Improving Quantum Recurrent Neural Networks with Amplitude Encoding Journal: Nuclear Physics B

Jack Morgan, Hamed Mohammadbagherpoor, Eric Ghysels · 2025

Read Improving Quantum Recurrent Neural Networks with Amplitude Encoding Journal: Nuclear Physics B by Jack Morgan, Hamed Mohammadbagherpoor, Eric Ghysels (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-16784.txt.

Description

Jack Morgan, Hamed Mohammadbagherpoor, and Eric Ghysels improve, in a 2025 preprint, quantum recurrent neural networks that read sequences over time. Three changes do the work. First, inputs packed into quantum amplitudes are joined by their pre-normalized sizes, extra numbers that help the model generalize. On two financial datasets, that tweak lifted performance. Second, the costly exact circuit that loads data can be swapped for a cheaper approximate one. A trained model keeps nearly the same accuracy with the swap.

Third, a new circuit layout matches the original model's math on a perfect machine while cutting depth a lot. Shallower circuits suffer less from today's noisy hardware. Together, the three changes make the quantum recurrent model cheaper to run and better behaved on real data. The authors frame them as practical fixes other quantum learning teams can adopt piecemeal.

Catalog record

CREATORJack Morgan, Hamed Mohammadbagherpoor, Eric Ghysels
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE30 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-16784

Browse the full library · About the Tower of Babel