RECORD
Continuous-Time Quantum Walks based Graph Neural Network
Artificial Intelligence · Submitted 21 Aug 2026
abstract compiled · 2026-09-23 · no language-model review
The authors' main claim We propose \textbf{CTQW-GNN}, a GNN based on Continuous-Time Quantum Walks (CTQW), to address both issues with theoretical justification.
The compiled record
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
Open the compiled record
The abstract, sentence by sentence
Graph Neural Networks (GNNs) are widely used on graph-structured data, but most suffer from two key weaknesses. First, message passing behaves as a low-pass filter under the homophily assumption, leading to poor performance on heterophilic graphs. Second, stacking layers drives node features toward constants, causing over-smoothing. Existing methods usually address these issues separately, while the few joint solutions rely largely on empirical heuristics, and many over-smoothing remedies sacrifice model expressiveness. We propose \textbf{CTQW-GNN}, a GNN based on Continuous-Time Quantum Walks (CTQW), to address both issues with theoretical justification. Its design exploits two properties of the CTQW propagator $e^{-\mathrm{i}Ht}$. First, it is unitary and has eigenvalues on the unit circle, so no frequency component is damped, counteracting the low-pass bias. Second, unitarity preserves feature norms and prevents the Dirichlet energy from decaying exponentially with depth, thereby mitigating over-smoothing. CTQW-GNN combines three complementary aggregation modules. \textit{CTQW-based Aggregation} evolves node features through the unitary propagator, preserving mid- and high-frequency signals for heterophilic graphs while preventing Dirichlet-energy collapse. \textit{CTQW-Attention Aggregation} constructs a multi-hop neighbor graph from CTQW amplitudes and applies attention over it, enabling access to distant homophilic nodes missed by single-hop aggregation. \textit{LF Aggregation} uses a standard low-pass GAT branch to retain strong performance on homophilic graphs, where pure CTQW aggregation can be suboptimal. We further provide a spectral-gap analysis explaining energy preservation and a Lieb--Robinson-type bound that gives a principled rule for selecting the walk time $t$.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | Graph Neural Networks (GNNs) are widely used on graph-structured data, but most suffer from two key weaknesses. | background 0.97 | 0.05 | 0.02 | asserted 1.00 |
| s1 | First, message passing behaves as a low-pass filter under the homophily assumption, leading to poor performance on heterophilic graphs. | background 0.95 | 0.05 | 0.02 | asserted 1.00 |
| s2 | Second, stacking layers drives node features toward constants, causing over-smoothing. | background 0.98 | 0.05 | 0.03 | asserted 1.00 |
| s3 | Existing methods usually address these issues separately, while the few joint solutions rely largely on empirical heuristics, and many over-smoothing remedies sacrifice model expressiveness. | gap 0.89 | 0.05 | 0.03 | asserted 0.99 |
| s4 | We propose \textbf{CTQW-GNN}, a GNN based on Continuous-Time Quantum Walks (CTQW), to address both issues with theoretical justification. | contribution 0.99 | 0.93 | 0.35 | asserted 1.00 |
| s5 | Its design exploits two properties of the CTQW propagator $e^{-\mathrm{i}Ht}$. | method 0.89 | 0.72 | 0.06 | asserted 0.99 |
| s6 | First, it is unitary and has eigenvalues on the unit circle, so no frequency component is damped, counteracting the low-pass bias. | method 0.81 | 0.34 | 0.07 | asserted 1.00 |
| s7 | Second, unitarity preserves feature norms and prevents the Dirichlet energy from decaying exponentially with depth, thereby mitigating over-smoothing. | method 0.72 | 0.63 | 0.13 | asserted 1.00 |
| s8 | CTQW-GNN combines three complementary aggregation modules. | method 0.91 | 0.78 | 0.04 | asserted 1.00 |
| s9 | \textit{CTQW-based Aggregation} evolves node features through the unitary propagator, preserving mid- and high-frequency signals for heterophilic graphs while preventing Dirichlet-energy collapse. | method 0.97 | 0.83 | 0.14 | asserted 1.00 |
| s10 | \textit{CTQW-Attention Aggregation} constructs a multi-hop neighbor graph from CTQW amplitudes and applies attention over it, enabling access to distant homophilic nodes missed by single-hop aggregation. | method 0.96 | 0.88 | 0.05 | asserted 0.97 |
| s11 | \textit{LF Aggregation} uses a standard low-pass GAT branch to retain strong performance on homophilic graphs, where pure CTQW aggregation can be suboptimal. | method 0.90 | 0.47 | 0.05 | asserted 0.83 |
| s12 | We further provide a spectral-gap analysis explaining energy preservation and a Lieb--Robinson-type bound that gives a principled rule for selecting the walk time $t$. | contribution 0.87 | 0.94 | 0.08 | asserted 1.00 |
What was examined, and what was not
- arXiv listing abstract v1
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
The paper
Original abstract
Graph Neural Networks (GNNs) are widely used on graph-structured data, but most suffer from two key weaknesses. First, message passing behaves as a low-pass filter under the homophily assumption, leading to poor performance on heterophilic graphs. Second, stacking layers drives node features toward constants, causing over-smoothing. Existing methods usually address these issues separately, while the few joint solutions rely largely on empirical heuristics, and many over-smoothing remedies sacrifice model expressiveness. We propose \textbf{CTQW-GNN}, a GNN based on Continuous-Time Quantum Walks (CTQW), to address both issues with theoretical justification. Its design exploits two properties of the CTQW propagator $e^{-\mathrm{i}Ht}$. First, it is unitary and has eigenvalues on the unit circle, so no frequency component is damped, counteracting the low-pass bias. Second, unitarity preserves feature norms and prevents the Dirichlet energy from decaying exponentially with depth, thereby mitigating over-smoothing. CTQW-GNN combines three complementary aggregation modules. \textit{CTQW-based Aggregation} evolves node features through the unitary propagator, preserving mid- and high-frequency signals for heterophilic graphs while preventing Dirichlet-energy collapse. \textit{CTQW-Attention Aggregation} constructs a multi-hop neighbor graph from CTQW amplitudes and applies attention over it, enabling access to distant homophilic nodes missed by single-hop aggregation. \textit{LF Aggregation} uses a standard low-pass GAT branch to retain strong performance on homophilic graphs, where pure CTQW aggregation can be suboptimal. We further provide a spectral-gap analysis explaining energy preservation and a Lieb--Robinson-type bound that gives a principled rule for selecting the walk time $t$.
Human review
How the literature uses it
What this paper does with the arXiv papers it cites
What it cites · 3 sentences without a confident label
Uses the method or tool of 1 paper
-
Specifically, we adopt the attention mechanism from Graph Transformer [ref] to aggregate distant nodes via CTQW-induced multi-hop connectivity, thereby capturing long-range homophilic relationships.
Masked Label Prediction: Unified Message Passing Model for Semi-Supervised Classification · Introduction · 0.96
Uses data or values from 1 paper
-
We use five homophilic datasets (Cora [ref] , Citeseer [ref] , PubMed [ref] , Computers, Photo [ref] ) and nine heterophilic datasets: the classical four (Chameleon, Squirrel, Actor, Texas) [ref] and the five new datasets of [ref] (Roman-empire, Amazon-ratings, Minesweeper, Tolokers, Wiki-cooc).
Geom-GCN: Geometric Graph Convolutional Networks · Experimental Setup · 0.99
Cites as support 1 paper
-
However, previous study posits that while some methods can mitigate over-smoothing, this comes at the cost of sacrificing the expressive performance of GNNs [ref] .
A Survey on Oversmoothing in Graph Neural Networks · Introduction · 0.96
Cites as background 1 paper
-
Numerous methods have been proposed to mitigate it mainly from Drop Operation [ref] , Normalization [ref] , and modifying the dynamical systems of GNNs [ref] .
DropEdge: Towards Deep Graph Convolutional Networks on Node Classification · Related Work · 0.88
Discussion
Know something about this paper? Leave a comment, or highlight any passage above to comment on it.