A Large Model Based Data Agent for Data Science and Artificial Intelligence Education
For many students, the leap into data science feels like staring at an unscalable wall of complex syntax and rigid coding requirements. This barrier is particularly daunting for those in the humanities and social sciences, who often find their innovative ideas sidelined by a lack of programming expertise. Simultaneously, educators are stretched thin, struggling to provide personalised guidance in large-scale classrooms while keeping pace with the relentless evolution of AI. Traditional commercial tools often exacerbate these issues, functioning as "black boxes" that obscure the learning process and raise significant concerns regarding data privacy and customisability.
Led by Prof. HUANG Jian, Chair Professor of Data Science and Analytics, Department of Data Science and Artificial Intelligence and the Department of Applied Mathematics, LAMBDA is a large model-based data agent designed to democratise the digital frontier. LAMBDA functions as an open-source, code-free multi-agent platform that breathes life into the concept of inclusive education. By employing a collaborative system, the platform allows students to focus on high-level conceptual thinking rather than the minutiae of debugging. A dedicated "programmer" agent translates natural language instructions into executable code, while a vigilant "inspector" agent automatically refines the output. This synergy enables students from any academic background to perform intricate data preprocessing, sophisticated visualisation, and rigorous model evaluation through simple dialogue.
The brilliance of LAMBDA lies in its adaptability for the modern academic. Through a unique Knowledge Integration Mechanism, professors can seamlessly weave the latest AI research and external algorithms into the platform, ensuring that the curriculum remains at the bleeding edge of the industry. This transparency not only fosters a deeper understanding of how AI arrives at its conclusions but also empowers educators to scale their impact, reducing the manual burden of code review and allowing more time for deep, pedagogical engagement. Early pilots have already revealed a surge in student confidence, proving that when the technical "noise" is removed, genuine intellectual curiosity flourishes.
This project stands as a cornerstone of PolyU Education 4.0, embodying an AI-empowered, student-centred model that redefines the classroom experience. By transitioning from passive consumption to active, agent-led discovery, LAMBDA reinforces PolyU’s standing as a world-class innovative institution. It ensures that every student, regardless of their primary discipline, is equipped to lead in a data-driven future, firmly establishing Hong Kong as a global hub for inclusive technological excellence.
This project is awarded under the Teaching Development Grant 2025–28 in the Theme-based Projects category (Individual Project).
