Advancing STEM cognition with current AI landscape and systems

Application of AI explores the potential of algorithms to ensure fairness, accuracy, and efficiency in grading students' performance, offering valuable insights into their strengths and areas for improvement. While the current AI landscape showcases remarkable progress, there are several areas ripe for exploration. One such avenue is AI steps to consider in STEM, wherein researchers aim to develop specialised steps/models to understand and generate domain-specific STEM content. The systematic literature review highlighted the importance of domain adaptation techniques for enhancing STEM comprehension by fine-tuning transformer-based language models like BERT. Integrating domain knowledge through ontology-based and context of STEM disciplines. Future research should focus on building domain-specific annotated datasets to improve the performance models in STEM comprehension. Additionally, exploring unsupervised domain adaptation techniques and leveraging domain-specific knowledge graphs can further enhance the NLP models’ adaptability to diverse STEM domains.