Artificial Intelligence-Driven Construction of Virtual Experimental Environments for Information Technology Teaching

Abstract

The problems of the traditional virtual simulation environment for teaching experimental science with information technology are context rigidity, lack of interaction and high cognitive load. To solve the problems above, we have put forward a framework for building an intelligent educational form-driven virtual experimental environment, and the three parts of this framework are intelligent educational forms, multi-agent computational models and self-evolving semantic associations. First, we will construct virtualized representations and multimodal interaction modes to clarify the generation mechanism of dynamic experimental scenarios and the spatial logic for adaptive regulation of cognitive load; second, based on the concept of multi-agent cooperation, we will build a computational model to achieve behavioural modelling of experimental objects, evolve state graphs driven by heterogeneous data, and establish a reinforcement learning framework for intention recognition and feedback generation; finally, we will design a self-evolution mechanism for the virtual experimental environment that integrates dynamic resource organisation according to domain knowledge graphs, adaptation of virtual instrument operation logic via generative adversarial networks, and semantic annotation and path optimisation for multimodal interaction trajectories. The framework above is now a complete system of theory and technology for virtual laboratory environments in information-technology education.