DATA COLLECTION PHASE: OPEN

Advancing the Science of
Education with Generative AI.

A research initiative investigating the efficacy of Large Language Models in adaptive learning environments, educational data mining, and learning analytics.

Adaptive Tutor v4.0

Session ID: #8X92-EXP

AI

Research Prompt: Explain the concept of 'Zone of Proximal Development'.

Evaluating scaffolding techniques in real-time.

Core Research Areas

Our multidisciplinary approach combines cognitive science, machine learning, and pedagogy to evaluate AI efficacy.

Collaborative Learning

Developing AI-powered tutors and peers to facilitate authenticate collaborative learning environments, especially for distance education.

Feedback Efficacy

Comparative analysis of diverse forms of learning feedback on student learning behaviors and learning outcomes.

Multimodal Analytics

Aggregating clickstream data, eye-tracking, affection, and natural language inputs to build comprehensive models of student engagement.

analysis_notebook.ipynb
import pandas as pd
import educational_mining_lib as edm
# Load interaction dataset
df = pd.read_csv('student_interactions_v4.csv')
# Calculate retention correlation
correlation = edm.correlate(df['ai_hints'], df['post_test_score'])
print(f"Pearson Coefficient: {correlation:.4f}")
> Pearson Coefficient: 0.8421
> p-value: < 0.001 (Significant)
> Model Convergence: 98.2%
Quantitative Analysis

Quantifiable learning processes, not just the outcomes.

We move beyond simple focus on outcomes and final products. Our vision is to fully unpack learning processes in an AI-mediated learning environment.

  • Log Trace Data Interactive events captured in an online learning platform.
  • Affective Status Real-time affection detection to understand how changes in affective states affect learning.
  • Natural Language Conversation Conversation histories in natural language to make sense of AI's roles in learning.