Focusing on Human-AI Interaction through Neuroscience and Intelligent Systems
Welcome to Jin Labs
Research Overview
Dr. Jin Xu is a faculty member at the School of Informatics and Cybersecurity, Technological University Dublin. His interdisciplinary research integrates time-series data processing, neuroscience and AI to address challenges in healthcare, neurotechnology and trustworthy intelligent systems.
- Advanced DSP & Linear Predictive Coding (LPC)
Dr. Xu is interested in foundational DSP techniques, with a focus on refining Linear Predictive Coding (LPC) for high-noise environments. He has developed a novel LPC Pole Processing Method to numerically extract dominant spectral features from non-stationary signals, such as EEG data. This method enhances signal interpretability and robustness and holds significant potential for future DSP-based applications in healthcare and beyond.
- Neuroengineering & Wearable Biosensing
With a strong foundation in EEG signal analysis, Dr. Xu has proposed the use of center frequencies to describe the dominant EEG activies instead of traditional fixed EEG bands (Delta, Theta, Alpha, Beta, Gamma) . He has also developed a multi-modal framework for detecting mental fatigue, leveraging signals such as EEG, Heart Rate Variability (HRV) and Skin Conductance (SC). Dr. Xu is committed to translating this research into an enterprise-level, real-time wearable system for applications in occupational safety, neurorehabilitation and Brain Computer Interfaces (BCIs).
- Intelligent and Trustworthy Medical Diagnosis
Dr. Xu is designing trustworthy, deployable medical AI assistants to address limitations in existing chatbots. Dr. Xu aims to develop a trustworthy, deployable medical chatbot powered by fine-tuned large language models (LLMs) with multimodal inputs, including text, medical images, audio and structured clinical knowledge. The system aims to enhance diagnostic accuracy, clinician workflows, and patient communication. This research prioritises explainability, clinical relevance and regulatory compliance for real-world healthcare integration.
- Trustworthy AI in Recommender Systems
Dr. Xu also investigates fairness and trustworthiness in AI-based recommender systems, particularly in the context of short-form video platforms. He is developing a simulation environment to evaluate how different recommendation algorithms influence user behavior across diverse demographic groups. This framework supports the development of transparent, user-centric AI systems aligned with EU regulations and societal values.
