We are pleased to announce a new paper accepted at QCNC 2025:
G. Maragkopoulos, N. Stefanakos, A. Mandilara & D. Syvridis
“Applications of Hybrid Machine Learning Methods to Large Datasets: A Case Study”
🔗 Access the preprint here: https://arxiv.org/abs/2504.06892
Key Contributions:
- Proposes a hybrid ML framework combining a classical deep network and a variational quantum circuit (VQC) on a single qudit to classify time-series data.
- Utilizes a trainable autoencoder embedding to map classical data into the quantum circuit’s space efficiently.
- Demonstrates that classification performance can be maintained, while drastically reducing the number of trainable parameters—a path toward resource-efficient quantum-augmented learning.
- Highlights how tailored preprocessing and embedding strategies are crucial for maximizing the benefit of hybrid quantum-classical architectures.
This work aligns with the Allegro EU Project’s mission to explore quantum-enhanced data processing methods in communication and network contexts.