IEEE transactions on visualization and computer graphics最新文献

Understanding the behavior of large language models (LLMs) is crucial for ensuring their safe and reliable use. However, existing explainable AI (XAI) methods for LLMs primarily rely on word-level explanations, which are often computationally inefficient and misaligned with human reasoning processes. Moreover, these methods often treat explanation as a one-time output, overlooking its inherently interactive and iterative nature. In this paper, we present LLM Analyzer, an interactive visualization system that addresses these limitations by enabling intuitive and efficient exploration of LLM behaviors through counterfactual analysis. Our system features a novel algorithm that generates fluent and semantically meaningful counterfactuals via targeted removal and replacement operations at user-defined levels of granularity. These counterfactuals are used to compute feature attribution scores, which are then integrated with concrete examples in a table-based visualization, supporting dynamic analysis of model behavior. A user study with LLM practitioners and interviews with experts demonstrate the system's usability and effectiveness, emphasizing the importance of involving humans in the explanation process as active participants rather than passive recipients.

Overdraw is inevitable in large-scale scatterplots. Current scatterplot abstraction methods lose features in medium-to-low density regions. We propose a visual abstraction method designed to provide better feature preservation across arbitrary abstraction levels for large-scale scatterplots, particularly in medium-to-low density regions. The method consists of three closely interconnected steps: first, we partition the scatterplot into iso-density regions and equalize visual density; then, we allocate pixels for different classes within each region; finally, we reconstruct the data distribution based on pixels. User studies, quantitative and qualitative evaluations demonstrate that, compared to previous methods, our approach better preserves features and exhibits a special advantage when handling ultra-high dynamic range data distributions.

Connectomics, a subfield of neuroscience, aims to map and analyze synapse-level wiring diagrams of the nervous system. While recent advances in deep learning have accelerated automated neuron and synapse segmentation, reconstructing accurate connectomes still demands extensive human proofreading to correct segmentation errors. We present SynAnno, an interactive tool designed to streamline and enhance the proofreading of synaptic annotations in large-scale connectomics datasets. SynAnno integrates into existing neuroscience workflows by enabling guided, neuron-centric proofreading. To address the challenges posed by the complex spatial branching of neurons, it introduces a structured workflow with an optimized traversal path and a 3D mini-map for tracking progress. In addition, SynAnno incorporates fine-tuned machine learning models to assist with error detection and correction, reducing the manual burden and increasing proofreading efficiency. We evaluate SynAnno through a user and case study involving seven neuroscience experts. Results show that SynAnno significantly accelerates synapse proofreading while reducing cognitive load and annotation errors through structured guidance and visualization support. The source code and interactive demo are available at: https://github.com/PytorchConnectomics/SynAnno.

Random forests are a machine learning method used to automatically classify datasets and consist of a multitude of decision trees. While these random forests often have higher performance and generalize better than a single decision tree, they are also harder to interpret. This paper presents a visualization method and system to increase interpretability of random forests. We cluster similar trees which enables users to interpret how the model performs in general without needing to analyze each individual decision tree in detail, or interpret an oversimplified summary of the full forest. To meaningfully cluster the decision trees, we introduce a new distance metric that takes into account both the decision rules as well as the predictions of a pair of decision trees. We also propose two new visualization methods that visualize both clustered and individual decision trees: (1) The Feature Plot, which visualizes the topological position of features in the decision trees, and (2) the Rule Plot, which visualizes the decision rules of the decision trees. We demonstrate the efficacy of our approach through a case study on the "Glass" dataset, which is a relatively complex standard machine learning dataset, as well as a small user study.

Storyline visualizations represent character interactions over time. When these characters belong to different groups, a new research question emerges: how can we balance optimization of readability across the groups while preserving the overall narrative structure of the story? Traditional algorithms that optimize global readability metrics (like minimizing crossings) can introduce quality biases between the different groups based on their cardinality and other aspects of the data. Visual consequences of these biases are: making characters of minority groups disproportionately harder to follow, and visually deprioritizing important characters when their curves become entangled with numerous secondary characters. We present F2Stories, a modular framework that addresses these challenges in storylines by offering three complementary optimization modes: (1) fairnessMode ensures that no group bears a disproportionate burden of visualization complexity regardless of their representation in the story; (2) focusMode allows prioritizing a group of characters while maintaining good readability for secondary characters; and (3) standardMode globally optimizes classical aesthetic metrics. Our approach is based on Mixed Integer Linear Programming (MILP), offering optimality guarantees, precise balancing of competing metrics through weighted objectives, and the flexibility to incorporate complex fairness concepts as additional constraints without the need to redesign the entire algorithm. We conducted an extensive experimental analysis to demonstrate how F2Stories enables more fair or focus group-prioritized storyline visualizations while maintaining adherence to established layout constraints. Our evaluation includes comprehensive results from a detailed case study that shows the effectiveness of our approach in real-world narrative contexts. An open access copy of this paper and all supplemental materials are available at osf.io/e2qvy.

Multimodal vision-language models (VLMs) continue to achieve ever-improving scores on chart understanding benchmarks. Yet, we find that this progress does not fully capture the breadth of visual reasoning capabilities essential for interpreting charts. We introduce EncQA, a novel benchmark informed by the visualization literature, designed to provide systematic coverage of visual encodings and analytic tasks that are crucial for chart understanding. EncQA provides 2,076 synthetic question-answer pairs, enabling balanced coverage of six visual encoding channels (position, length, area, color quantitative, color nominal, and shape) and eight tasks (find extrema, retrieve value, find anomaly, filter values, compute derived value exact, compute derived value relative, correlate values, and correlate values relative). Our evaluation of 9 state-of-the-art VLMs reveals that performance varies significantly across encodings within the same task, as well as across tasks. Contrary to expectations, we observe that performance does not improve with model size for many task-encoding pairs. Our results suggest that advancing chart understanding requires targeted strategies addressing specific visual reasoning gaps, rather than solely scaling up model or dataset size.

Creating aesthetically pleasing data visualizations remains challenging for users without design expertise or familiarity with visualization tools. To address this gap, we present DataWink, a system that enables users to create custom visualizations by adapting high-quality examples. Our approach combines large multimodal models (LMMs) to extract data encoding from existing SVG-based visualization examples, featuring an intermediate representation of visualizations that bridges primitive SVG and visualization programs. Users may express adaptation goals to a conversational agent and control the visual appearance through widgets generated on demand. With an interactive interface, users can modify both data mappings and visual design elements while maintaining the original visualization's aesthetic quality. To evaluate DataWink, we conduct a user study (N=12) with replication and free-form exploration tasks. As a result, DataWink is recognized for its learnability and effectiveness in personalized authoring tasks. Our results demonstrate the potential of example-driven approaches for democratizing visualization creation.

Tactile graphics are widely used to present maps and statistical diagrams to blind and low vision (BLV) people, with accessibility guidelines recommending their use for graphics where spatial relationships are important. Their use is expected to grow with the advent of commodity refreshable tactile displays. However, in stark contrast to visual information graphics, we lack a clear understanding of the benefts that well-designed tactile information graphics offer over text descriptions for BLV people. To address this gap, we introduce a framework considering the three components of encoding, perception and cognition to examine the known benefts for visual information graphics and explore their applicability to tactile information graphics. This work establishes a preliminary theoretical foundation for the tactile-frst design of information graphics and identifes future research avenues.

We present VizGenie, a self-improving, agentic framework that advances scientific visualization through large language model (LLM) by orchestrating of a collection of domain-specific and dynamically generated modules. Users initially access core functionalities-such as threshold-based filtering, slice extraction, and statistical analysis-through pre-existing tools. For tasks beyond this baseline, VizGenie autonomously employs LLMs to generate new visualization scripts (e.g., VTK Python code), expanding its capabilities on-demand. Each generated script undergoes automated backend validation and is seamlessly integrated upon successful testing, continuously enhancing the system's adaptability and robustness. A distinctive feature of VizGenie is its intuitive natural language interface, allowing users to issue high-level feature-based queries (e.g., "visualize the skull" or "highlight tissue boundaries"). The system leverages image-based analysis and visual question answering (VQA) via fine-tuned vision models to interpret these queries precisely, bridging domain expertise and technical implementation. Additionally, users can interactively query generated visualizations through VQA, facilitating deeper exploration. Reliability and reproducibility are further strengthened by Retrieval-Augmented Generation (RAG), providing context-driven responses while maintaining comprehensive provenance records. Evaluations on complex volumetric datasets demonstrate significant reductions in cognitive overhead for iterative visualization tasks. By integrating curated domain-specific tools with LLM-driven flexibility, VizGenie not only accelerates insight generation but also establishes a sustainable, continuously evolving visualization practice. The resulting platform dynamically learns from user interactions, consistently enhancing support for feature-centric exploration and reproducible research in scientific visualization.

Traditional volume visualization (VolVis) methods, like direct volume rendering, suffer from rigid transfer function designs and high computational costs. Although novel view synthesis approaches enhance rendering efficiency, they require additional learning effort for non-experts and lack support for semantic-level interaction. To bridge this gap, we propose NLI4VolVis, an interactive system that enables users to explore, query, and edit volumetric scenes using natural language. NLI4VolVis integrates multi-view semantic segmentation and vision-language models to extract and understand semantic components in a scene. We introduce a multi-agent large language model architecture equipped with extensive function-calling tools to interpret user intents and execute visualization tasks. The agents leverage external tools and declarative VolVis commands to interact with the VolVis engine powered by 3D editable Gaussians, enabling open-vocabulary object querying, real-time scene editing, best-view selection, and 2D stylization. We validate our system through case studies and a user study, highlighting its improved accessibility and usability in volumetric data exploration. We strongly recommend readers check out our case studies, demo video, and source code at https://nli4volvis.github.io/.