Hey, this is Shuai. I am a PhD candidate in HCI Lab at Hong Kong University of Science and Technology (HKUST). I'm fortunate to be advised by Prof. Xiaojuan Ma.
My research goal is to design human-centered interactive systems that users can easily understand and use, appropriately rely on, and effectively collaborate with.
Shuai Ma, Qiaoyi Chen, Xinru Wang, Chengbo Zheng, Zhenhui Peng, Ming Yin, Xiaojuan Ma. (Working Paper)
[PDF]
In AI-assisted decision-making, humans often accept or reject AI suggestions passively, facing challenges in analytical engagement and communicating disagreements effectively. To tackle this, we propose Human-AI Deliberation, fostering reflective discussion on conflicting opinions. This framework involves dimension-level opinion elicitation, deliberative discussion, and updates, guided by human deliberation theories. Deliberative AI, utilizing large language models, enhances interactions between humans and domain-specific models, improving decision-making such as in graduate admissions. Evaluation shows Deliberative AI outperforms traditional explainable AI assistants, enhancing reliance and task performance, informing future AI-assisted decision tool designs.
Shuai Ma, Chenyi Zhang, Xinru Wang, Xiaojuan Ma, Ming Yin. (Working Paper)
[PDF]
AI is often used as a Recommender in decision-making, but this can reduce human analytical thinking and cause over-reliance on AI. Unlike AI, human advisors play diverse roles in decisions. This paper examines three AI roles—Recommender, Analyzer, and Devil’s Advocate—at two performance levels. Results show each role's unique strengths and weaknesses in task performance, reliance, and user experience. Notably, the Analyzer role can be more effective than the Recommender when AI performance is low. These findings inform the design of adaptive AI assistants.
Shuai Ma, Xinru Wang, Ying Lei, Chuhan Shi, Ming Yin, Xiaojuan Ma. (CHI 2024)
[PDF]
This paper investigates human self-confidence calibration in AI-assisted decision-making, conducting three user studies to analyze its effect on human-AI collaboration. The research examines the impact of self-confidence on AI reliance and tests three calibration mechanisms. Findings indicate that proper self-confidence calibration improves rational behavior and appropriate reliance in human-AI teams, offering insights into enhancing collaboration efficiency.
Chengbo Zheng, Kangyu Yuan, Bingcan Guo, Reza Hadi Mogavi, Zhenhui Peng, Shuai Ma, Xiaojuan Ma. (CHI 2024)
[PDF]
This study investigates using AI usage data for learning assessment in project-based education. Through workshops with college students, we explored how AI data can reflect students' skills and contributions. The findings highlight the potential and challenges of integrating AI data in educational assessments, informing the development of future educational tools for data analysis and presentation.
Ying Lei, Shuai Ma, Yuling Sun. (CHI 2024)
[PDF]
This paper explores the ICT-mediated social connections of late-life migrants, focusing on the dynamic changes in their social networks and the roles of ICT in these shifts. Utilizing the social convoy model, we examine the evolving support roles within migrants' networks, alongside the challenges and expectations related to ICT-supported social connections. Our findings offer in-depth insights into the social connections and support systems of late-life migrants, culminating in design implications for future ICT-based support systems for this demographic.
Qian Zhu, Dakuo Wang, Shuai Ma, April Wang, Zixin Chen, Udayan Khurana, Xiaojuan Ma. (DIS 2024)
[PDF]
How do data scientists view suggestions from other peers and AI when performing feature engineering? When the two are presented to data scientists at the same time, how will they choose? What are the key deficiencies in today’s AI when recommending features? In this article, we designed a human-AI collaborative feature engineering framework and invited 14 data scientists to conduct experiments.
Shuai Ma, Ying Lei, Xinru Wang, Chengbo Zheng, Chuhan Shi, Ming Yin, Xiaojuan Ma. (CHI 2023)
[PDF]
[Code]
[Live Demo]
[Video]
We proposed to promote humans' appropriate trust based on the correctness likelihood of both sides at a task-instance level. Results from a between-subjects experiment (N=293) showed that our CL exploitation strategies promoted more appropriate human trust in AI, compared with only using AI confidence.
Chuhan Shi, Yicheng Hu, Shenan Wang, Shuai Ma, Chengbo Zheng, Xiaojuan Ma, Qiong Luo. (CHI 2023)
[PDF]
Targeting Multi-step Human-AI Collaboration task for chemists, we proposed a human-AI collaborative system, RetroLens, through a participatory design process. AI can contribute by two approaches: joint action and algorithm-inthe-loop.
Chengbo Zheng, Yuheng Wu, Chuhan Shi, Shuai Ma, Jiehui Luo, Xiaojuan Ma. (CHI 2023)
[PDF]
What will happen if AI participate equally in human group decision-making? We studied this problem in teacher-AI collaborative decision-making. We find that although the voice of AI is considered valuable, AI still plays a secondary role in the group because it cannot fully follow the dynamics of the discussion and make progressive contributions. Moreover, the divergent opinions of our participants regarding an "equal AI" shed light on the possible future of human-AI relations.
Shuai Ma, Mingfei Sun, Xiaojuan Ma. (TOCHI 2022)
[PDF]
[Code]
[Live Demo]
[Video]
For human-robot teaching scenario, we propose an adaptive modeling and expression method to facilitate the transparent communication of robots' learning statuses during human demonstration.
Shuai Ma, Taichang Zhou, Fei Nie, Xiaojuan Ma. (CHI 2022)
[PDF]
[Code]
[Live Demo]
[Video]
Focusing on synchronous online classes (e.g., real-time Zoom-based classes), Glancee address instructors’ difficulty to observe students’ learning status due to students’ unwillingness to show their videos. Specifically, we mitigate the gap that lack of empirical investigation on instructors’ preferences and lack of exploration of designing adaptable systems to meet the needs of individual instructors.
Liuping Wang, Dakuo Wang, Feng Tian, Zhenhui Peng, Xiangmin Fan, Zhan Zhang, Shuai Ma, Mo Yu, Xiaojuan Ma, Hongan Wang. (CSCW 2021)
[PDF]
We investigated how chatbots can be designed to provide information and emotional support for pregnant women in an online health community.
Shuai Ma, Zijun Wei, Feng Tian, Xiangmin Fan, Jianming Zhang, Xiaohui Shen, Zhe Lin, Jin Huang, Radomir Mech, Dimitris Samaras, Hongan Wang. (CHI 2019)
[PDF]
How to effectively personalize a general model? We proposed a user preference modeling method based on interactive machine learning and designed a confidence-based integration framework to personalize a deep neural network to cater to users' individual preferences in photo composition. Based on the proposed algorithm, we designed SmartEye, which can gradually learn users' preferences as the interaction goes on.
Shuai Ma*, Qian Zhu*, Cuixia Ma (UIST 2019 Adjunct)
[PDF]
Video editing is challenging for both professionals and amateurs due to the time-consuming task of screening useful clips from raw footage. To address these difficulties, we conducted a pilot study involving surveys and interviews with 20 participants. Based on our findings, we developed Pre-screen, an innovative tool that offers global and detailed video analysis, as well as intelligent material screening features using advanced video processing and visualization methods.
Shuai Ma*, Qian Zhu* (UIST 2019 Adjunct)
[PDF]
In Massive Open Online Courses (MOOCs), learners often face distractions leading to divided attention (DA). This paper introduces Reminder, a system that detects DA using cameras on PC and mobile devices. It predicts attention scores with a regression model and adapts to individual users. Reminder also offers visualizations to help learners easily review missed content. User studies demonstrate its effectiveness in detecting and assisting learners with missed course content.
Yunzhi Li, Liuping Wang, Shuai Ma, Xiangmin Fan, Zijun Wang, Junfeng Jiao, Dakuo Wang (CHI 2019 Workshop Paper)
[PDF]
This position paper introduces three ongoing research projects focused on designing, developing, and evaluating systems for human-AI interaction in healthcare. Collaborating with local government administrators, hospitals, clinics, and doctors in a Beijing suburb, we study how AI technologies are transforming healthcare delivery and reception. We aim to foster discussion at the workshop and establish collaborations within the health informatics community.
Jing Gao, Feng Tian, Junjun Fan, Dakuo Wang, Xiangmin Fan, Yicheng Zhu, Shuai Ma, Jin Huang, Hongan Wang. (CHI 2018 LBW)
[PDF]
In this work, we explored the feasibility and accuracy of detecting motor impairment in Parkinson’s disease (PD) via implicitly sensing and analyzing users’ everyday interactions with their smartphones. Through a 42 subjects study, our approach achieved an overall accuracy of 88.1% (90.0%/86.4% sensitivity/specificity) in discriminating PD subjects from age-matched healthy controls.
Liuping Wang, Xiangmin Fan, Feng Tian, Lingjia Deng, Shuai Ma, Jin Huang, Hongan Wang (CHI 2018 LBW)
[PDF]
We present mirrorU, a mobile system that supports users to reflect on and write about their daily emotional experience. While prior work has focused primarily on providing memory triggers or affective cues, mirrorU provides in-situ assessment and interactive feedback to scaffold reflective writing. It automatically and continuously monitors the composition process in three dimensions (i.e., level of detail, overall valence, and cognitive engagement) and provides relevant feedback to support reflection.
Program Committee: ACM FAccT '24, ACM CHI '24 LBW, ACM CHI '23 LBW
Conference Review: ACM CHI '24 (1), '23 (2), '22, '20, '19, CSCW '23, UIST '22, CHI EA '23, '22, WWW '21
Journal Review: ACM TOCHI, ACM TiiS, CCF TOPCI
Volunteer: CHI' 23 (3), CHI' 22 (4)
(1) (2) received Special Recognition for Outstanding Reviews for CHI 2023, CHI 2024
(3) (4) received Student Volunteer Award for CHI 2022, CHI 2023