Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

GuideAI: A Real-time Personalized Learning Solution with Adaptive Interventions

Ananya Shukla, Chaitanya Modi, Satvik Bajpai, Siddharth Siddharth

TL;DR

GuideAI presents a biosensor-augmented, real-time adaptive learning framework that integrates eye tracking, HRV, posture, and note-taking signals to infer six cognitive states and deliver tone- and modality-aware interventions across text, image, audio, and video formats. In a controlled within-subject study (N=25), GuideAI reduced cognitive load and improved problem-solving and recall relative to a non-adaptive baseline, while also enhancing perceived performance and engagement. The work contributes a closed-loop architecture, semantic state abstractions for robust LLM prompting, and modality-specific interventions, demonstrating the feasibility of cognition-aware education at scale. While promising, the authors acknowledge privacy, deployment feasibility, and content reliability challenges, and call for longitudinal, diverse-cohort validations and accessible sensing solutions to broaden adoption.

Abstract

Large Language Models (LLMs) have emerged as powerful learning tools, but they lack awareness of learners' cognitive and physiological states, limiting their adaptability to the user's learning style. Contemporary learning techniques primarily focus on structured learning paths, knowledge tracing, and generic adaptive testing but fail to address real-time learning challenges driven by cognitive load, attention fluctuations, and engagement levels. Building on findings from a formative user study (N=66), we introduce GuideAI, a multi-modal framework that enhances LLM-driven learning by integrating real-time biosensory feedback including eye gaze tracking, heart rate variability, posture detection, and digital note-taking behavior. GuideAI dynamically adapts learning content and pacing through cognitive optimizations (adjusting complexity based on learning progress markers), physiological interventions (breathing guidance and posture correction), and attention-aware strategies (redirecting focus using gaze analysis). Additionally, GuideAI supports diverse learning modalities, including text-based, image-based, audio-based, and video-based instruction, across varied knowledge domains. A preliminary study (N = 25) assessed GuideAI's impact on knowledge retention and cognitive load through standardized assessments. The results show statistically significant improvements in both problem-solving capability and recall-based knowledge assessments. Participants also experienced notable reductions in key NASA-TLX measures including mental demand, frustration levels, and effort, while simultaneously reporting enhanced perceived performance. These findings demonstrate GuideAI's potential to bridge the gap between current LLM-based learning systems and individualized learner needs, paving the way for adaptive, cognition-aware education at scale.

GuideAI: A Real-time Personalized Learning Solution with Adaptive Interventions

TL;DR

GuideAI presents a biosensor-augmented, real-time adaptive learning framework that integrates eye tracking, HRV, posture, and note-taking signals to infer six cognitive states and deliver tone- and modality-aware interventions across text, image, audio, and video formats. In a controlled within-subject study (N=25), GuideAI reduced cognitive load and improved problem-solving and recall relative to a non-adaptive baseline, while also enhancing perceived performance and engagement. The work contributes a closed-loop architecture, semantic state abstractions for robust LLM prompting, and modality-specific interventions, demonstrating the feasibility of cognition-aware education at scale. While promising, the authors acknowledge privacy, deployment feasibility, and content reliability challenges, and call for longitudinal, diverse-cohort validations and accessible sensing solutions to broaden adoption.

Abstract

Large Language Models (LLMs) have emerged as powerful learning tools, but they lack awareness of learners' cognitive and physiological states, limiting their adaptability to the user's learning style. Contemporary learning techniques primarily focus on structured learning paths, knowledge tracing, and generic adaptive testing but fail to address real-time learning challenges driven by cognitive load, attention fluctuations, and engagement levels. Building on findings from a formative user study (N=66), we introduce GuideAI, a multi-modal framework that enhances LLM-driven learning by integrating real-time biosensory feedback including eye gaze tracking, heart rate variability, posture detection, and digital note-taking behavior. GuideAI dynamically adapts learning content and pacing through cognitive optimizations (adjusting complexity based on learning progress markers), physiological interventions (breathing guidance and posture correction), and attention-aware strategies (redirecting focus using gaze analysis). Additionally, GuideAI supports diverse learning modalities, including text-based, image-based, audio-based, and video-based instruction, across varied knowledge domains. A preliminary study (N = 25) assessed GuideAI's impact on knowledge retention and cognitive load through standardized assessments. The results show statistically significant improvements in both problem-solving capability and recall-based knowledge assessments. Participants also experienced notable reductions in key NASA-TLX measures including mental demand, frustration levels, and effort, while simultaneously reporting enhanced perceived performance. These findings demonstrate GuideAI's potential to bridge the gap between current LLM-based learning systems and individualized learner needs, paving the way for adaptive, cognition-aware education at scale.
Paper Structure (57 sections, 2 equations, 10 figures, 2 tables)

This paper contains 57 sections, 2 equations, 10 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Physiological and Cognitive States as Determinants of Learning Efficiency
  4. Biosensors as Biomarkers for Personalized and Adaptive Learning
  5. From Rule-based Systems to LLMs: Toward Multi-modal Personalization in Education
  6. Formative Assessment
  7. Pre-study Survey
  8. AI Usage and Perception
  9. Learning Modality Preferences
  10. AI Learning Tools: Usage Patterns and Limitations
  11. Desired Interventions
  12. Design Implications
  13. GuideAI Solution
  14. Overview
  15. Data Capture
  16. ...and 42 more sections

Figures (10)

  • Figure 1: Participant openness to working with a personalized AI learning assistant on a scale from 1-10. The distribution shows strong receptivity, with 50 out of 66 participants (75.7%) rating their openness at 7 or higher, and a median rating of 8.
  • Figure 2: Desired interventions from personalized AI (N = 66). Top priorities: personalized learning paths (68.2%), tailored practice questions (68.2%), and learning style adaptation (65.2%). Nearly half (47.0%) wanted break reminders and alternative methods. Multiple selections allowed.
  • Figure 3: GuideAI System Architecture with three modules: Sensor Module (left) collects eye tracking, heart activity, posture, and notes data. Processing Module (center) transforms raw signals into physiological and behavioral metrics. Inference Module (right) integrates these to assess six cognitive dimensions (cognitive load, attention, engagement, understanding, stress, fatigue) and deliver personalized interventions across learning modalities (text, image, audio, video).
  • Figure 4: Multi-modal biosensory setup integrating eye tracking, heart rate monitoring, note-taking analysis, and posture detection.
  • Figure 5: GuideAI's biosensing-driven intervention framework linking physiological triggers to adaptive strategies. (a) High cognitive load detected via pupil dilation, elevated heart rate, and fixation duration triggers content pausing and chunking; (b) Understanding difficulties from erroneous notes initiate conceptual correction; (c) High attention states enable complexity increases; (d) Stress detection triggers graduated interventions from simplification to breathing exercises; (e) Low engagement adapts content pacing; (f) Fatigue prompts strategic breaks. This represents a subset of GuideAI's real-time pedagogical adaptations.
  • ...and 5 more figures