arXiv subjects

Shihui Feng

Publications and source records attributed to Shihui Feng.

10 recordsLinked to original sources

arXiv2026

Human-AI Collaboration Reconfigures Group Regulation from Socially Shared to Hybrid Co-Regulation

Generative AI (GenAI) is increasingly used in collaborative learning, yet its effects on how groups regulate collaboration remain unclear. Effective collaboration depends not only on what groups discuss, but on how they jointly manage goals, participation, strategy use, monitoring, and repair through co-regulation and socially shared regulation. We compared collaborative regulation between Human-AI and Human-Human groups in a parallel-group randomised experiment with 71 university students completing the same collaborative tasks with GenAI either available or unavailable. Focusing on human discourse, we used statistical analyses to examine differences in the distribution of collaborative regulation across regulatory modes, regulatory processes, and participatory focuses. Results showed that GenAI availability shifted regulation away from predominantly socially shared forms towards more hybrid co-regulatory forms, with selective increases in directive, obstacle-oriented, and affective regulatory processes. Participatory-focus distributions, however, were broadly similar across conditions. These findings suggest that GenAI reshapes the distribution of regulatory responsibility in collaboration and offer implications for the human-centred design of AI-supported collaborative learning.

cs.AI↗

arXiv2026

Single-agent vs. Multi-agents for Automated Video Analysis of On-Screen Collaborative Learning Behaviors

On-screen learning behavior provides valuable insights into how students seek, use, and create information during learning. Analyzing on-screen behavioral engagement is essential for capturing students' cognitive and collaborative processes. The recent development of Vision Language Models (VLMs) offers new opportunities to automate the labor-intensive manual coding often required for multimodal video data analysis. In this study, we compared the performance of both leading closed-source VLMs (Claude-3.7-Sonnet, GPT-4.1) and open-source VLM (Qwen2.5-VL-72B) in single- and multi-agent settings for automated coding of screen recordings in collaborative learning contexts based on the ICAP framework. In particular, we proposed and compared two multi-agent frameworks: 1) a three-agent workflow multi-agent system (MAS) that segments screen videos by scene and detects on-screen behaviors using cursor-informed VLM prompting with evidence-based verification; 2) an autonomous-decision MAS inspired by ReAct that iteratively interleaves reasoning, tool-like operations (segmentation/ classification/ validation), and observation-driven self-correction to produce interpretable on-screen behavior labels. Experimental results demonstrated that the two proposed MAS frameworks achieved viable performance, outperforming the single VLMs in scene and action detection tasks. It is worth noting that the workflow-based agent achieved best on scene detection, and the autonomous-decision MAS achieved best on action detection. This study demonstrates the effectiveness of VLM-based Multi-agent System for video analysis and contributes a scalable framework for multimodal data analytics.

cs.AI↗

arXiv2026

Comparing the Impact of Pedagogy-Informed Custom and General-Purpose GAI Chatbots on Students' Science Problem-Solving Processes and Performance Using Heterogeneous Interaction Network Analysis

Problem solving plays an essential role in science education, and generative AI (GAI) chatbots have emerged as a promising tool for supporting students' science problem solving. However, general-purpose chatbots (e.g., ChatGPT), which often provide direct, ready-made answers, may lead to students' cognitive offloading. Prior research has rarely focused on custom chatbots for facilitating students' science problem solving, nor has it examined how they differently influence problem-solving processes and performance compared to general-purpose chatbots. To address this gap, we developed a pedagogy-informed custom GAI chatbot grounded in the Socratic questioning method, which supports students by prompting them with guiding questions. This study employed a within-subjects counterbalanced design in which 48 secondary school students used both custom and general-purpose chatbot to complete two science problem-solving tasks. 3297 student-chatbot dialogues were collected and analyzed using Heterogeneous Interaction Network Analysis (HINA). The results showed that: (1) students demonstrated significantly higher interaction intensity and cognitive interaction diversity when using custom chatbot than using general-purpose chatbot; (2) students were more likely to follow custom chatbot's guidance to think and reflect, whereas they tended to request general-purpose chatbot to execute specific commands; and (3) no statistically significant difference was observed in students' problem-solving performance evaluated by solution quality between two chatbot conditions. This study provides novel theoretical insights and empirical evidence that custom chatbots are less likely to induce cognitive offloading and instead foster greater cognitive engagement compared to general-purpose chatbots. This study also offers insights into the design and integration of GAI chatbots in science education.

cs.SI↗

arXiv2026

Heterogeneous Interaction Network Analysis (HINA): A New Learning Analytics Approach for Modelling, Analyzing, and Visualizing Complex Interactions in Learning Processes

Existing learning analytics approaches, which often model learning processes as sequences of learner actions or homogeneous relationships, are limited in capturing the distributed, multi-faceted nature of interactions in contemporary learning environments. To address this, we propose Heterogeneous Interaction Network Analysis (HINA), a novel multi-level learning analytics framework for modeling complex learning processes across diverse entities (e.g., learners, behaviours, AI agents, and task designs). HINA integrates a set of original methods, including summative measures and a new non-parametric clustering technique, with established practices for statistical testing and interactive visualization to provide a flexible and powerful analytical toolkit. In this paper, we first detail the theoretical and mathematical foundations of HINA for individual, dyadic, and meso-level analysis. We then demonstrate HINA's utility through a case study on AI-mediated small-group collaborative learning, revealing students' interaction profiles with peers versus AI; distinct engagement patterns that emerge from these interactions; and specific types of learning behaviors (e.g., asking questions, planning) directed to AI versus peers. By transforming process data into Heterogeneous Interaction Networks (HINs), HINA introduces a new paradigm for modeling learning processes and provides the dedicated, multi-level analytical methods required to extract meaning from them. It thereby moves beyond a single process data type to quantify and visualize how different elements in a learning environment interact and co-influence each other, opening new avenues for understanding complex educational dynamics.

cs.SI↗

arXiv2025

Where is AIED Headed? Key Topics and Emerging Frontiers (2020-2024)

In this study, we analyze 2,398 research articles published between 2020 and 2024 across eight core venues related to the field of Artificial Intelligence in Education (AIED). Using a three-step knowledge co-occurrence network analysis, we analyze the knowledge structure of the field, the evolving knowledge clusters, and the emerging frontiers. Our findings reveal that AIED research remains strongly technically focused, with sustained themes such as intelligent tutoring systems, learning analytics, and natural language processing, alongside rising interest in large language models (LLMs) and generative artificial intelligence (GenAI). By tracking the bridging keywords over the past five years, we identify four emerging frontiers in AIED--LLMs, GenAI, multimodal learning analytics, and human-AI collaboration. The current research interests in GenAI are centered around GAI-driven personalization, self-regulated learning, feedback, assessment, motivation, and ethics.The key research interests and emerging frontiers in AIED reflect a growing emphasis on co-adaptive, human-centered AI for education. This study provides the first large-scale field-level mapping of AIED's transformation in the GenAI era and sheds light on the future research development and educational practices.

cs.SI↗

arXiv2025

LLM-Assisted Automated Deductive Coding of Dialogue Data: Leveraging Dialogue-Specific Characteristics to Enhance Contextual Understanding

Dialogue data has been a key source for understanding learning processes, offering critical insights into how students engage in collaborative discussions and how these interactions shape their knowledge construction. The advent of Large Language Models (LLMs) has introduced promising opportunities for advancing qualitative research, particularly in the automated coding of dialogue data. However, the inherent contextual complexity of dialogue presents unique challenges for these models, especially in understanding and interpreting complex contextual information. This study addresses these challenges by developing a novel LLM-assisted automated coding approach for dialogue data. The novelty of our proposed framework is threefold: 1) We predict the code for an utterance based on dialogue-specific characteristics -- communicative acts and communicative events -- using separate prompts following the role prompts and chain-of-thoughts methods; 2) We engaged multiple LLMs including GPT-4-turbo, GPT-4o, DeepSeek in collaborative code prediction; 3) We leveraged the interrelation between events and acts to implement consistency checking using GPT-4o. In particular, our contextual consistency checking provided a substantial accuracy improvement. We also found the accuracy of act predictions was consistently higher than that of event predictions. This study contributes a new methodological framework for enhancing the precision of automated coding of dialogue data as well as offers a scalable solution for addressing the contextual challenges inherent in dialogue analysis.

cs.CL↗

arXiv2025

Analyzing Students' Emerging Roles Based on Quantity and Heterogeneity of Individual Contributions in Small Group Online Collaborative Learning Using Bipartite Network Analysis

Understanding students' emerging roles in computer-supported collaborative learning (CSCL) is critical for promoting regulated learning processes and supporting learning at both individual and group levels. However, it has been challenging to disentangle individual performance from group-based deliverables. This study introduces new learning analytic methods based on student -- subtask bipartite networks to gauge two conceptual dimensions -- quantity and heterogeneity of individual contribution to subtasks -- for understanding students' emerging roles in online collaborative learning in small groups. We analyzed these two dimensions and explored the changes of individual emerging roles within seven groups of high school students ($N = 21$) in two consecutive collaborative learning projects. We found a significant association in the changes between assigned leadership roles and changes in the identified emerging roles between the two projects, echoing the importance of externally facilitated regulation scaffolding in CSCL. We also collected qualitative data through a semi-structured interview to further validate the quantitative analysis results, which revealed that student perceptions of their emerging roles were consistent with the quantitative analysis results. This study contributes new learning analytic methods for collaboration analytics as well as a two-dimensional theoretical framework for understanding students' emerging roles in small group CSCL.

cs.SI↗

arXiv2023

Heterogenous Network Analytics of Small Group Teamwork: Using Multimodal Data to Uncover Individual Behavioral Engagement Strategies

Individual behavioral engagement is an important indicator of active learning in collaborative settings, encompassing multidimensional behaviors mediated through various interaction modes. Little existing work has explored the use of multimodal process data to understand individual behavioral engagement in face-to-face collaborative learning settings. In this study we bridge this gap, for the first time, introducing a heterogeneous tripartite network approach to analyze the interconnections among multimodal process data in collaborative learning. Students' behavioral engagement strategies are analyzed based on their interaction patterns with various spatial locations and verbal communication types using a heterogeneous tripartite network. The multimodal collaborative learning process data were collected from 15 teams of four students. We conducted stochastic blockmodeling on a projection of the heterogeneous tripartite network to cluster students into groups that shared similar spatial and oral engagement patterns. We found two distinct clusters of students, whose characteristic behavioural engagement strategies were identified by extracting interaction patterns that were statistically significant relative to a multinomial null model. The two identified clusters also exhibited a statistically significant difference regarding students' perceived collaboration satisfaction and teacher-assessed team performance level. This study advances collaboration analytics methodology and provides new insights into personalized support in collaborative learning.

cs.SI↗

arXiv2020

Online geolocalized emotion across US cities during the COVID crisis: Universality, policy response, and connection with local mobility

As the COVID-19 pandemic began to sweep across the US it elicited a wide spectrum of responses, both online and offline, across the population. To aid the development of effective spatially targeted interventions in the midst of this turmoil, it is important to understand the geolocalization of these online emotional responses, as well as their association with offline behavioral responses. Here, we analyze around 13 million geotagged tweets in 49 cities across the US from the first few months of the pandemic to assess regional dependence in online sentiments with respect to a few major topics, and how these sentiments correlate with policy development and human mobility. Surprisingly, we observe universal trends in overall and topic-based sentiments across cities over the time period studied, with variability primarily seen only in the immediate impact of federal guidelines and local lockdown policies. We also find that these local sentiments are highly correlated with and predictive of city-level mobility, while the correlations between sentiments and local cases and deaths are relatively weak. Our findings point to widespread commonalities in the online public emotional responses to COVID across the US, both temporally and relative to offline indicators, in contrast with the high variability seen in early local containment policies. This study also provides new insights into the use of social media data in crisis management by integrating offline data to gain an in-depth understanding of public emotional responses, policy development, and local mobility.

physics.soc-ph↗

arXiv2020

Mixing Patterns in Interdisciplinary Collaboration Networks: Assessing Interdisciplinarity Through Multiple Lenses

There are inherent challenges to interdisciplinary research collaboration, such as bridging cognitive gaps and balancing transaction costs with collaborative benefits. This raises the question: Does interdisciplinary research necessarily result in interdisciplinary collaborations? This study aims to explore this question and assess collaboration preferences in interdisciplinary research at the individual, dyadic, and team level by examining mixing patterns in a collaboration network. Using a network of over 2,000 researchers from the field of artificial intelligence in education, we find that "interdisciplinarity" is demonstrated by diverse research experiences of individual researchers rather than diversity among researchers within collaborations. We also examine intergroup mixing by applying a novel approach to classify the active and non-active researchers in the collaboration network based on participation in multiple teams. We find a significant difference in indicators of academic performance and experience between the clusters of active and non-active researchers, suggesting intergroup mixing as a key factor in academic success. Our results shed light on the nature of team formation in interdisciplinary research, as well as highlight the importance of interdisciplinary programs.

cs.SI↗