Human Still Wins over LLM: An Empirical Study of Active Learning on Domain-Specific Annotation Tasks.
Yuxuan LuBingsheng YaoShao ZhangYun WangPeng ZhangTun LuToby Jia-Jun LiDakuo WangPublished in: CoRR (2023)
Keyphrases
- batch mode
- active learning
- domain specific
- learning algorithm
- semi supervised
- annotation effort
- human users
- supervised learning
- transfer learning
- domain independent
- general purpose
- cost sensitive
- relevance feedback
- learning process
- labeled data
- human operators
- random sampling
- training examples
- human behavior
- human experts
- information extraction
- metadata