Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Ontology-driven Reinforcement Learning for Personalized Student Support

Ryan Hare, Ying Tang

TL;DR

The paper tackles personalized education amid resource constraints by proposing an ontology-driven framework that fuses semantic knowledge with a multi-agent reinforcement learning (RL) brain. It formalizes domain structure as a DAG-based ontology $G=\langle C,E,\rho,D,\alpha\rangle$ and represents student state per topic as vectors $\mathbf{x_{c_i}}$, enabling a multi-agent MDP $\langle S,A,R,p,\gamma\rangle$ for learning. A data-transformation layer maps heterogeneous inputs to a common state and a personalized-assistance generator translates RL outputs into tutoring content, potentially leveraging rule-based systems or large language models. The approach is modular and adaptable to various virtual educational systems, offering scalable AI augmentation of instructors; future work includes empirical validation and deeper RL optimizations.

Abstract

In the search for more effective education, there is a widespread effort to develop better approaches to personalize student education. Unassisted, educators often do not have time or resources to personally support every student in a given classroom. Motivated by this issue, and by recent advancements in artificial intelligence, this paper presents a general-purpose framework for personalized student support, applicable to any virtual educational system such as a serious game or an intelligent tutoring system. To fit any educational situation, we apply ontologies for their semantic organization, combining them with data collection considerations and multi-agent reinforcement learning. The result is a modular system that can be adapted to any virtual educational software to provide useful personalized assistance to students.

Ontology-driven Reinforcement Learning for Personalized Student Support

TL;DR

The paper tackles personalized education amid resource constraints by proposing an ontology-driven framework that fuses semantic knowledge with a multi-agent reinforcement learning (RL) brain. It formalizes domain structure as a DAG-based ontology and represents student state per topic as vectors , enabling a multi-agent MDP for learning. A data-transformation layer maps heterogeneous inputs to a common state and a personalized-assistance generator translates RL outputs into tutoring content, potentially leveraging rule-based systems or large language models. The approach is modular and adaptable to various virtual educational systems, offering scalable AI augmentation of instructors; future work includes empirical validation and deeper RL optimizations.

Abstract

In the search for more effective education, there is a widespread effort to develop better approaches to personalize student education. Unassisted, educators often do not have time or resources to personally support every student in a given classroom. Motivated by this issue, and by recent advancements in artificial intelligence, this paper presents a general-purpose framework for personalized student support, applicable to any virtual educational system such as a serious game or an intelligent tutoring system. To fit any educational situation, we apply ontologies for their semantic organization, combining them with data collection considerations and multi-agent reinforcement learning. The result is a modular system that can be adapted to any virtual educational software to provide useful personalized assistance to students.
Paper Structure (9 sections, 3 equations, 3 figures)

This paper contains 9 sections, 3 equations, 3 figures.

Table of Contents

  1. Introduction
  2. Proposed System Architecture
  3. Ontologies for Education
  4. Data Collection and Storage
  5. Data Transformation
  6. Multi-agent Reinforcement Learning
  7. Generating Personalized Assistance
  8. Example Implementation
  9. Conclusions

Figures (3)

  • Figure 1: Full diagram of the proposed system showing all four components and connections to a virtual educational system.
  • Figure 2: A simplified example ontology that focuses around mathematical functions, showing a few possible sub-concepts and how they're related.
  • Figure 3: An example transformation graph showing how some unknown metrics from $y$ with varied descriptors are mapped to $x$, with appropriate transformation functions assigned along the connecting edges.