From rule-based to generative AI: a systematic review of algorithmic evolution and multimodal fusion in VR interview training systems

Engineering notes

Progressing from rule-based systems and machine learning to generative artificial intelligence and large language models, this evolution has significantly enhanced the systems’ interactivity, personalization, and adaptive feedback capabilities.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为端到端自动驾驶、BEV感知、3D目标检测、轨迹预测、路径规划、LiDAR感知等高价值论文补充中文说明。

Original abstract

The interview performance is crucial for employment success. The combination of artificial intelligence and virtual reality provides job seekers with a safe and repeatable immersive training environment. Following PRISMA 2020 guidelines, we conducted a systematic literature review of immersive virtual reality interview training systems across IEEE Xplore, ACM Digital Library, Web of Science, and Google Scholar from 2015 to 2025. Based on our inclusion and exclusion criteria, 23 unique studies were identified for full synthesis. We find that studies increased after 2020, with over half published between 2023 and 2025. This growth was driven by hardware maturation around 2019, and the rise of generative artificial intelligence, notably large language models such as ChatGPT since late 2022. Notably, our findings reveal a technological evolution centered on the adaptive intervention loop of interaction, perception, and decision. Starting with the interaction module initiating interview stimuli, the system utilizes the perception module to capture real-time multimodal data integrating speech, behaviors (e.g., eye movements and facial expressions), and physiological signals (e.g., heart rate and electrodermal activity), allowing the algorithm module to perform intelligent evaluation and decision-making, which in turn drives the interaction module to execute dynamic interventions. Progressing from rule-based systems and machine learning to generative artificial intelligence and large language models, this evolution has significantly enhanced the systems’ interactivity, personalization, and adaptive feedback capabilities. Furthermore, we identify challenges in hardware costs, algorithmic biases, and privacy security, providing a reference for the future development of a more equitable and efficient next-generation intelligent interview system.

Links and sources

• PDF from original source

Comments