Hi, I’m Noah Chou. I work as noahisarider. My research lives where large models meet media and people — multimodal fake-news detection, LLM alignment, multi-agent simulation, human–AI collaboration, and recommender systems. My development side leans into small web instruments and the interfaces that make these models usable.
I am an undergraduate at South China University of Technology, majoring in software engineering with a minor in business administration. IELTS 8.0/9.0. GPA 3.95/4.00, ranked first in my year.
News communication and large models — with Prof. Quanyi Zou, 2024.05–
How large models behave in a news environment: detecting fake news across text and image, the moderation layer underneath it, and the alignment work that decides what a model is willing to say. Three papers, one accepted.
Towards Secure Blockchain-based IoT: Accurate and Explainable Smart Contract Vulnerability Detection via Counterfactual Contrastive Explanations (under review), IEEE Internet of Things Journal
Multi-task fake news detection (in progress)
One patent, pending (expected 2026)
Multimodal AI mental health coaching — with Prof. Yuanyuan Dang, 2026.05–
A coaching system built to be tested in a randomised controlled trial: retrieval over a curated knowledge base, tool calls over MCP, and an evaluation harness that keeps the model honest. Contributed to the manuscript in preparation.
AIGC clinical case generation — with Prof. Yuanyuan Dang, 2024.07–2024.12
Fine-tuned Whisper and Qwen to turn clinical dictation into structured patient cases.
Tourism trajectory data mining — with Prof. Zikun Deng, 2025.08–
Mining tourist trajectory data to test whether what people post steers where they go.
Tourism trajectory simulation and visitor-flow prediction (in progress)
Agricultural automatic evaluation system — 2025.07–2026.02, seven-person team
An applied project: drone and camera imagery in, crop and pest assessments out, with an LLM writing the field report.
A red-team audit framework for LLM and agentic recommendation and shopping assistants: a threat taxonomy mapped to OWASP ASI 2026, an attack library, a dual-channel judge, defense plugins, a leaderboard and compliance reports.
An on-chain governance research toolchain: DAO and Web3 data collection from Snapshot, Tally, EVM and Steemit into research-ready panels, then causal inference — difference-in-differences, PSM-DID, placebo and event study — with the datasets published alongside.
Open-source AI heart murmur detection for primary care: multi-location masked-attention fusion, scored on the official PhysioNet-2022 challenge metric.
The classic models of Mankiw’s Principles of Economics, from the production possibilities frontier to the Phillips curve, turned into runnable and testable Python.