arXiv subjects

Jiaman He

Publications and source records attributed to Jiaman He.

6 recordsLinked to original sources

arXiv2026

SearchLog: A Web Browser Extension for Capturing Search Logs in Laboratory Studies

Natural search logs are valuable for studying search behavior in information seeking settings. We present SearchLog, an easy-to-install web browser extension for collecting natural search logs during lab-based studies. SearchLog enables participants to search the open web in a browser while recording structured interaction data from the mouse, keyboard, search activity, and browser state modules. The extension captures clicks, scrolling, hovered text, typed words, search queries, result rankings, AI-generated summaries when available, tab activity, and window changes. A local Flask backend stores each session as an ordered JSON event stream, including HTML snapshots and preprocessed search result data for later analysis. These logs can be used to derive measures such as query reformulation, page visits, dwell time, scroll behavior, tab switching, search path complexity, and exposure to AI-generated search content. By supporting natural browser-based search with structured experimental metadata, SearchLog provides a reusable resource to study search behavior across traditional and AI-enhanced search interfaces. SearchLog is available at https://github.com/peanutH/SearchLog-chromium-extension.

cs.IR↗

arXiv2026

EduGage: Methods and Dataset for Sensor-Based Momentary Assessment of Engagement in Self-Guided Video Learning

Engagement, which links to attentional, emotional, and cognitive dimensions, plays an important role in learning. In online and video-based learning environments, learners often need to regulate their own interactions with instructional materials. Measuring and reflecting on engagement can therefore support both learners and adaptive learning systems. In this study, we use wearable and camera-based sensing devices to collect physiological and motion signals, including PPG, ECG, EDA, EEG, IMU, heart rate, temperature, and eye-tracking data, to estimate learner engagement. We conducted a user study with 16 participants in a video-based learning scenario, where participants completed learning tasks and provided repeated in-situ self-reports of engagement through brief probes. We develop and evaluate a system for engagement estimation, compare different sensing modalities, and further analyze the feasibility and effectiveness of multimodal modeling for characterizing learner engagement. Across participant-based cross-validation, our model achieves an MAE of 0.81, 83.75% within-1 accuracy, 73.93% binary accuracy, and 68.45% binary Macro-F1, outperforming sensor-free, statistical, deep temporal, foundation-model, and LLM-based baselines. Our results suggest that fine-grained engagement estimation is feasible but inherently noisy, and that practical systems should prioritize lightweight combinations of behavioral and physiological signals over full multimodal instrumentation. We release the EduGage dataset, including synchronized multimodal sensor signals, probe-aligned momentary engagement labels, video metadata, quizzes, and study materials, to support reproducible research on fine-grained sensor-based engagement modeling in self-guided learning.

cs.HC↗

arXiv2026

Characterizing Personality from Eye-Tracking: The Role of Gaze and Its Absence in Interactive Search Environments

Personality traits influence how individuals engage, behave, and make decisions during the information-seeking process. However, few studies have linked personality to observable search behaviors. This study aims to characterize personality traits through a multimodal time-series model that integrates eye-tracking data and gaze missingness-periods when the user's gaze is not captured. This approach is based on the idea that people often look away when they think, signaling disengagement or reflection. We conducted a user study with 25 participants, who used an interactive application on an iPad, allowing them to engage with digital artifacts from a museum. We rely on raw gaze data from an eye tracker, minimizing preprocessing so that behavioral patterns can be preserved without substantial data cleaning. From this perspective, we trained models to predict personality traits using gaze signals. Our results from a five-fold cross-validation study demonstrate strong predictive performance across all five dimensions: Neuroticism (Macro F1 = 77.69%), Conscientiousness (74.52%), Openness (77.52%), Agreeableness (73.09%), and Extraversion (76.69%). The ablation study examines whether the absence of gaze information affects the model performance, demonstrating that incorporating missingness improves multimodal time-series modeling. The full model, which integrates both time-series signals and missingness information, achieves 10-15% higher accuracy and macro F1 scores across all Big Five traits compared to the model without time-series signals and missingness. These findings provide evidence that personality can be inferred from search-related gaze behavior and demonstrate the value of incorporating missing gaze data into time-series multimodal modeling.

cs.HC↗

arXiv2025

Can We Hide Machines in the Crowd? Quantifying Equivalence in LLM-in-the-loop Annotation Tasks

Many evaluations of large language models (LLMs) in text annotation focus primarily on the correctness of the output, typically comparing model-generated labels to human-annotated ``ground truth'' using standard performance metrics. In contrast, our study moves beyond effectiveness alone. We aim to explore how labeling decisions -- by both humans and LLMs -- can be statistically evaluated across individuals. Rather than treating LLMs purely as annotation systems, we approach LLMs as an alternative annotation mechanism that may be capable of mimicking the subjective judgments made by humans. To assess this, we develop a statistical evaluation method based on Krippendorff's $\alpha$, paired bootstrapping, and the Two One-Sided t-Tests (TOST) equivalence test procedure. This evaluation method tests whether an LLM can blend into a group of human annotators without being distinguishable. We apply this approach to two datasets -- MovieLens 100K and PolitiFact -- and find that the LLM is statistically indistinguishable from a human annotator in the former ($p = 0.004$), but not in the latter ($p = 0.155$), highlighting task-dependent differences. It also enables early evaluation on a small sample of human data to inform whether LLMs are suitable for large-scale annotation in a given application.

cs.IR↗

arXiv2025

AgentSense: Virtual Sensor Data Generation Using LLM Agents in Simulated Home Environments

A major challenge in developing robust and generalizable Human Activity Recognition (HAR) systems for smart homes is the lack of large and diverse labeled datasets. Variations in home layouts, sensor configurations, and individual behaviors further exacerbate this issue. To address this, we leverage the idea of embodied AI agents -- virtual agents that perceive and act within simulated environments guided by internal world models. We introduce AgentSense, a virtual data generation pipeline in which agents live out daily routines in simulated smart homes, with behavior guided by Large Language Models (LLMs). The LLM generates diverse synthetic personas and realistic routines grounded in the environment, which are then decomposed into fine-grained actions. These actions are executed in an extended version of the VirtualHome simulator, which we augment with virtual ambient sensors that record the agents' activities. Our approach produces rich, privacy-preserving sensor data that reflects real-world diversity. We evaluate AgentSense on five real HAR datasets. Models pretrained on the generated data consistently outperform baselines, especially in low-resource settings. Furthermore, combining the generated virtual sensor data with a small amount of real data achieves performance comparable to training on full real-world datasets. These results highlight the potential of using LLM-guided embodied agents for scalable and cost-effective sensor data generation in HAR. Our code is publicly available at https://github.com/ZikangLeng/AgentSense.

cs.CV↗

arXiv2025

Characterising Topic Familiarity and Query Specificity Using Eye-Tracking Data

Eye-tracking data has been shown to correlate with a user's knowledge level and query formulation behaviour. While previous work has focused primarily on eye gaze fixations for attention analysis, often requiring additional contextual information, our study investigates the memory-related cognitive dimension by relying solely on pupil dilation and gaze velocity to infer users' topic familiarity and query specificity without needing any contextual information. Using eye-tracking data collected via a lab user study (N=18), we achieved a Macro F1 score of 71.25% for predicting topic familiarity with a Gradient Boosting classifier, and a Macro F1 score of 60.54% with a k-nearest neighbours (KNN) classifier for query specificity. Furthermore, we developed a novel annotation guideline -- specifically tailored for question answering -- to manually classify queries as Specific or Non-specific. This study demonstrates the feasibility of eye-tracking to better understand topic familiarity and query specificity in search.

cs.IR↗