Transforming Education With Autonomous and Adaptive Agentic AI Systems

The research evaluates how effective agentic AI tutoring systems with their customized and adaptive teaching methods function to enhance student academic performance. The researchers used experimental research methods which included testing Deep Reinforcement Learning, Multi-Agent Systems, Hybrid models, Traditional Intelligent Tutoring Systems, and a Control group. The researchers conducted their analysis by examining data which included pre- and post-test results along with engagement logs and knowledge tracing accuracy data. AI-driven methods lead to better results than traditional methods, which deep reinforcement learning technology produces the most effective learning gains of 54% together with multi-agent systems at 49% and hybrid models at 45% while control group participants showed only 16% advancement. Knowledge tracing models reached 98% predictive accuracy which confirmed that all sessions showed strong performance growth. The researchers discovered that students who spent more time on their studies achieved better academic results.

Перевод пока недоступен