Published November 16, 2026 | Version v1

Data, Configurations, and Results for QRNN Time-Series Benchmark

  • 1. ROR icon Centro de Supercomputación de Galicia
  • 2. ROR icon Universidade de Santiago de Compostela

Description

This repository contains the datasets, shared experimental configurations, and numerical results for QRNN Time-Series Benchmark. The study compares four variational quantum recurrent neural network (QRNN) architectures with three classical recurrent neural network baselines. The comparison covers three time-series tasks: short-term memory (STM), nonlinear autoregressive moving average (NARMA), and Mackey–Glass (MG) forecasting. The analysis focuses on memory and nonlinear processing capacity, the dependence of performance on memory size and task difficulty, training convergence, and the stationarity of the quantum recurrent dynamics.

Files

qrnn_benchmark_results_repo.zip

Files (9.2 MB)

Name Size
md5:f423ae2f5257dfef8a7fcec9fd374e37
9.2 MB Preview Download

Additional details

Dates

Available
2026-11-16