A personal academic profile with research motivation and long-term direction
This page keeps the tone scholarly and direct, with room for a portrait, a concise narrative, and a clear view of technical strengths.
Conference photo
Background
I am an Information and Computing Science undergraduate at Xi'an Jiaotong-Liverpool University. My recent work includes audio question answering, modality importance analysis, and challenge-oriented audio system development.
Research motivation
What motivates me most is understanding what kinds of abstractions audio models learn, how those abstractions transfer across tasks, and how to diagnose failures when performance depends too heavily on superficial cues rather than the intended level of audio understanding.
Technical strengths
Research proposal direction
My research proposal focuses on hierarchical audio intelligence: a unified, testable representation and evaluation protocol that identifies abstraction levels, measures cross-task transfer, and connects hierarchy to compute-efficient adaptation and robustness.