Want To Reduce Labeling Cost? GPT-3 Can Help
This addresses the problem of expensive data labeling for NLP practitioners, offering a generalizable and cost-effective methodology, though it is incremental as it builds on existing GPT-3 capabilities.
The paper tackles the high cost of data annotation in NLP by using GPT-3 as a low-cost labeler, finding it reduces labeling costs by 50% to 96% while maintaining performance on NLU and NLG tasks, and proposes a framework combining GPT-3 labels with human labels for better results with limited budgets.
Data annotation is a time-consuming and labor-intensive process for many NLP tasks. Although there exist various methods to produce pseudo data labels, they are often task-specific and require a decent amount of labeled data to start with. Recently, the immense language model GPT-3 with 175 billion parameters has achieved tremendous improvement across many few-shot learning tasks. In this paper, we explore ways to leverage GPT-3 as a low-cost data labeler to train other models. We find that, to make the downstream model achieve the same performance on a variety of NLU and NLG tasks, it costs 50% to 96% less to use labels from GPT-3 than using labels from humans. Furthermore, we propose a novel framework of combining pseudo labels from GPT-3 with human labels, which leads to even better performance with limited labeling budget. These results present a cost-effective data labeling methodology that is generalizable to many practical applications.