International Journal of Information Technology

Vol. 30, No. 2, 2024

Editorial for

International Journal for Information Technology (IJIT)
Special Issue on Advances in Large Language Models

Large Language Models (LLMs) have rapidly evolved into a foundational paradigm for intelligent systems that can reason, generate, retrieve, and interact across diverse modalities. This special issue examines recent advances in LLMs and LLM-enabled systems, with attention to both technical capability and real-world applicability. From retrieval-augmented generation to multimodal learning, audio perception, and healthcare decision support, the contributions reflect a shift toward more reliable, context-aware, and domain-aligned intelligence.

The works featured in this issue span methodological reflection and applied innovation. Research on retrieval-augmented generation revisits when retrieval improves LLM performance and when it introduces failure modes, encouraging a more nuanced understanding of knowledge integration. In education, multimodal LLMs for gamified learning demonstrate how these systems can support engagement, personalization, and active learning in higher education, particularly in computer science contexts.

This issue also highlights the importance of robust interfaces between LLMs and complex real-world data. Work on music source separation using Gaussian noise injection explores how perception systems can provide more reliable audio inputs for LLM-facing applications. In healthcare, a relationship-aware copilot illustrates the potential of LLM-based decision support designed around clinician needs, contextual relationships, and responsible human-AI collaboration.

Together, these contributions invite broader reflection on the design, evaluation, and deployment of LLM technologies. Beyond raw capability, the emphasis is shifting toward reliability, interpretability, resilience, and domain sensitivity. By bringing together perspectives from retrieval, education, audio processing, and healthcare, this special issue advances a vision of LLM systems that are not only powerful, but also trustworthy, adaptable, and responsive to real human and institutional needs.

Yours Sincerely

Zhang Chengqi, Special Issue Editor