Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:2610.04588

Variational Quantum Attention for Molecular Graph Learning

Yu-Cheng Lin, Yu-Chao Hsu, Tai-Yue Li, Nan-Yow Chen, Samuel Yen-Chi Chen

quant-phcs.AIcs.LGSubmitted Oct 3, 2026• Announced Oct 6, 2026• 0 citations
arXivHTMLPDF

Abstract

Molecular property prediction is central to computational drug discovery, where graph neural networks learn to weight neighboring atomic environments during message passing. Yet it remains unclear how variational quantum circuits alter learned attention behavior in molecular graphs. We introduce an edge-aware variational quantum attention mechanism for molecular graph learning, in which the receiving atom, neighboring atom, and connecting bond jointly determine the quantum attention state. Across five molecular property and bioactivity prediction tasks, QGAT achieves competitive performance relative to GATv2, with a consistent improvement on BBBP across all six evaluated circuit ansatzes. We further compare how the quantum and classical attention scores weight molecular structure beyond accuracy. In the Verubecestat BACE1 inhibitor series, QGAT achieves a higher Spearman correlation than GATv2 and assigns positive attributions to several structural changes consistent with reported structure-activity relationships (SARs). This case study shows that the two attention mechanisms can exhibit different prediction and attribution behavior across structurally related BACE1 analogues, while broader validation is required to determine how consistently these differences generalize across chemical series and targets. Circuit ablations further show that performance depends on the circuit design. Together, these results show that variational quantum attention can serve as a viable alternative molecular attention parameterization while inducing circuit- and chemistry-dependent behavior distinct from a matched classical scorer.

AI Summary

The paper asks whether variational quantum circuits can replace classical attention scorers in molecular graph networks while changing predictive and molecular-attribution behavior. Its edge-aware quantum graph attention network (QGAT) combines projected receiving-atom, neighboring-atom, and bond features, amplitude-encodes the resulting 64- or 128-dimensional vector into 6 or 7 qubits, and uses scaled Pauli-$Z$ expectation values as attention logits. Message passing, aggregation, and prediction remain classical. Evaluation covers five molecular classification and regression datasets using a fixed 10% scaffold-held-out test set and five training/validation scaffold folds. Circuits are trained using exact state-vector simulation. The strongest evaluated QGAT configuration per task is competitive with a matched GATv2 baseline: BBBP ROC-AUC increases from $0.6808\pm0.0111$ to $0.7157\pm0.0095$, with all six tested ansatzes exceeding the baseline mean. Other differences are smaller or mixed; FreeSolv slightly favors GATv2. QGAT uses approximately 0.5–1.3% fewer trainable parameters. Circuit ablations show task-dependent performance, with optimal depths of one to three blocks rather than monotonic improvement with depth. Substructure masking, attention visualization, and integrated gradients characterize differences in molecular weighting. On the external Verubecestat BACE1 inhibitor series, QGAT achieves Spearman correlation 0.8264 versus 0.7954, but worse RMSE. Several attribution directions agree with reported structure–activity relationships, while both models miss others. Stated limitations include ligand-only inputs, no hardware evaluation, and primarily qualitative chemical interpretation centered on one series; the results do not establish generally superior interpretability.

Little sign of AI-generated textIncrementalMissing comparison

Error found: The headline prediction table reports the strongest observed QGAT configuration for each task without documenting independent validation-based configuration selection, leaving its best-configuration gains vulnerable to selection bias.

Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.

Related papers

Quantum Graph Attention Networks: Trainable Quantum Encoders for Inductive Graph Learning

Arthur Mendonça Faria, Mehdi Djellabi, Igor O. Sokolov +1 more

We introduce Quantum Graph Attention Networks (QGATs) as trainable quantum encoders for inductive learning on graphs, extending the Quantum Graph Neural Networks (QGNN) framework. QGATs leverage parameterized quantum circuits to encode node features and neighborhood structures, with quantum attention mechanisms modulating the contribution of each neighbor via dynamically learned unitaries. This allows for expressive, locality-aware quantum representations that can generalize across unseen graph instances. We evaluate our approach on the QM9 dataset, targeting the prediction of various chemical properties. Our experiments compare classical and quantum graph neural networks-with and without attention layers-demonstrating that attention consistently improves performance in both paradigms. Notably, we observe that quantum attention yields increasing benefits as graph size grows, with QGATs significantly outperforming their non-attentive quantum counterparts on larger molecular graphs. Furthermore, for smaller graphs, QGATs achieve predictive accuracy comparable to classical GAT models, highlighting their viability as expressive quantum encoders. These results show the potential of quantum attention mechanisms to enhance the inductive capacity of QGNN in chemistry and beyond.

arXiv:2509.11390•Sep 14, 2025•2 citations

Hybrid Quantum Graph Neural Networks for Molecular Toxicity Classification

Anurag Rai, G Nithish, Dasiga Venkata Ashish Kumar +1 more

Drug discovery depends heavily on predicting molecular toxicity early and classical ML still struggles here, because the interactions that make a molecule toxic are fundamentally quantum-mechanical. This paper builds a Hybrid Quantum Graph Neural Network (QGNN) that feeds GCN-learned molecular embeddings into a Variational Quantum Circuit (VQC). A 3-layer GCN produces graphlevel embeddings, which are angle-encoded into 4-8 qubit quantum states, and Pauli-Z expectations are concatenated with classical features before the final classifier. A quantum edge embedding is also introduced, encoding bondtype information directly into entangling gate parameters. On MoleculeNet's Tox21 (7,831 compounds, 12 endpoints) and ClinTox (1,478 compounds), 5-fold cross-validation shows the hybrid 4-qubit model with edge embedding achieves 0.762 ROC-AUC on Tox21 (+3.1% over 0.739 classical baseline), while the 8-qubit hybrid model reaches 0.741 on ClinTox (+6.3% over 0.697 baseline), both with batch size 64. Smaller batches (64) consistently outperform larger batches (256) at$3-17 \times$computational cost. Quantum edge embedding yields its largest improvement with only 15 additional parameters: +16.4% on ClinTox (batch 64). Quantum-only models achieve 83.7% parameter reduction while landing consistently just below the classical baseline. Across five folds, however, every one of these differences falls inside overlapping 95% confidence intervals$(p>0.05)$, so the gains reported here are consistent in direction but not individually significant at this sample size. Total overhead for hybrid models is 10-13% extra parameters with training time ranging from 1.2 to 54.6 seconds per epoch. External validation on the held-out BBBP benchmark (2,039 compounds), performed with the same architecture and no retuning, confirms that the gain transfers: the hybrid 4-qubit model reaches 0.702 ROC-AUC against 0.649 for the classical baseline (+8.2%), the largest relative gain across the three datasets. A mechanistic interpretability analysis attributes the edge-embedding effect to bond order reaching the model only through the entangling gates, where a learnable convex combination of perbond-type angles tunes entanglement continuously instead of pinning it at the CNOT maximum; the effect is positive on reactivity-driven toxicity endpoints and negative on passive permeability, matching the chemistry of each target.

International Conference on Circuits, Control, Communication and Computing (I4C)•Sep 1, 2026•0 citations

Quantum drug discovery: a hybrid quantum graph neural network–variational quantum eigensolver framework for serine neutralization

Javier Villalba-Díez

Abstract Quantum computing has emerged as a powerful tool for modeling molecular systems with high accuracy, accelerating the search for novel therapeutic targets. In this study, we present a two-stage hybrid workflow that combines a Quantum Graph Neural Network with a Variational Quantum Eigensolver to identify potent serine neutralizers in the QM9 dataset. In the first stage, an advanced quantum graph neural network architecture incorporating attention layers, self-distillation, and an adaptive learning-rate schedule is trained to predict ionization potentials and binding free energies. Across five independent random-seed trials, our adaptive-thresholded QGNN–VQE pipeline achieves an average $$R^2$$ R 2 of $$0.990 \pm 0.008$$ 0.990 ± 0.008 and a mean absolute error of $$0.034 \pm 0.001$$ 0.034 ± 0.001 eV ( $$\approx 0.79 \pm 0.03$$ ≈ 0.79 ± 0.03 kcal/mol) on the QM9 validation set, demonstrating robust, chemical-accuracy-level predictions. Building on these high-precision predictions, the second stage employs a QAOA-inspired hybrid ranking scheme that merges quantum graph neural network outputs, feature-space similarity (via and cosine similarity), and variational quantum eigensolver-derived energy stability. This $$\alpha $$ α -weighted ( $$\alpha = 0.95$$ α = 0.95 ) scoring framework is applied to over 133,000 molecules, efficiently parsed in parallel. The final ranked list is validated through an improved bijection test, featuring an adaptive similarity threshold to ensure both structural and functional robustness. The top-scoring candidate identified by our pipeline is 5,6,7-tetrahydro-4 H-pyrazolo[4,3-c]pyridin-4-one, highlighting the efficacy of quantum-enhanced modeling in pinpointing complex pharmacological targets. These findings underscore the transformative potential of hybrid quantum-classical workflows in drug discovery, offering a flexible blueprint for advanced screening of other critical biomolecular interactions in the rapidly expanding field of quantum biology. GraphicAbstract VQE for Serine Neutralization

The European Physical Journal D•Jul 1, 2025•8 citations

Hybrid VQE–Graph Neural Network Framework for Scalable Binding-Affinity Prediction

Neeraja P, J Umamageswaran

The Variational Quantum Eigensolver (VQE) has emerged as a viable method for calculating molecular energies through its combination of quantum and classical computing power on NISQ systems. The existing methods for predicting binding affinities in intricate systems face difficulties because of two major problems which are noise issues and the inability to handle large systems. The Hybrid VQE-Graph Neural Network (VQE-GNN) solution which we present enables us to achieve accurate binding affinity predictions by using quantum molecular feature learning together with graph neural network technology. Our method combines Hamiltonian Fragmentation with a noise-resilient VQE approach and an attention-steered Graph Neural Network. The research project aims to create a solution which combines precise quantum chemistry results with practical machine learning models. The experiments which researchers conducted on both the PDBbind and BindingDB datasets demonstrated that our approach accomplished better prediction results together with improved correlation and ligand ranking performance when compared to standard GNNs which proves that quantum-classical hybrid systems hold considerable promise for drug discovery research.

International Conference on Computational Robotics, Testing and Engineering Evaluation (ICCRTEE)•May 14, 2026•0 citations

Quantum-Enhanced Graph Neural Networks for Complex Decision Making in Cognitive Autonomous Systems

Sai Krishna Thota

Graph Neural Networks (GNNs) are now one of the key tools to model relational and non-Euclidean data in multiagent systems, infrastructure networks and even molecular modeling applications. Nevertheless, in very dense or noisy graph topologies, classical message-passing architectures may experience scalability and representational degeneration, especially due to over-smoothing and insufficient higher-order dependency information. The recent developments in quantum machine learning indicate that hybrid quantum-classical models can also provide an alternative feature representation by projecting relational data to parameterized quantum circuits. This article presents a Quantum-Enhanced Graph Neural Network (QGNN) architecture that extends variational quantum circuits with graph learning pipelines, based on classical learning. Parameters are represented as parameterized qubits, and relational dependencies are represented as structured entanglement mappings. We benchmark the regimes where quantum-enhanced relational encoding proves to be more robust and stable through the swarm coordination, molecular property prediction, and financial transaction network. We also elaborate on the issue of scalability, hardware issues, and practical aspects of near-term hybrid implementations and give a critical view of the changing role of quantum methods in computational intelligence systems.

•Feb 27, 2026•0 citations

A Hybrid Quantum-Classical Machine Learning Framework for Accelerated Drug Discovery

Sriramadasu Sri Sainadh, Saigeetha S.B., Jeeru Bharath Reddy +4 more

The pharmaceutical industry faces mounting challenges in drug discovery, where exploring a chemical space exceeding 10 60 possible compounds remains computationally prohibitive. Traditional in-silico screening approaches struggle to capture the quantum mechanical interactions that determine molecular behavior. To overcome these limitations, this study introduces a hybrid quantum–classical machine learning framework that uniquely integrates molecular graph theory with quantum-enhanced learning for accelerated drug discovery. Using the QM9 benchmark dataset containing 133,885 molecules of low-to-medium structural complexity, we extract Laplacian eigenvalue spectra as topological descriptors that preserve both connectivity and electronic information. These features are jointly processed by Random Forest (RF), Support Vector Machine (SVM), and a Variational Quantum Classifier (VQC) in a weighted ensemble architecture designed for balanced generalization. The Quantum Graph Neural Network (QGNN) component achieves 97.5% accuracy, while the hybrid ensemble attains 100.0% accuracy, confirming strong predictive reliability. Although quantum training requires roughly 3.2× longer computation, it captures high-dimensional correlations that classical methods overlook. Overall, the proposed framework bridges the gap between theoretical quantum computing and practical pharmaceutical modeling, providing a scalable, physics-inspired pathway toward intelligent molecular screening and next-generation drug discovery systems.

Procedia Computer Science•Jan 1, 2026•0 citations
Molecular property prediction is central to computational drug discovery, where graph neural networks learn to weight neighboring atomic environments during message passing. Yet it remains unclear how variational quantum circuits alter learned attention behavior in molecular graphs. We introduce an edge-aware variational quantum attention mechanism for molecular graph learning, in which the receiving atom, neighboring atom, and connecting bond jointly determine the quantum attention state. Across five molecular property and bioactivity prediction tasks, QGAT achieves competitive performance relative to GATv2, with a consistent improvement on BBBP across all six evaluated circuit ansatzes. We further compare how the quantum and classical attention scores weight molecular structure beyond accuracy. In the Verubecestat BACE1 inhibitor series, QGAT achieves a higher Spearman correlation than GATv2 and assigns positive attributions to several structural changes consistent with reported structure-activity relationships (SARs). This case study shows that the two attention mechanisms can exhibit different prediction and attribution behavior across structurally related BACE1 analogues, while broader validation is required to determine how consistently these differences generalize across chemical series and targets. Circuit ablations further show that performance depends on the circuit design. Together, these results show that variational quantum attention can serve as a viable alternative molecular attention parameterization while inducing circuit- and chemistry-dependent behavior distinct from a matched classical scorer.
The paper asks whether variational quantum circuits can replace classical attention scorers in molecular graph networks while changing predictive and molecular-attribution behavior. Its edge-aware quantum graph attention network (QGAT) combines projected receiving-atom, neighboring-atom, and bond features, amplitude-encodes the resulting 64- or 128-dimensional vector into 6 or 7 qubits, and uses scaled Pauli-ZZZ expectation values as attention logits. Message passing, aggregation, and prediction remain classical.
Evaluation covers five molecular classification and regression datasets using a fixed 10% scaffold-held-out test set and five training/validation scaffold folds. Circuits are trained using exact state-vector simulation. The strongest evaluated QGAT configuration per task is competitive with a matched GATv2 baseline: BBBP ROC-AUC increases from 0.6808±0.01110.6808\pm0.01110.6808±0.0111 to 0.7157±0.00950.7157\pm0.00950.7157±0.0095, with all six tested ansatzes exceeding the baseline mean. Other differences are smaller or mixed; FreeSolv slightly favors GATv2. QGAT uses approximately 0.5–1.3% fewer trainable parameters. Circuit ablations show task-dependent performance, with optimal depths of one to three blocks rather than monotonic improvement with depth.
Substructure masking, attention visualization, and integrated gradients characterize differences in molecular weighting. On the external Verubecestat BACE1 inhibitor series, QGAT achieves Spearman correlation 0.8264 versus 0.7954, but worse RMSE. Several attribution directions agree with reported structure–activity relationships, while both models miss others. Stated limitations include ligand-only inputs, no hardware evaluation, and primarily qualitative chemical interpretation centered on one series; the results do not establish generally superior interpretability.
We introduce Quantum Graph Attention Networks (QGATs) as trainable quantum encoders for inductive learning on graphs, extending the Quantum Graph Neural Networks (QGNN) framework. QGATs leverage parameterized quantum circuits to encode node features and neighborhood structures, with quantum attention mechanisms modulating the contribution of each neighbor via dynamically learned unitaries. This allows for expressive, locality-aware quantum representations that can generalize across unseen graph instances. We evaluate our approach on the QM9 dataset, targeting the prediction of various chemical properties. Our experiments compare classical and quantum graph neural networks-with and without attention layers-demonstrating that attention consistently improves performance in both paradigms. Notably, we observe that quantum attention yields increasing benefits as graph size grows, with QGATs significantly outperforming their non-attentive quantum counterparts on larger molecular graphs. Furthermore, for smaller graphs, QGATs achieve predictive accuracy comparable to classical GAT models, highlighting their viability as expressive quantum encoders. These results show the potential of quantum attention mechanisms to enhance the inductive capacity of QGNN in chemistry and beyond.
Drug discovery depends heavily on predicting molecular toxicity early and classical ML still struggles here, because the interactions that make a molecule toxic are fundamentally quantum-mechanical. This paper builds a Hybrid Quantum Graph Neural Network (QGNN) that feeds GCN-learned molecular embeddings into a Variational Quantum Circuit (VQC). A 3-layer GCN produces graphlevel embeddings, which are angle-encoded into 4-8 qubit quantum states, and Pauli-Z expectations are concatenated with classical features before the final classifier. A quantum edge embedding is also introduced, encoding bondtype information directly into entangling gate parameters. On MoleculeNet's Tox21 (7,831 compounds, 12 endpoints) and ClinTox (1,478 compounds), 5-fold cross-validation shows the hybrid 4-qubit model with edge embedding achieves 0.762 ROC-AUC on Tox21 (+3.1% over 0.739 classical baseline), while the 8-qubit hybrid model reaches 0.741 on ClinTox (+6.3% over 0.697 baseline), both with batch size 64. Smaller batches (64) consistently outperform larger batches (256) at3−17×3-17 \times3−17×computational cost. Quantum edge embedding yields its largest improvement with only 15 additional parameters: +16.4% on ClinTox (batch 64). Quantum-only models achieve 83.7% parameter reduction while landing consistently just below the classical baseline. Across five folds, however, every one of these differences falls inside overlapping 95% confidence intervals(p>0.05)(p>0.05)(p>0.05), so the gains reported here are consistent in direction but not individually significant at this sample size. Total overhead for hybrid models is 10-13% extra parameters with training time ranging from 1.2 to 54.6 seconds per epoch. External validation on the held-out BBBP benchmark (2,039 compounds), performed with the same architecture and no retuning, confirms that the gain transfers: the hybrid 4-qubit model reaches 0.702 ROC-AUC against 0.649 for the classical baseline (+8.2%), the largest relative gain across the three datasets. A mechanistic interpretability analysis attributes the edge-embedding effect to bond order reaching the model only through the entangling gates, where a learnable convex combination of perbond-type angles tunes entanglement continuously instead of pinning it at the CNOT maximum; the effect is positive on reactivity-driven toxicity endpoints and negative on passive permeability, matching the chemistry of each target.
Abstract Quantum computing has emerged as a powerful tool for modeling molecular systems with high accuracy, accelerating the search for novel therapeutic targets. In this study, we present a two-stage hybrid workflow that combines a Quantum Graph Neural Network with a Variational Quantum Eigensolver to identify potent serine neutralizers in the QM9 dataset. In the first stage, an advanced quantum graph neural network architecture incorporating attention layers, self-distillation, and an adaptive learning-rate schedule is trained to predict ionization potentials and binding free energies. Across five independent random-seed trials, our adaptive-thresholded QGNN–VQE pipeline achieves an average R2R^2R2 R 2 of 0.990±0.0080.990 \pm 0.0080.990±0.008 0.990 ± 0.008 and a mean absolute error of 0.034±0.0010.034 \pm 0.0010.034±0.001 0.034 ± 0.001 eV ( ≈0.79±0.03\approx 0.79 \pm 0.03≈0.79±0.03 ≈ 0.79 ± 0.03 kcal/mol) on the QM9 validation set, demonstrating robust, chemical-accuracy-level predictions. Building on these high-precision predictions, the second stage employs a QAOA-inspired hybrid ranking scheme that merges quantum graph neural network outputs, feature-space similarity (via and cosine similarity), and variational quantum eigensolver-derived energy stability. This α\alpha α α -weighted ( α=0.95\alpha = 0.95α=0.95 α = 0.95 ) scoring framework is applied to over 133,000 molecules, efficiently parsed in parallel. The final ranked list is validated through an improved bijection test, featuring an adaptive similarity threshold to ensure both structural and functional robustness. The top-scoring candidate identified by our pipeline is 5,6,7-tetrahydro-4 H-pyrazolo[4,3-c]pyridin-4-one, highlighting the efficacy of quantum-enhanced modeling in pinpointing complex pharmacological targets. These findings underscore the transformative potential of hybrid quantum-classical workflows in drug discovery, offering a flexible blueprint for advanced screening of other critical biomolecular interactions in the rapidly expanding field of quantum biology. GraphicAbstract VQE for Serine Neutralization
The Variational Quantum Eigensolver (VQE) has emerged as a viable method for calculating molecular energies through its combination of quantum and classical computing power on NISQ systems. The existing methods for predicting binding affinities in intricate systems face difficulties because of two major problems which are noise issues and the inability to handle large systems. The Hybrid VQE-Graph Neural Network (VQE-GNN) solution which we present enables us to achieve accurate binding affinity predictions by using quantum molecular feature learning together with graph neural network technology. Our method combines Hamiltonian Fragmentation with a noise-resilient VQE approach and an attention-steered Graph Neural Network. The research project aims to create a solution which combines precise quantum chemistry results with practical machine learning models. The experiments which researchers conducted on both the PDBbind and BindingDB datasets demonstrated that our approach accomplished better prediction results together with improved correlation and ligand ranking performance when compared to standard GNNs which proves that quantum-classical hybrid systems hold considerable promise for drug discovery research.
Graph Neural Networks (GNNs) are now one of the key tools to model relational and non-Euclidean data in multiagent systems, infrastructure networks and even molecular modeling applications. Nevertheless, in very dense or noisy graph topologies, classical message-passing architectures may experience scalability and representational degeneration, especially due to over-smoothing and insufficient higher-order dependency information. The recent developments in quantum machine learning indicate that hybrid quantum-classical models can also provide an alternative feature representation by projecting relational data to parameterized quantum circuits. This article presents a Quantum-Enhanced Graph Neural Network (QGNN) architecture that extends variational quantum circuits with graph learning pipelines, based on classical learning. Parameters are represented as parameterized qubits, and relational dependencies are represented as structured entanglement mappings. We benchmark the regimes where quantum-enhanced relational encoding proves to be more robust and stable through the swarm coordination, molecular property prediction, and financial transaction network. We also elaborate on the issue of scalability, hardware issues, and practical aspects of near-term hybrid implementations and give a critical view of the changing role of quantum methods in computational intelligence systems.
The pharmaceutical industry faces mounting challenges in drug discovery, where exploring a chemical space exceeding 10 60 possible compounds remains computationally prohibitive. Traditional in-silico screening approaches struggle to capture the quantum mechanical interactions that determine molecular behavior. To overcome these limitations, this study introduces a hybrid quantum–classical machine learning framework that uniquely integrates molecular graph theory with quantum-enhanced learning for accelerated drug discovery. Using the QM9 benchmark dataset containing 133,885 molecules of low-to-medium structural complexity, we extract Laplacian eigenvalue spectra as topological descriptors that preserve both connectivity and electronic information. These features are jointly processed by Random Forest (RF), Support Vector Machine (SVM), and a Variational Quantum Classifier (VQC) in a weighted ensemble architecture designed for balanced generalization. The Quantum Graph Neural Network (QGNN) component achieves 97.5% accuracy, while the hybrid ensemble attains 100.0% accuracy, confirming strong predictive reliability. Although quantum training requires roughly 3.2× longer computation, it captures high-dimensional correlations that classical methods overlook. Overall, the proposed framework bridges the gap between theoretical quantum computing and practical pharmaceutical modeling, providing a scalable, physics-inspired pathway toward intelligent molecular screening and next-generation drug discovery systems.