Guardado en:
Detalles Bibliográficos
Autores principales: Chen, Jun-Hao, Hung, Ming-Kai, Tsai, Yun-Cheng, Chen, Samuel Yen-Chi
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:https://arxiv.org/abs/2512.21820
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866908732016820224
author Chen, Jun-Hao
Hung, Ming-Kai
Tsai, Yun-Cheng
Chen, Samuel Yen-Chi
author_facet Chen, Jun-Hao
Hung, Ming-Kai
Tsai, Yun-Cheng
Chen, Samuel Yen-Chi
contents We compare two quantum sequence models, QLSTM and QFWP, under an Equal Parameter Count (EPC) and adjoint differentiation setup on daily EUR USD forecasting as a controlled one dimensional time series case study. Across 10 random seeds and batch sizes from 4 to 64, we measure component wise runtimes including train forward, backward, full train, and inference, as well as accuracy including RMSE and directional accuracy. Batched forward scales well by about 2.2 to 2.4 times, but backward scales modestly, with QLSTM about 1.01 to 1.05 times and QFWP about 1.18 to 1.22 times, which caps end to end training speedups near 2 times. QFWP achieves lower RMSE and higher directional accuracy at all batch sizes, supported by a Wilcoxon test with p less than or equal to 0.004 and a large Cliff delta, while QLSTM reaches the highest throughput at batch size 64, revealing a clear speed accuracy Pareto frontier. We provide an EPC aligned, numerically checked benchmarking pipeline and practical guidance on batch size choices, while broader datasets and hardware and noise settings are left for future work.
format Preprint
id arxiv_https___arxiv_org_abs_2512_21820
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Batched Training for QLSTM vs. QFWP: A System-Oriented Approach to EPC-Aware RMSE-DA
Chen, Jun-Hao
Hung, Ming-Kai
Tsai, Yun-Cheng
Chen, Samuel Yen-Chi
Computational Engineering, Finance, and Science
We compare two quantum sequence models, QLSTM and QFWP, under an Equal Parameter Count (EPC) and adjoint differentiation setup on daily EUR USD forecasting as a controlled one dimensional time series case study. Across 10 random seeds and batch sizes from 4 to 64, we measure component wise runtimes including train forward, backward, full train, and inference, as well as accuracy including RMSE and directional accuracy. Batched forward scales well by about 2.2 to 2.4 times, but backward scales modestly, with QLSTM about 1.01 to 1.05 times and QFWP about 1.18 to 1.22 times, which caps end to end training speedups near 2 times. QFWP achieves lower RMSE and higher directional accuracy at all batch sizes, supported by a Wilcoxon test with p less than or equal to 0.004 and a large Cliff delta, while QLSTM reaches the highest throughput at batch size 64, revealing a clear speed accuracy Pareto frontier. We provide an EPC aligned, numerically checked benchmarking pipeline and practical guidance on batch size choices, while broader datasets and hardware and noise settings are left for future work.
title Batched Training for QLSTM vs. QFWP: A System-Oriented Approach to EPC-Aware RMSE-DA
topic Computational Engineering, Finance, and Science
url https://arxiv.org/abs/2512.21820