Publications

2026

Who Benefits from Wikipedia's Growth Mentorship?

Yubo Zhou, Pengyu Chen, Ceren Budak · International AAAI Conference on Web and Social Media (ICWSM), 2026 · Conference Paper

A causal study of 35K+ newcomer conversations on Wikipedia showing that a mentor's reply raises retention (+1.7pp overall, +2.4pp for technical questions), with response speed and topical relevance as the strongest correlates.

Rethinking Message Passing Neural Networks with Diffusion Distance-Guided Stress Majorization

Haoran Zheng, Renchi Yang, Yubo Zhou, Jianliang Xu · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (SIGKDD), 2026 · Conference Paper

DDSM, a graph neural network that reframes message passing as diffusion-distance-guided stress majorization to mitigate over-smoothing, benchmarked against 15 baselines across 11 graphs spanning homophilic and heterophilic structures.

Paper · Preprint · Code

Lookover: Scaffolding Student Participation in Social Annotation

Yubo Zhou, Xinyue Xu, Zesen Zhao, Anhong Guo, Xinyue Chen · ACM Symposium on User Interface Software and Technology (UIST) Adjunct Proceedings, 2026 · Poster

An AI tool that helps students participate more actively in peer discussion on social-annotation platforms while keeping the thinking their own, evaluated in a graduate course.

2024

CogSimulator: A Model for Simulating User Cognition and Behavior with Minimal Data

Weizhen Bian, Yubo Zhou, Yuanhang Luo, Ming Mo, Siyan Liu, Yikai Gong, Renjie Wan, Ziyuan Luo, Aobo Wang · Annual Meeting of the Cognitive Science Society (CogSci), 2024 · Conference Paper

A model that simulates user cognition and behavior from minimal data, aimed at predicting user responses without large-scale data collection.

Paper · Preprint

EmoSpeech: A Corpus of Emotionally Rich and Contextually Detailed Speech Annotations

Weizhen Bian, Yubo Zhou, Kaitai Zhang, Xiaohan Gu · International Symposium on Chinese Spoken Language Processing (ISCSLP), 2024 · Conference Paper

A speech corpus with emotion-rich, natural-language annotations mined from film and TV dialogue, built to support training of emotion-controllable speech systems.

Paper