Graph Neural Networks (GNNs) have emerged as a dominant paradigm for graph classification. Specifically, most existing GNNs mainly rely on the message passing strategy between neighbor nodes, where the expressivity is limited by the 1-dimensional Weisfeiler-Lehman (1-WL) test. Although a number of k-WL-based GNNs have been proposed to overcome this limitation, their computational cost increases rapidly with k, significantly restricting the practical applicability. Moreover, since the k-WL models mainly operate on node tuples, these k-WL-based GNNs cannot retain fine-grained node- or edge-level semantics required by attribution methods (e.g., Integrated Gradients), leading to the less interpretable problem. To overcome the above shortcomings, in this paper, we propose a novel Line Graph Aggregation Network (LGAN), that constructs a line graph from the induced subgraph centered at each node to perform the higher-order aggregation. We theoretically prove that the LGAN not only possesses the greater expressive power than the 2-WL under injective aggregation assumptions, but also has lower time complexity. Empirical evaluations on benchmarks demonstrate that the LGAN outperforms state-of-the-art k-WL-based GNNs, while offering better interpretability.

Chinese calligraphy is a quintessential element of Chinese cultural heritage. Analyzing and comparing calligraphic styles not only enhances the appreciation, learning, and advancement of calligraphy but also provides valuable insights into ancient China. However, such analysis remains challenging due to the limited scalability and possible inconsistencies of qualitative methods, as well as usability and misalignment issues in conventional quantitative approaches. We propose Calli-VA, a visual analytics system, to address these challenges. Calli-VA extracts character images and their corresponding strokes from original works and characterizes each character using systematic criteria. During analysis, the system defines the analysis scope by overview and uncovers relationships between characters. Explanation and recommendation mechanisms are integrated to help users understand patterns and guide further exploration. A documentation feature allows users to record and share their findings. We demonstrate the effectiveness of Calli-VA through three case studies and expert feedback.

Team-based combat scenarios are prevalent in various real-world applications like video gaming. Analyzing tactics in these scenarios is essential for gaining insights into game processes and improving combat behaviors. The decision-making data in team-based combat include character actions, movement trajectories, and event sequences. Existing studies face challenges in visualizing and analyzing combat tactics due to the complexity and the multifaceted characteristics of the decision-making data. To address these challenges, we introduce TactiVis, a visual analytics system designed for analyzing combat decision-making behavior. Using MOBA game as a representative case of team-based combat, TactiVis adopts a macro-to-micro tactics visual analytics framework consisting of three stages: match-level analysis, event-level understanding, and character-level comparison. In the TactiVis system, we introduce the v-storyline visualization, which encodes positions along the vertical axis to reveal tactical patterns. Case studies and a usability study demonstrate the utility and usability of TactiVis for helping analysts understand combat patterns and analyze tactics.

Automating the generation of multiple linked view visualization is imperative for improving data analysis efficiency. Large Language Models (LLMs) offer substantial potential for enabling this automation, yet they encounter notable challenges in understanding complex queries and producing relevant interactive visualizations. To tackle these challenges, we introduce SmartMLVs, a system designed to harness LLMs for automatic interactive multiple linked views generation with human guidance. First, we analyze the challenges LLMs may encounter when designing visualizations in place of experts. To address these challenges, we gather the essential domain knowledge required for visual analysis process and propose a framework consisting of decomposition, visualization and linking. The decomposition process applies a human-AI interaction method to clarify user requirements. For each decomposed question, the generation process handles chart type selection, data processing and visualization generation. Finally, the linking process adds interactions for views and provides users with data insights. For better human-AI collaboration, we design a system for data exploration. Our system applies the entire framework, supporting users’ interactive exploration with multiple linked views, and can iteratively generate linked views based on user feedback. We examine the effectiveness of our method through usage scenarios and evaluations.

In astronomical spectral analysis, class recognition is essential and fundamental for subsequent scientific research. The experts often perform the visual inspection after automatic classification to deal with low-quality spectra to improve accuracy. However, given the enormous spectral volume and inadequacy of the current inspection practice, such inspection is tedious and time-consuming. This article presents a visual analytics system named SpectrumVA to promote the efficiency of visual inspection while guaranteeing accuracy. We abstract inspection as a visual parameter space analysis process, using redshifts and spectral lines as parameters. Different navigation strategies are employed in the "selection-inspection-promotion" workflow. At the selection stage, we help the experts identify a spectrum of interest through spectral representations and auxiliary information. Several possible redshifts and corresponding important spectral lines are also recommended through a global-to-local strategy to provide an appropriate entry point for the inspection. The inspection stage adopts a variety of instant visual feedback to help the experts adjust the redshift and select spectral lines in an informed trial-and-error manner. Similar spectra to the inspected one rather than different ones are visualized at the promotion stage, making the inspection process more fluent. We demonstrate the effectiveness of SpectrumVA through a quantitative algorithmic assessment, a case study, interviews with domain experts, and a user study.

In Chinese archaeological research, analyzing the evolution of motifs in ancient pottery is crucial for studying the spread and growth of cultures across various eras and regions. However, such analyses are often challenging due to the complexities of identifying motifs with evolutionary connections that may manifest concurrent changes in appearance, space, and time, compounded by ineffective documentation. We propose PM-Vis, a visual analytics system for tracing and analyzing the evolution of pottery motifs. PM-Vis is anchored in a "selection-organization-documentation" workflow. In the selection stage, we design a three-fold projection paired with a motif-based search mechanism, displaying the appearance similarity and temporal and spatial proximities of all motifs or a specific motif, aiding users in selecting motifs with evolutionary connections. The organization stage helps users establish the evolutionary sequence and segment the selected motifs into distinct evolutionary phases. Finally, the documentation stage enables users to record their observations and insights through various forms of annotation. We demonstrate the usefulness and effectiveness of PM-Vis through two case studies, expert feedback, and a user study.