Anna Maria Gianni, Nikolaos Nikolakis, Nikolaos Antoniadis
{"title":"基于LLM的游戏化XR自适应反馈机制学习框架","authors":"Anna Maria Gianni, Nikolaos Nikolakis, Nikolaos Antoniadis","doi":"10.1016/j.cexr.2025.100116","DOIUrl":null,"url":null,"abstract":"<div><div>Rapid technological advancements present challenges in computer science education, as traditional instructional approaches often fail to maintain learner engagement or adapt effectively to diverse learning needs. To address these limitations, this study proposes an innovative adaptive learning framework integrating real-time feedback from large language models (LLMs), personalized learning via model-agnostic meta-learning (MAML), and game-theoretic incentives in an immersive XR environment. Learners are modeled as strategic agents whose individual and collaborative behaviors dynamically align with course objectives. Preliminary evaluation in a real-world computer science course demonstrated a 22 % increase in student-reported motivation and over 40 % fewer task retries compared to a traditional digital baseline. These early findings highlight the framework's practical potential to significantly enhance engagement, personalization, and effectiveness in technical education.</div></div>","PeriodicalId":100320,"journal":{"name":"Computers & Education: X Reality","volume":"7 ","pages":"Article 100116"},"PeriodicalIF":0.0000,"publicationDate":"2025-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"An LLM based learning framework for adaptive feedback mechanisms in gamified XR\",\"authors\":\"Anna Maria Gianni, Nikolaos Nikolakis, Nikolaos Antoniadis\",\"doi\":\"10.1016/j.cexr.2025.100116\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><div>Rapid technological advancements present challenges in computer science education, as traditional instructional approaches often fail to maintain learner engagement or adapt effectively to diverse learning needs. To address these limitations, this study proposes an innovative adaptive learning framework integrating real-time feedback from large language models (LLMs), personalized learning via model-agnostic meta-learning (MAML), and game-theoretic incentives in an immersive XR environment. Learners are modeled as strategic agents whose individual and collaborative behaviors dynamically align with course objectives. Preliminary evaluation in a real-world computer science course demonstrated a 22 % increase in student-reported motivation and over 40 % fewer task retries compared to a traditional digital baseline. These early findings highlight the framework's practical potential to significantly enhance engagement, personalization, and effectiveness in technical education.</div></div>\",\"PeriodicalId\":100320,\"journal\":{\"name\":\"Computers & Education: X Reality\",\"volume\":\"7 \",\"pages\":\"Article 100116\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2025-09-15\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Computers & Education: X Reality\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S2949678025000248\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Computers & Education: X Reality","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2949678025000248","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
An LLM based learning framework for adaptive feedback mechanisms in gamified XR
Rapid technological advancements present challenges in computer science education, as traditional instructional approaches often fail to maintain learner engagement or adapt effectively to diverse learning needs. To address these limitations, this study proposes an innovative adaptive learning framework integrating real-time feedback from large language models (LLMs), personalized learning via model-agnostic meta-learning (MAML), and game-theoretic incentives in an immersive XR environment. Learners are modeled as strategic agents whose individual and collaborative behaviors dynamically align with course objectives. Preliminary evaluation in a real-world computer science course demonstrated a 22 % increase in student-reported motivation and over 40 % fewer task retries compared to a traditional digital baseline. These early findings highlight the framework's practical potential to significantly enhance engagement, personalization, and effectiveness in technical education.