![[2503.01927] QCS-ADME: Quantum Circuit Search for Drug Property Prediction with Imbalanced Data and Regression Adaptation](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2503.01927] QCS-ADME: Quantum Circuit Search for Drug Property Prediction with Imbalanced Data and Regression Adaptation
Summary
The paper presents QCS-ADME, a novel quantum circuit search framework for predicting drug properties, addressing challenges in imbalanced data and regression adaptation.
Why It Matters
As quantum machine learning (QML) gains traction in biomedical applications, this research highlights a significant advancement in predicting ADME properties, which are crucial for drug development. The proposed methods could enhance the accuracy of predictions in scenarios where traditional machine learning struggles, particularly with imbalanced datasets.
Key Takeaways
- Introduces a training-free scoring mechanism for evaluating QML circuits.
- Demonstrates significant correlation between scoring metrics and performance in imbalanced classification tasks.
- Develops methods to quantify relationships between quantum states for regression tasks.
- Validates the approach on various ADME tasks, outperforming baseline methods.
- Addresses the dual challenges of classification and regression in drug property prediction.
Quantum Physics arXiv:2503.01927 (quant-ph) [Submitted on 2 Mar 2025 (v1), last revised 25 Feb 2026 (this version, v2)] Title:QCS-ADME: Quantum Circuit Search for Drug Property Prediction with Imbalanced Data and Regression Adaptation Authors:Kangyu Zheng, Tianfan Fu, Zhiding Liang View a PDF of the paper titled QCS-ADME: Quantum Circuit Search for Drug Property Prediction with Imbalanced Data and Regression Adaptation, by Kangyu Zheng and 2 other authors View PDF HTML (experimental) Abstract:The biomedical field is beginning to explore the use of quantum machine learning (QML) for tasks traditionally handled by classical machine learning, especially in predicting ADME (absorption, distribution, metabolism, and excretion) properties, which are essential in drug evaluation. However, ADME tasks pose unique challenges for existing quantum computing systems (QCS) frameworks, as they involve both classification with unbalanced dataset and regression problems. These dual requirements make it necessary to adapt and refine current QCS frameworks to effectively address the complexities of ADME predictions. We propose a novel training-free scoring mechanism to evaluate QML circuit performance on imbalanced classification and regression tasks. Our mechanism demonstrates significant correlation between scoring metrics and test performance on imbalanced classification tasks. Additionally, we develop methods to quantify continuous similarity relationships between quantum states, enablin...
