Quantum Machine Learning Algorithms for Drug Discovery Applications

Abstract

The growing quantity of public and private data sets focused on small molecules screened against biological targets or whole organisms provides a wealth of drug discovery relevant data. This is matched by the availability of machine learning algorithms such as Support Vector Machines (SVM) and Deep Neural Networks (DNN) that are computationally expensive to perform on very large data sets with thousands of molecular descriptors. Quantum computer (QC) algorithms have been proposed to offer an approach to accelerate quantum machine learning over classical computer (CC) algorithms, however with significant limitations. In the case of cheminformatics, which is widely used in drug discovery, one of the challenges to overcome is the need for compression of large numbers of molecular descriptors for use on a QC. Here, we show how to achieve compression with data sets using hundreds of molecules (SARS-CoV-2) to hundreds of thousands of molecules (whole cell screening data sets for plague and M. tuberculosis) with SVM and the data reuploading classifier (a DNN equivalent algorithm) on a QC benchmarked against CC and hybrid approaches. This study illustrates the steps needed in order to be “quantum computer ready” in order to apply quantum computing to drug discovery and to provide the foundation on which to build this field.

QNet: A Scalable and Noise-Resilient Quantum Neural Network Architecture for Noisy Intermediate-Scale Quantum Computers

“…PCA + Baseline: A popular quantum-classical hybrid QNN model targetted for smaller quantum devices uses classical algorithm (e.g., PCA) to reduce data dimension to a level that is tractable for the "Baseline" model [15,17,23,24] (Figure 4A). We refer to this model as "PCA + Baseline."…”

“…In (A), the high dimensional data is transformed to a lower dimension using a classical algorithm (e.g., PCA, LDA, etc.). The lower-dimensional data is used as inputs to the Baseline QNN model [17,23,24]. In (B), classical features are loaded repeatedly in a singlequbit using a series of rotation operations [25][26][27].…”

“…Numerous choices exist for the encoding circuits, PQC, and measurements to build a QNN [9,[11][12][13]. Recent works have demonstrated application of QNN in image classification [14,15], drug discovery [16,17], finance [18,19], and many other industrial problems [20].…”

“…Two variants of the conventional QNN model are widely used to address the above scalability issue: 1) reduction of data dimension using a classical algorithm (e.g., PCA) before training a QML [17,23,24], and 2) repeatedly uploading the FIGURE 1 | Conventional quantum neural network (QNN) and the proposed architecture (QNet). Each conventional QNN consists of a classical-to-quantum data encoding circuit followed by a parametric quantum circuit and measurement operations.…”

Quantum version of the k-NN classifier based on a quantum sorting algorithm

“…In 2017, Biamonte et al show the complexity and main characteristics of various quantum machine learning algorithms [22]. More recently, in 2021, proposals on quantum machine learning included: encoding patterns for quantum algorithms [23], financial applications [24], federated quantum machine learning [25], an algorithm for knowledge graphs [26] and algorithms for drug discovery applications [27].…”