arXiv subjects

Linhao Meng

Publications and source records attributed to Linhao Meng.

4 recordsLinked to original sources

arXiv2026

LangLasso: Interactive Cluster Descriptions through LLM Explanation

Dimensionality reduction is a powerful technique for revealing structure and potential clusters in data. However, as the axes are complex, non-linear combinations of features, they often lack semantic interpretability. Existing visual analytics (VA) methods support cluster interpretation through feature comparison and interactive exploration, but they require technical expertise and intense human effort. We present \textit{LangLasso}, a novel method that complements VA approaches through interactive, natural language descriptions of clusters using large language models (LLMs). It produces human-readable descriptions that make cluster interpretation accessible to non-experts and allow integration of external contextual knowledge beyond the dataset. We systematically evaluate the reliability of these explanations and demonstrate that \langlasso provides an effective first step for engaging broader audiences in cluster interpretation. The tool is available at https://langlasso.vercel.app

cs.HC↗

arXiv2025

Towards Difficulty-Aware Analysis of Deep Neural Networks

Traditional instance-based model analysis focuses mainly on misclassified instances. However, this approach overlooks the varying difficulty associated with different instances. Ideally, a robust model should recognize and reflect the challenges presented by intrinsically difficult instances. It is also valuable to investigate whether the difficulty perceived by the model aligns with that perceived by humans. To address this, we propose incorporating instance difficulty into the deep neural network evaluation process, specifically for supervised classification tasks on image data. Specifically, we consider difficulty measures from three perspectives -- data, model, and human -- to facilitate comprehensive evaluation and comparison. Additionally, we develop an interactive visual tool, DifficultyEyes, to support the identification of instances of interest based on various difficulty patterns and to aid in analyzing potential data or model issues. Case studies demonstrate the effectiveness of our approach.

cs.HC↗

arXiv2023

Class-constrained t-SNE: Combining Data Features and Class Probabilities

Data features and class probabilities are two main perspectives when, e.g., evaluating model results and identifying problematic items. Class probabilities represent the likelihood that each instance belongs to a particular class, which can be produced by probabilistic classifiers or even human labeling with uncertainty. Since both perspectives are multi-dimensional data, dimensionality reduction (DR) techniques are commonly used to extract informative characteristics from them. However, existing methods either focus solely on the data feature perspective or rely on class probability estimates to guide the DR process. In contrast to previous work where separate views are linked to conduct the analysis, we propose a novel approach, class-constrained t-SNE, that combines data features and class probabilities in the same DR result. Specifically, we combine them by balancing two corresponding components in a cost function to optimize the positions of data points and iconic representation of classes -- class landmarks. Furthermore, an interactive user-adjustable parameter balances these two components so that users can focus on the weighted perspectives of interest and also empowers a smooth visual transition between varying perspectives to preserve the mental map. We illustrate its application potential in model evaluation and visual-interactive labeling. A comparative analysis is performed to evaluate the DR results.

cs.HC↗

arXiv2020

Bubble Storytelling with Automated Animation: A Brexit Hashtag Activism Case Study

Hashtag data are common and easy to acquire. Thus, they are widely used in studies and visual data storytelling. For example, a recent story by China Central Television Europe (CCTV Europe) depicts Brexit as a hashtag movement displayed on an animated bubble chart. However, creating such a story is usually laborious and tedious, because narrators have to switch between different tools and discuss with different collaborators. To reduce the burden, we develop a prototype system to help explore the bubbles' movement by automatically inserting animations connected to the storytelling of the video creators and the interaction of viewers to those videos. We demonstrate the usability of our method through both use cases and a semi-structured user study.

cs.CY↗