ProPACT: A Proactive AI-Driven Adaptive Collaborative Tutor for Pair Programming

2026-05-04

AI Summary

This study addresses the limitations of existing adaptive learning systems, which predominantly focus on individual learners and react passively, thereby struggling to effectively regulate the dynamic collaboration inherent in pair programming. To overcome this, the work proposes the first proactive, collaboration-centered adaptive intervention mechanism. It introduces a multimodal dyadic learning model that integrates shared visual attention with both shared and individual cognitive load metrics. Leveraging XGBoost-based time-series prediction, the system anticipates suboptimal collaboration states in advance and delivers minimally intrusive scaffolding through a hierarchical adaptive strategy. Evaluated in a controlled experiment involving 26 programming pairs, the approach significantly improves debugging success rates, task efficiency, feedback adoption, and post-intervention levels of shared attention and cognitive synergy.

Abstract

Effective pair programming depends on coordination of attention, cognitive effort, and joint regulation over time, yet most adaptive learning systems remain individual-centric and reactive. This paper introduces ProPACT, a proactive AI-driven adaptive collaborative tutor that treats collaboration itself as the object of instruction. ProPACT constructs a multimodal dyadic learner model based on Joint Visual Attention (JVA), Joint Mental Effort (JME), and individual mental effort, and employs an XGBoost-based forecasting model to predict emerging suboptimal collaboration states up to 30 seconds in advance. These predictions drive a hierarchical adaptive policy that delivers minimally intrusive scaffolds while fading support during productive collaboration. A within-subject study with 26 pair-programming dyads shows that proactive feedback significantly improves debugging success, task efficiency, feedback uptake, and post-intervention gains in JVA and JME, demonstrating the potential of forecast-driven dyadic adaptivity for real-time collaborative learning regulation.

Problem

• pair programming
• collaborative learning
• adaptive tutoring
• joint regulation
• proactive AI

Innovation

• proactive adaptation
• dyadic learner modeling
• joint visual attention
• collaborative regulation
• forecast-driven scaffolding