidth, remains unclear. To examine the existence and value of such a limit, we systematically studied the SB together with the cross-correlation function, or equivalently, the signal-to-background ratio (SBR). We ultimately derive an upper bound on SB that applies universally to all types of HSP sources. A newly defined quantity governs this limit, the quality factor, which is the product of SBR and effective SB. The quality factor indicates how closely an HSP source approaches an ideal noise-free source. Furthermore, by employing an HSP source based on hot atomic vapor, we achieved an SB of $(8.5\pm0.3)$$\times$$10^5$ pairs/s/MHz and a quality factor of $0.73\pm0.02$ under the single-photon criterion. Both values represent the highest reported performance to date among all HSP platforms. These results provide a unified benchmark for evaluating and optimizing HSP sources.

quant-ph↗

arXiv2025

Temporally-long C-band heralded single photons generated from hot atoms

C-band photons are recognized for having the lowest loss coefficient in optical fibers, making them highly favorable for optical fiber-based communication. In this study, we systematically investigated the temporal width of C-band heralded single photons and developed a theoretical model for biphoton generation via the spontaneous four-wave mixing (SFWM) process using a diamond-type transition scheme, which has not been previously reported. Our experimental data on temporal width closely aligns with the predictions of this model. Additionally, we introduced a new concept: the atomic velocity group relating to the two-photon resonance condition and the one-photon detuning in this atomic frame. These two parameters are crucial for understanding the behavior of the biphoton source. The concept indicates that the hot-atom source behaves similarly to the cold-atom source. Guided by our theoretical model, we observed 1529-nm (C-band) heralded single photons with a temporal width of 28.3$\pm$0.6 ns, corresponding to a linewidth of 11.0$\pm$0.2 MHz. For comparison, the ultimate linewidth limit is 6.1 MHz, determined by the natural linewidth of the atoms. Among all atom-based sources of 1300 to 1550 nm heralded single photons utilizing either cold or hot atoms, the temporal width achieved in this work represents the first instance of a width exceeding 10 ns, making it (or its linewidth) the longest (or narrowest) record to date. This work significantly enhances our understanding of diamond-type or cascade-type SFWM biphoton generation and marks an important milestone in achieving greater temporal width in atom-based sources of C-band heralded single photons.

quant-ph↗

arXiv2025

Protecting Heralded Single Photons Generated from Double-$\Lambda$ Biphoton Sources with Doppler-Broadened Atomic Media

Biphoton sources that use room-temperature or hot atoms are valuable for real-world applications in long-distance quantum communication and photonic quantum computation. The heralded single photons produced by biphoton sources using the double-$\Lambda$ spontaneous four-wave mixing (SFWM) process offer advantages of narrow linewidth, stable frequency, and tunable linewidth -- qualities not found in other types of biphoton sources. In this study, we investigated a hot-atom SFWM double-$\Lambda$ biphoton source. We discovered that, under the condition counterintuitive to the present theory, heralded single photons of the source enhanced their generation rate by a factor of 3.6, heralding probability by a factor of 3.0, temporal width by 2.1, and spectral brightness by a factor of 10. These unexpected findings led us to propose a new theoretical framework for a previously unexplored physical mechanism. Our proposed theory effectively explains the observed results. Traditionally, similar spectral brightness (SB) from atom-based sources resulted in a lower signal-to-background ratio (SBR) than crystal- or chip-based biphoton sources, mainly due to poorer heralding probabilities. In our work, we experimentally demonstrated that the SBR improved by a factor of 4.8 while maintaining a comparable SB. As a result, the SBR performance of the atom-based biphoton source is now on par with that of crystal- or chip-based sources. This research introduces a new tuning parameter for double-$\Lambda$ SFWM biphoton sources, enhances our understanding of biphoton generation, and opens new avenues for improving the performance of these sources.

quant-ph↗

arXiv2022

Performance Prediction in Major League Baseball by Long Short-Term Memory Networks

Player performance prediction is a serious problem in every sport since it brings valuable future information for managers to make important decisions. In baseball industries, there already existed variable prediction systems and many types of researches that attempt to provide accurate predictions and help domain users. However, it is a lack of studies about the predicting method or systems based on deep learning. Deep learning models had proven to be the greatest solutions in different fields nowadays, so we believe they could be tried and applied to the prediction problem in baseball. Hence, the predicting abilities of deep learning models are set to be our research problem in this paper. As a beginning, we select numbers of home runs as the target because it is one of the most critical indexes to understand the power and the talent of baseball hitters. Moreover, we use the sequential model Long Short-Term Memory as our main method to solve the home run prediction problem in Major League Baseball. We compare models' ability with several machine learning models and a widely used baseball projection system, sZymborski Projection System. Our results show that Long Short-Term Memory has better performance than others and has the ability to make more exact predictions. We conclude that Long Short-Term Memory is a feasible way for performance prediction problems in baseball and could bring valuable information to fit users' needs.

cs.LG↗

arXiv2018

An Application of HodgeRank to Online Peer Assessment

Bias and heterogeneity in peer assessment can lead to the issue of unfair scoring in the educational field. To deal with this problem, we propose a reference ranking method for an online peer assessment system using HodgeRank. Such a scheme provides instructors with an objective scoring reference based on mathematics.

stat.ML↗