A quick overview of my background and research
If we’re meeting for a collaboration or interview, this page is a fast way to understand what I’ve worked on: education, publications, research experience, and selected projects. The PDF version below includes the full details.
Education
BSc in Information and Computing Science
Xi'an Jiaotong-Liverpool University · Sep 2022 - Jul 2026 (expected)
Current GPA: 3.63/4.00. Capstone project: Listening or Reading? An Empirical Study of Modality Importance Analysis Across AQA Question Types.
Publications
Listening or Reading? An Empirical Study of Modality Importance Analysis Across AQA Question Types
DCASE 2025 Workshop · 2025
Zeyu Yin, Yiqiang Cai, Pingsong Deng, Xinyang Lyu, Shengchen Li
ECHOTWIN-QA: A Dual-Tower BEATSBERT System for DCASE 2025 Task 5 Audio Question Answering
DCASE 2025 Challenge (Task 5) · 2025
Zeyu Yin, Ziyang Zhou, Yiqiang Cai, Shengchen Li, Xi Shao
ADAPTF-SEPNET: AudioSet-Driven Adaptive Pre-training of TF-SEPNet for Multi-device Acoustic Scene Classification
DCASE 2025 Challenge · 2025
Ziyang Zhou, Zeyu Yin, Yiqiang Cai, Shengchen Li, Xi Shao
EmoSound: A Multimodal AI Agent for Emotion-Aware Audio Accompaniment of Emoticons
BICS 2025 · 2025
Jianghui Sun, Haosen Shi, Zeyu Yin, Wansu Mo, Hongyi Ding, Yiming Hu, Xi Yang, Yuyao Yan
Research experience
Participant / Researcher, DCASE 2025 Challenge & Workshop
2025 · Barcelona, Spain
Participated in DCASE 2025 Task 5: Audio Question Answering and the DCASE 2025 Workshop. Developed an end-to-end AQA system and conducted experiments and ablation studies to analyze modality importance across question types.
Undergraduate Researcher, Summer Undergraduate Research Fellowship (SURF)
Summer 2024 · Suzhou, China
Worked on Expressive Timing Modelling in Performed Classical Piano Music.
Projects
Listening or Reading? An Empirical Study of Modality Importance Analysis Across AQA Question Types
Audio Question Answering · modality weighting · Acoustic Reasoning · DCASE 2025
A research case study investigating how audio QA systems trade off acoustic evidence against semantic and text-derived shortcuts across question types.
ECHOTWIN-QA: A Dual-Tower BEATSBERT System for DCASE 2025 Task 5 Audio Question Answering
Challenge system · BEATSBERT · PyTorch
A full end-to-end AQA system built for the DCASE 2025 challenge, including training, evaluation, ablations, and technical reporting.
ADAPTF-SEPNET: AudioSet-Driven Adaptive Pre-training of TF-SEPNet for Multi-device Acoustic Scene Classification
Acoustic scene classification · Challenge · Evaluation
A collaborative challenge project focused on model development, experimental evaluation, and analysis for acoustic scene classification.
EmoSound: A Multimodal AI Agent for Emotion-Aware Audio Accompaniment of Emoticons
Multimodal agent · BICS 2025 · Evaluation
A multimodal AI agent project where I implemented components for the agent and evaluation pipeline, and contributed to experiments and writing.
Skills
Programming, ML, web, and systems
Python, Java, JavaScript/TypeScript, SQL, Bash, PyTorch, HuggingFace, NestJS, Spring Boot, Node.js
Also experienced with scikit-learn, pandas, NumPy, MySQL, Postgres, Linux, Docker, Git, Cloudflare Tunnel, SLURM, LaTeX, and Markdown.
Honors
University Academic Excellence Award
2023 · Full Scholarship
Received the university academic excellence award with full scholarship support.
National Second Prize, Future Engineer Competition (Smart Invention)
2021, 2022
Won the national second prize twice; developed an Arduino-based intelligent system integrating AI technology.