题目:Bringing the Unobserved into View: Mitigating Selection Bias in Survey Research via Large Language Models
主讲人:许嘉蓉 青年副研究员 信息管理与商业智能系
时间:2025-10-14 13:30-14:30
地点:
线下:思源楼 724室 线上:腾讯822994061/密码722722
内容摘要:
Questionnaire-based surveys often suffer from selection bias when certain populations remain unobserved due to coverage gaps or self-selection. This study investigates whether large language models (LLMs), trained on vast and diverse text corpora, can help bring these unobserved populations into view. We propose a generative framework to address two key questions: Who are the missing respondents, and how might they have responded? This framework enables us to generate the unobserved and evaluate how their inclusion can improve the representativeness of survey research.
活动讲座
新闻动态
微信头条
招生咨询
媒体视角
瞰见云课堂