GamED.AI : A Hierarchical Multi-Agent Framework for Automated Educational Game Generation

Shiven Agarwal, Yash Shah, Ashish Raj Shekhar, Priyanuj Bordoloi, Vivek Gupta, Demo, Video, Code · 2026

Read GamED.AI : A Hierarchical Multi-Agent Framework for Automated Educational Game Generation by Shiven Agarwal, Yash Shah, Ashish Raj Shekhar, Priyanuj Bordoloi, Vivek Gupta, Demo, Video, Code (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-23947.txt.

Description

Agarwal and four colleagues present GamED.AI, a hierarchical multi-agent framework that turns teacher questions into playable, pedagogically grounded games checked by contracts. Built on LangGraph sub-graphs, quality gates, and Pydantic schemas, it covers two template families with fifteen interaction mechanics spanning procedural execution, higher-order Bloom's taxonomy objectives, and spatial reasoning. On 200 questions across five subject domains it reaches a 90 percent validation pass rate and 98.3 percent schema compliance, with 73 percent token reduction over ReAct agents, about 73,500 down to about 19,900 tokens per game, at 46 cents per game.

Catalog record

CREATORShiven Agarwal, Yash Shah, Ashish Raj Shekhar, Priyanuj Bordoloi, Vivek Gupta, Demo, Video, Code
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE28121
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.23947

Browse the full library · About the Tower of Babel