Summary of Balancing Label Quantity and Quality For Scalable Elicitation, by Alex Mallen and Nora Belrose
Balancing Label Quantity and Quality for Scalable Elicitation
by Alex Mallen, Nora Belrose
First submitted to arxiv on: 17 Oct 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary Medium Difficulty summary: The paper explores scalable oversight methods for training and evaluating AI systems in domains where human judgment is unreliable or expensive. It considers the complementary problem of training models with low-quality labels, finding that large pretrained models often have an inductive bias towards producing correct answers. The study focuses on the microeconomics of the quantity-quality tradeoff on binary NLP classification tasks used in previous work. It finds three regimes of eliciting classification knowledge from pretrained models using supervised finetuning: quantity-dominant, quality-dominant, and a mixed regime involving low- and high-quality data together to attain higher accuracy at a lower cost. The paper establishes a Pareto frontier of scalable elicitation methods that optimally trade off labeling cost and classifier performance. By adding a few-shot prompt to make use of the model’s existing knowledge of the task, the accuracy of supervised fine-tuning can be improved by up to 5 percentage points at a fixed labeling budget. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Low Difficulty summary: This paper looks at how we can train AI systems when human judgment is hard or expensive. It explores what happens when we use models that were trained on low-quality data, and finds that these models often produce correct answers because of their existing knowledge. The study focuses on the problem of getting the right balance between having enough data and having good quality data. It finds that there are different ways to approach this tradeoff, and that adding a few extra pieces of information can help improve the accuracy of the model’s predictions. |
Keywords
» Artificial intelligence » Classification » Few shot » Fine tuning » Nlp » Prompt » Supervised
