Discretization-Aware Fine-Tuning for Quantum Machine Learning with Chemical Foundation Models
Abstract
A key challenge in practical quantum machine learning (QML), particularly for discriminative tasks such as classification, is the limited capacity of near-term quantum devices to encode high-dimensional classical data into small quantum registers. In optimized basis-encoded (bit-bit) settings, this constraint leads to cross-class collisions, where samples with different labels are mapped to the same discrete bit-string and thus become indistinguishable to any downstream model. In this work, we investigate how data representation affects QML performance under such severe information bottlenecks. We introduce discretization-aware fine-tuning (DAFT), a method that adapts a pre-trained chemical foundation model to produce representations that remain informative after quantization. DAFT reduces collision probability through a differentiable soft collision loss. We evaluate both quantum and classical models under a controlled setting in which they receive identical discretized bit-string inputs, isolating the effect of representation from model architecture. On the blood-brain barrier penetration (BBBP) molecular property prediction benchmark using ChemBERTa-77M, DAFT reduces collision counts by several orders of magnitude and improves quantum classification accuracy by more than 12 percentage points compared to a frozen backbone. Importantly, without DAFT, classical models outperform QML under the same input constraints. With DAFT, however, this comparison reverses at higher qubit counts. At 10 qubits, the quantum model surpasses a matched classical baseline trained on identical bit-strings (0.883 vs. 0.855, p = 0.026). These results show that, in information-constrained regimes, achieving a quantum advantage critically depends on aligning continuous representations with discrete quantum encodings.
Related
- Harnessing Machine Learning and Composite Quantum Chemical Methods for Computing Accurate Bond Dissociation Energies of PFAS
- When Quantum Meets AI: Quantum Methods for Machine Learning and Machine Learning Methods for Quantum Systems
- Application a quantum machine learning algorithm to the applied problem of modeling the thermal regime of a servers
- Photonic Time-Delayed Quantum Extreme Learning Machine
- Machine learning quantum data to accelerate quantum Monte Carlo
BibTeX
@misc{matsuura2026discretizationaware,
title = {{Discretization-Aware Fine-Tuning for Quantum Machine Learning with Chemical Foundation Models}},
author = {Shunji Matsuura and Sonika Johri},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.03220},
eprint = {2609.03220},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.03220},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Discretization-Aware Fine-Tuning for Quantum Machine Learning with Chemical Foundation Models
AU - Shunji Matsuura
AU - Sonika Johri
PY - 2026
DA - 2026/09/02
PB - arXiv
DO - 10.48550/arxiv.2609.03220
UR - https://arxiv.org/abs/2609.03220
AB - A key challenge in practical quantum machine learning (QML), particularly for discriminative tasks such as classification, is the limited capacity of near-term quantum devices to encode high-dimensional classical data into small quantum registers. In optimized basis-encoded (bit-bit) settings, this constraint leads to cross-class collisions, where samples with different labels are mapped to the same discrete bit-string and thus become indistinguishable to any downstream model. In this work, we investigate how data representation affects QML performance under such severe information bottlenecks. We introduce discretization-aware fine-tuning (DAFT), a method that adapts a pre-trained chemical foundation model to produce representations that remain informative after quantization. DAFT reduces collision probability through a differentiable soft collision loss. We evaluate both quantum and classical models under a controlled setting in which they receive identical discretized bit-string inputs, isolating the effect of representation from model architecture. On the blood-brain barrier penetration (BBBP) molecular property prediction benchmark using ChemBERTa-77M, DAFT reduces collision counts by several orders of magnitude and improves quantum classification accuracy by more than 12 percentage points compared to a frozen backbone. Importantly, without DAFT, classical models outperform QML under the same input constraints. With DAFT, however, this comparison reverses at higher qubit counts. At 10 qubits, the quantum model surpasses a matched classical baseline trained on identical bit-strings (0.883 vs. 0.855, p = 0.026). These results show that, in information-constrained regimes, achieving a quantum advantage critically depends on aligning continuous representations with discrete quantum encodings.
ER -