Date
AI Overview
Why it matters
Traditional personalized learning produces static paths, while large language models struggle to capture long-term progress or align with evolving student goals and motivations.
Study at a glance
- Question: Can reinforcement learning optimize artificial intelligence models to plan personalized, goal-driven learning paths for students?
- Setting and data: 300 online learning sessions with 14,584 interaction turns; 22 undergraduate students in a controlled trial.
- Method: Machine learning model development and evaluation, alongside a controlled user study.
- AI tools: Pxplore, GPT-4o, GPT-4o-mini, Qwen3-8B, Llama-3.1-8B-Instruct, Claude-3.7-Sonnet
Key findings
- Undergraduate students using the Pxplore system increased their test scores by 28.28 percentage points, outperforming the control group that used standard content-matching.
- The reinforcement learning-trained model achieved a 65.47% pedagogical alignment rate, outperforming both supervised fine-tuning and proprietary models like GPT-4o.
- Students using Pxplore reported higher satisfaction, motivation, perceived personalization, and relevance compared to students using a standard retrieval baseline.
Summary drafted with AI from the full paper and may contain errors. See the full paper for details.
Personalized Learning Path Planning (PLPP) aims to design adaptive learning paths that align with individual goals. While large language models (LLMs) show potential in personalizing learning experiences, existing approaches often lack mechanisms for goal-aligned planning. We introduce Pxplore, a novel framework for PLPP that integrates a reinforcement-based training paradigm and an LLM-driven educational architecture. We design a structured learner state model and an automated reward function that transforms abstract objectives into computable signals. We train the policy combining supervised fine-tuning (SFT) and Group Relative Policy Optimization (GRPO), and deploy it within a real-world learning platform. Extensive experiments validate Pxplore's effectiveness in producing coherent, personalized, and goal-driven learning paths. We release our code and dataset at https://github.com/Pxplore/pxplore-algo.
Research Categories
What is the application?
Who is the user?
Who age?
Why use AI?
Publication Source
Publisher
arXiv
