pith. sign in

arxiv: 2409.15645 · v1 · submitted 2024-09-24 · 🪐 quant-ph · stat.ML

Quantum Machine Learning in Drug Discovery: Applications in Academia and Pharmaceutical Industries

Anthony M. Smaldone , Yu Shee , Gregory W. Kyro , Chuzhi Xu , Nam P. Vu , Rishab Dutta , Marwa H. Farag , Alexey Galda
show 3 more authors
Sandeep Kumar Elica Kyoseva Victor S. Batista
This is my paper

Pith reviewed 2026-05-23 21:10 UTC · model grok-4.3

classification 🪐 quant-ph stat.ML
keywords quantum machine learningdrug discoveryquantum neural networksmolecular property predictionmolecular generationvariational quantum circuitshybrid quantum-classicalgate-based quantum computers
0
0 comments X

The pith

Quantum neural networks on gate-based computers can improve molecular property prediction and generation for drug discovery.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper reviews the foundations and applications of quantum machine learning, focusing on quantum neural networks implemented via variational circuits on gate-based quantum hardware. It examines how data encoding into quantum states and hybrid quantum-classical methods support tasks in chemistry relevant to drug discovery. A sympathetic reader would care because these methods could accelerate the identification of molecular properties and the creation of new drug candidates if they scale beyond current hardware limits. The review covers both the theoretical setup and real-world challenges in academia and industry settings.

Core claim

The paper claims that the combination of quantum computing and machine learning through quantum neural networks on gate-based devices establishes a pathway for advancements in chemistry applications to drug discovery, specifically by enabling more effective molecular property prediction and molecular generation, supported by data encoding techniques, variational quantum circuits, and hybrid approaches, while noting the need to address practical implementation barriers.

What carries the argument

Variational quantum circuits within quantum neural networks, which encode classical molecular data into quantum states and optimize parameters in a hybrid quantum-classical loop to perform prediction and generation tasks.

If this is right

  • Molecular property prediction becomes feasible through quantum variational methods applied to chemical structures.
  • Molecular generation tasks gain from quantum generative models built on the same circuit architectures.
  • Hybrid quantum-classical workflows allow current hardware to contribute to drug discovery pipelines despite limited qubit counts.
  • Academic and pharmaceutical research can integrate these techniques once encoding and optimization steps are refined.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If the hybrid methods scale, they could reduce the time from molecular screening to candidate selection in industry settings.
  • The same circuit structures might extend to related problems such as protein folding or reaction pathway prediction.
  • Hardware improvements that lower error rates would directly test whether the predicted speedups materialize over classical baselines.

Load-bearing premise

Theoretical quantum machine learning methods including data encoding and variational circuits can be run effectively on near-term quantum hardware for actual drug discovery problems.

What would settle it

A direct comparison on a standard molecular dataset where a quantum neural network fails to match or exceed classical machine learning accuracy when executed on current gate-based quantum processors with realistic noise levels.

Figures

Figures reproduced from arXiv: 2409.15645 by Alexey Galda, Anthony M. Smaldone, Chuzhi Xu, Elica Kyoseva, Gregory W. Kyro, Marwa H. Farag, Nam P. Vu, Rishab Dutta, Sandeep Kumar, Victor S. Batista, Yu Shee.

Figure 1
Figure 1. Figure 1: Structure of a typical quantum neural network. The input is encoded into a [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Data encoding methods. (a) Quantum circuit to prepare the [1,0,1] vector with [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Generic three qubit quantum neural network using x-axis angle encoding (blue), [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Training a general variational quantum circuit (blue) via classical post-processing [PITH_FULL_IMAGE:figures/full_fig_p011_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: A classical graph neural network for extracting features from a molecule. [PITH_FULL_IMAGE:figures/full_fig_p011_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: QCNN architecture introduced by Cong et al.66 with log(N) parameters, where N is the number of qubits. small kernel and the input data. This is attractive, as it allows the quantum approach to load only a small amount of information at a time onto quantum devices, as determined by the kernel size, which is of paramount importance during the NISQ era. This feature of QCNNs can be particularly useful in a bi… view at source ↗
Figure 7
Figure 7. Figure 7: Hybrid Quantum-Convolutional Neural Network (HQCNN), adapted from Smal [PITH_FULL_IMAGE:figures/full_fig_p016_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Quantum CNN Summary: (Top) Quantum circuit by Smaldone and Batista [PITH_FULL_IMAGE:figures/full_fig_p017_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Optimal folded mRNA structure of the 42-nucleotide sequence computed using [PITH_FULL_IMAGE:figures/full_fig_p018_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Different types of Quantum Autoencoder (QAE). (a) QAE utilizing a fully quan [PITH_FULL_IMAGE:figures/full_fig_p019_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Classical architecture of a transformer decoder layer alongside the corresponding [PITH_FULL_IMAGE:figures/full_fig_p024_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: (Top) Schematic representation of a superconducting cavity resonator coupled to a [PITH_FULL_IMAGE:figures/full_fig_p027_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Circuit diagrams for the R-ECD encoding method. (a) General encoder archi [PITH_FULL_IMAGE:figures/full_fig_p030_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: The CUDA-Q software stack. CUDA-Q builds off of a core MLIR-based in [PITH_FULL_IMAGE:figures/full_fig_p032_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Logarithmic (log10) execution time for one ‘observe’ call (i.e., measure the observ￾able operator applied to the state vector/ wave-function, also known as expectation value) for each data-point (10 thousand data-points), i.e., in total there are ten thousand expec￾tation values, on a single CPU versus a single GPU for a one layer of the parameterized quantum circuit (PQC) similar to the PQC shown in [PI… view at source ↗
Figure 16
Figure 16. Figure 16: An example of a parameterized quantum circuit employed in QNNs. [PITH_FULL_IMAGE:figures/full_fig_p033_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Gate fusion fuses multiple gates into one larger gate. [PITH_FULL_IMAGE:figures/full_fig_p035_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: Execution time for ‘observe’ call (expectation value) made for ten thousand data [PITH_FULL_IMAGE:figures/full_fig_p036_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: An example of multi-QPU backend with multi-GPU. Here, there are two virtual [PITH_FULL_IMAGE:figures/full_fig_p039_19.png] view at source ↗
Figure 20
Figure 20. Figure 20: Single-qubit and two-qubit gates translate to rank-2 and rank-4 tensors, respec [PITH_FULL_IMAGE:figures/full_fig_p040_20.png] view at source ↗
Figure 21
Figure 21. Figure 21: Tensor diagram (left) and example of matrix like contractions (right). [PITH_FULL_IMAGE:figures/full_fig_p041_21.png] view at source ↗
read the original abstract

The nexus of quantum computing and machine learning - quantum machine learning - offers the potential for significant advancements in chemistry. This review specifically explores the potential of quantum neural networks on gate-based quantum computers within the context of drug discovery. We discuss the theoretical foundations of quantum machine learning, including data encoding, variational quantum circuits, and hybrid quantum-classical approaches. Applications to drug discovery are highlighted, including molecular property prediction and molecular generation. We provide a balanced perspective, emphasizing both the potential benefits and the challenges that must be addressed.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 1 minor

Summary. This manuscript is a literature review on quantum machine learning (QML) with a focus on applications in drug discovery. It examines the potential of quantum neural networks on gate-based quantum computers, covering theoretical foundations such as data encoding, variational quantum circuits, and hybrid quantum-classical approaches. The review highlights applications including molecular property prediction and molecular generation, while presenting a balanced discussion of benefits and challenges in both academic and pharmaceutical contexts.

Significance. If the literature synthesis is accurate and reasonably comprehensive, the paper offers a useful overview of an emerging interdisciplinary area at the intersection of quantum computing and chemistry. The explicit framing as future-oriented potential rather than near-term feasibility is a strength, as is the inclusion of challenges alongside opportunities.

minor comments (1)
  1. The abstract and introduction could more explicitly state the review's scope (e.g., time period of literature covered or selection criteria for cited works) to help readers assess completeness.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive assessment of the manuscript, recognition of its balanced perspective on opportunities and challenges, and recommendation to accept. We appreciate the note that the future-oriented framing is a strength.

Circularity Check

0 steps flagged

No significant circularity; literature review with no derivations

full rationale

This is a literature review paper with no original derivations, equations, fitted parameters, predictions, or ansatzes. The central claim concerns the potential of QML for drug discovery tasks and is explicitly framed as forward-looking while acknowledging challenges. No load-bearing steps reduce to self-citation chains, self-definitions, or fitted inputs called predictions. All referenced foundations are external citations, making the work self-contained against external benchmarks with no internal circularity possible.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

This is a review paper with no new derivations, so the ledger contains no free parameters, axioms, or invented entities introduced by the authors.

pith-pipeline@v0.9.0 · 5654 in / 987 out tokens · 26741 ms · 2026-05-23T21:10:53.994863+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

21 extracted references · 21 canonical work pages · 5 internal anchors

  1. [1]

    An approximate Fourier transform useful in quantum factoring

    (1) Coppersmith, D. An approximate Fourier transform useful in quantum factoring. 2002; http://arxiv.org/abs/quant-ph/0201067, arXiv:quant-ph/0201067. 46 (2) Kitaev, A. Y. Quantum measurements and the Abelian Stabilizer Problem. 1995; http://arxiv.org/abs/quant-ph/9511026, arXiv:quant-ph/9511026. (3) Shor, P. Algorithms for quantum computation: discrete l...

    work page internal anchor Pith review Pith/arXiv arXiv 2002

  2. [2]

    S.; Chow, J

    (7) Moll, N.; Barkoutsos, P.; Bishop, L. S.; Chow, J. M.; Cross, A.; Egger, D. J.; Filipp, S.; Fuhrer, A.; Gambetta, J. M.; Ganzhorn, M.; others Quantum optimization using variational algorithms on near-term quantum devices. Quantum Sci. Technol. 2018, 3, 030503. (8) McArdle, S.; Jones, T.; Endo, S.; Li, Y.; Benjamin, S. C.; Yuan, X. Variational ansatz- b...

    work page 2018

  3. [3]

    F.; Wang, C.-Z.; Ho, K.-M.; Orth, P

    (10) Gomes, N.; Zhang, F.; Berthusen, N. F.; Wang, C.-Z.; Ho, K.-M.; Orth, P. P.; Yao, Y. 47 Efficient Step-Merged Quantum Imaginary Time Evolution Algorithm for Quantum Chemistry. J. Chem. Theory Comput. 2020, 16, 6256–6266. (11) Motta, M.; Sun, C.; Tan, A. T. K.; O’Rourke, M. J.; Ye, E.; Minnich, A. J.; Brand˜ ao, F. G. S. L.; Chan, G. K.-L. Determining...

    work page arXiv 2020

  4. [4]

    (23) Varadi, M.; Velankar, S

    Nature 2024, 630, 493–500. (23) Varadi, M.; Velankar, S. The impact of AlphaFold Protein Structure Database on the fields of life sciences. PROTEOMICS 2023, 23, 2200128. (24) Niazi, S. K.; Mariam, Z. Recent Advances in Machine-Learning-Based Chemoinfor- matics: A Comprehensive Review. International Journal of Molecular Sciences 2023, 24, 11488. (25) Paul,...

    work page arXiv 2024

  5. [5]

    (50) Beer, K.; Khosla, M.; K¨ ohler, J.; Osborne, T. J. Quantum machine learning of graph-structured data. 2021; http://arxiv.org/abs/2103.10837, arXiv:2103.10837 [quant-ph]. (51) Liao, Y.; Zhang, X.-M.; Ferrie, C. Graph Neural Networks on Quantum Computers. 2024; http://arxiv.org/abs/2405.17060, arXiv:2405.17060 [quant-ph]. (52) Mernyei, P.; Meichanetzid...

    work page doi:10.1101/2023.08.06.552202 2021

  6. [6]

    C.; Huang, H.-Y.; Cerezo, M.; Sharma, K.; Sornborger, A.; Cincio, L.; Coles, P

    (58) Caro, M. C.; Huang, H.-Y.; Cerezo, M.; Sharma, K.; Sornborger, A.; Cincio, L.; Coles, P. J. Generalization in quantum machine learning from few training data. Na- ture Communications 2022, 13,

    work page 2022

  7. [7]

    Hybrid Quantum Graph Neural Network for Molecular Property Prediction

    (59) Vitz, M.; Mohammadbagherpoor, H.; Sandeep, S.; Vlasic, A.; Padbury, R.; Pham, A. Hybrid Quantum Graph Neural Network for Molecular Property Prediction. 2024; http://arxiv.org/abs/2405.05205, arXiv:2405.05205 [cond-mat, physics:quant-ph] version:

    work page arXiv 2024

  8. [8]

    Simplifying Graph Convolutional Networks

    (60) Dong, L.; Li, Y.; Liu, D.; Ji, Y.; Hu, B.; Shi, S.; Zhang, F.; Hu, J.; Qian, K.; Jin, X.; Wang, B. Prediction of Protein-Ligand Binding Affinity by a Hybrid Quantum-Classical Deep Learning Algorithm. Advanced Quantum Technologies2023, 6, 2300107. (61) Wu, F.; Zhang, T.; Souza Jr., A. H. d.; Fifty, C.; Yu, T.; Weinberger, K. Q. Sim- plifying Graph Con...

    work page internal anchor Pith review Pith/arXiv arXiv 2019

  9. [9]

    H.; Wood, K

    (69) Liu, J.; Lim, K. H.; Wood, K. L.; Huang, W.; Guo, C.; Huang, H.-L. Hybrid quantum- classical convolutional neural networks. Science China Physics, Mechanics & Astron- omy 2021, 64, 290311. (70) MacCormack, I.; Delaney, C.; Galda, A.; Aggarwal, N.; Narang, P. Branching quantum convolutional neural networks. Physical Review Research 2022, 4, 013117. (7...

    work page 2021

  10. [10]

    A Tutorial on Quantum Convolutional Neural Networks (QCNN)

    (72) Oh, S.; Choi, J.; Kim, J. A Tutorial on Quantum Convolutional Neural Networks (QCNN). 2020 International Conference on Information and Communication Tech- nology Convergence (ICTC). Jeju, Korea (South), 2020; pp 236–239. (73) https://pennylane.ai/qml/demos/tutorial_quanvolution/. 54 (74) https://qiskit-community.github.io/qiskit-machine-learning/tuto...

    work page arXiv 2020

  11. [11]

    V.; Aliper, A.; Zhavoronkov, A

    (98) Putin, E.; Asadulaev, A.; Vanhaelen, Q.; Ivanenkov, Y.; Aladinskaya, A. V.; Aliper, A.; Zhavoronkov, A. Adversarial threshold neural computer for molecular de novo design. Molecular pharmaceutics 2018, 15, 4386–4397. (99) Dallaire-Demers, P.-L.; Killoran, N. Quantum generative adversarial networks. Phys- ical Review A 2018, 98, 012324. 57 (100) Romer...

    work page arXiv 2018

  12. [12]

    Attention Is All You Need

    (107) Vaswani, A.; Shazeer, N.; Parmar, N.; Uszkoreit, J.; Jones, L.; Gomez, A. N.; Kaiser, L.; Polosukhin, I. Attention Is All You Need. 2023; http://arxiv.org/abs/ 1706.03762, arXiv:1706.03762 [cs]. 58 (108) Wang, J.; Hsieh, C.-Y.; Wang, M.; Wang, X.; Wu, Z.; Jiang, D.; Liao, B.; Zhang, X.; Yang, B.; He, Q.; others Multi-constraint molecular generation ...

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  13. [13]

    Kernel-elastic autoencoder for molecular design

    (112) Li, H.; Shee, Y.; Allen, B.; Maschietto, F.; Morgunov, A.; Batista, V. Kernel-elastic autoencoder for molecular design. PNAS Nexus 2024, 3, pgae168. (113) Xue, C.; Chen, Z.-Y.; Zhuang, X.-N.; Wang, Y.-J.; Sun, T.-P.; Wang, J.-C.; Liu, H.- Y.; Wu, Y.-C.; Wang, Z.-L.; Guo, G.-P. End-to-End Quantum Vision Transformer: Towards Practical Quantum Speedup ...

    work page arXiv 2024

  14. [14]

    V.; Sosnin, S

    (117) Khokhlov, I.; Krasnov, L.; Fedorov, M. V.; Sosnin, S. Image2SMILES: Transformer- Based Molecular Optical Recognition Engine**. Chemistry–Methods 2022, 2, e202100069. (118) Ding, W.; Chen, H.; Ji, H. Efficiently Predicting Reaction Rates and Revealing Reac- tive Sites with a Molecular Image-Vision Transformer and Fukui Function Validation. Industrial...

    work page arXiv 2022

  15. [15]

    Qudit machine learning.Machine Learn- ing: Science and Technology 2024, 5, 015057

    (135) Roca-Jerat, S.; Rom´ an-Roche, J.; Zueco, D. Qudit machine learning.Machine Learn- ing: Science and Technology 2024, 5, 015057. (136) Mandilara, A.; Dellen, B.; Jaekel, U.; Valtinos, T.; Syvridis, D. Classification of data with a qudit, a geometric approach. Quantum Machine Intelligence 2024, 6,

    work page 2024

  16. [16]

    M.; Jiang, L.; Mirrahimi, M.; Devoret, M

    (137) Ofek, N.; Petrenko, A.; Heeres, R.; Reinhold, P.; Leghtas, Z.; Vlastakis, B.; Liu, Y.; Frunzio, L.; Girvin, S. M.; Jiang, L.; Mirrahimi, M.; Devoret, M. H.; Schoelkopf, R. J. Extending the lifetime of a quantum bit with error correction in superconducting circuits. Nature 2016, 536, 441–445. (138) Sivak, V. V.; Eickbusch, A.; Royer, B.; Singh, S.; T...

    work page 2016

  17. [17]

    P.; Lyu, N.; Wang, C.; Batista, V

    (150) Dutta, R.; Vu, N. P.; Lyu, N.; Wang, C.; Batista, V. S. Simulating electronic structure on bosonic quantum computers. 2024; https://arxiv.org/abs/2404.10222. 63 (151) Wang, C. S.; Curtis, J. C.; Lester, B. J.; Zhang, Y.; Gao, Y. Y.; Freeze, J.; Batista, V. S.; Vaccaro, P. H.; Chuang, I. L.; Frunzio, L.; Jiang, L.; Girvin, S. M.; Schoelkopf, R. J. Ef...

    work page arXiv 2024

  18. [18]

    (167) Bayraktar, H. et al. cuQuantum SDK: A High-Performance Library for Accelerating Quantum Science. arXiv:2308.01999 [quant-ph] 2023, (168) https://github.com/cudaq-libraries/workshops/tree/2024-chem-review . 65 (169) https://quantumai.google/qsim/choose_hw. (170) Anoshin, M.; Sagingalieva, A.; Mansell, C.; Zhiganov, D.; Shete, V.; Pflitsch, M.; Mel- n...

    work page arXiv 2023

  19. [19]

    (183) Cohn, J.; Motta, M.; Parrish, R. M. Quantum Filter Diagonalization with Compressed Double-Factorized Hamiltonians. PRX Quantum 2021, 2, 040352. (184) https://github.com/cudaq-libraries/workshops/blob/2024-chem-review/ multi-GPU/mgpu_ghz.py. (185) https://www.nvidia.com/en-us/data-center/nvlink/. (186) https://docs.nvidia.com/cuda/cuda-runtime-api/in...

    work page internal anchor Pith review Pith/arXiv arXiv 2021

  20. [20]

    Larocca, S

    68 (204) Larocca, M.; Thanasilp, S.; Wang, S.; Sharma, K.; Biamonte, J.; Coles, P. J.; Cincio, L.; McClean, J. R.; Holmes, Z.; Cerezo, M. A Review of Barren Plateaus in Variational Quantum Computing. 2024; http://arxiv.org/abs/2405.00781, arXiv:2405.00781 [quant-ph, stat]. (205) Ragone, M.; Bakalov, B. N.; Sauvage, F.; Kemper, A. F.; Ortiz Marrero, C.; La...

    work page arXiv 2024

  21. [21]

    Quantum Circuit Learning

    (206) Mitarai, K.; Negoro, M.; Kitagawa, M.; Fujii, K. Quantum Circuit Learning. Physical Review A 2018, 98, 032309, arXiv:1803.00745 [quant-ph]. (207) Schuld, M.; Bergholm, V.; Gogolin, C.; Izaac, J.; Killoran, N. Evaluating analytic gra- dients on quantum hardware. Physical Review A 2019, 99, 032331, arXiv:1811.11184 [quant-ph]. (208) Abbas, A.; King, R...

    work page internal anchor Pith review Pith/arXiv arXiv 2018