In this work, we propose two novel quantum walk kernels, namely the Hierarchical Aligned Quantum Jensen-Shannon Kernels (HAQJSK), between un-attributed graph structures. Different from most classical proposed HAQJSK kernels can incorporate hierarchical aligned structure information graphs and transform of random sizes into fixed-size structures, i.e., Transitive Adjacency Matrix vertices Density Continuous-Time Walks (CTQW). With pairwise to hand, resulting are defined by computing...

In this paper, we develop a novel backtrackless aligned-spatial graph convolutional network (BASGCN) model to learn effective features for classification. Our idea is transform arbitrary-sized graphs into fixed-sized aligned grid structures and define new spatial convolution operation associated with the structures. We show that proposed BASGCN not only reduces problems of information loss imprecise representation arising in existing spatially-based (GCN) models, but also bridges theoretical...

Network representations are powerful tools to modeling the dynamic time-varying financial complex systems consisting of multiple co-evolving time series, e.g., stock prices. In this work, we develop a novel framework compute kernel-based similarity measure between networks. Specifically, explore whether proposed kernel can be employed understand structural evolution networks with associated standard machines. For set each vertex representing individual series different and edge pair absolute...

This paper proposes a new Quantum Spatial Graph Convolutional Neural Network (QSGCNN) model that can directly learn classification function for graphs of arbitrary sizes. Unlike state-of-the-art (GCNN) models, the proposed QSGCNN incorporates process identifying transitive aligned vertices between and transforms sized into fixed-sized vertex grid structures. In order to representative graph characteristics, quantum spatial convolution is employed extract multi-scale features, in terms...

Graph-based representations are powerful tools for analyzing structured data. In this paper, we propose a novel model to learn Deep Hierarchical Attention-based Kernelized Representations (DHAKR) graph classification. To end, commence by learning an assignment matrix hierarchically map the substructure invariants into set of composite invariants, resulting in hierarchical kernelized graphs. Moreover, introduce feature-channel attention mechanism capture interdependencies between different...

Financial security is crucial for the development of Chinese economy owing to complex interconnections between financial market and real economy. This study employs generalized variance decomposition method construct a two-way risk spillover model China’s It investigates trans- 4 mission hedging in different markets when facing external shocks over past four years. Firstly, following shocks, both overall market’s total individual markets’ exhibit characteristics effects across various...

To ease the process of building Knowledge Graphs (KGs) from scratch, a cost-effective method is required to enrich KG using triples extracted corpus. However, it challenging with newly since they contain noisy information. This paper proposes refine by leveraging information In particular, we first formulate task KGs as two coupled sub-tasks, namely join event extraction and knowledge graph fusion. We then propose collaborative fusion framework, which composed an explorer supervisor, allow...

We develop a novel method for measuring the similarity between complete weighted graphs, which are probed by means of discrete-time quantum walks. Directly probing graphs using walks is intractable due to cost simulating walk. overcome this problem extracting commute time minimum spanning tree from graph. The walk initialized version Perron-Frobenius operator. This naturally encapsulates edge weight information extracted original For each pair be compared, we simulate on corresponding trees...

In recent years, with the increasing proportion of wind power generation, impact generation on grid security is also growing. This makes prediction accuracy higher and higher. paper utilizes LSTM model deep learning domain to predict generation. Besides, Auto Encoder employed reduce data dimension, improve generalization ability model, shorten training time. Simulation experiments show that has better than other machine such as SVM.

Graph convolutional networks (GCNs) are powerful tools for graph structure data analysis. One main drawback arising in most existing GCN models is that of the oversmoothing problem, i.e., vertex features abstracted from convolution operation have previously tended to be indistinguishable if model has many layers (e.g., more than two layers). To address this article, we propose a family aligned network (AVCN) focus on learning multiscale local-level vertices classification. This done by...

The direct conversion of CO2 and CH4 into value-added chemicals at room temperature atmospheric pressure poses a significant challenge in the chemical field. Nonthermal plasma (NTP) exhibits unique properties, enabling provision energy up to 10 eV temperature. However, active species NTP are complex, making control liquid products difficult. To address this, Cu can be introduced γ-Al2O3 through solid ion exchange method, followed by secondary calcination obtain nonprecious metal catalyst...

Graph Neural Networks (GNNs) are powerful tools for graph classification. One important operation GNNs is the downsampling or pooling that can learn effective embeddings from node representations. In this paper, we propose a new hierarchical operation, namely Edge-Node Attention-based Differentiable Pooling (ENADPool), to Unlike classical based on unclear assignment and simply computes averaged feature over nodes of each cluster, proposed ENADPool not only employs hard clustering strategy...