Clarify: Improving Model Robustness With Natural Language Corrections
arXiv:2402.03715 · doi:10.1145/3654777.3676362
Abstract
The standard way to teach models is by feeding them lots of data. However, this approach often teaches models incorrect ideas because they pick up on misleading signals in the data. To prevent such misconceptions, we must necessarily provide additional information beyond the training data. Prior methods incorporate additional instance-level supervision, such as labels for misleading features or additional labels for debiased data. However, such strategies require a large amount of labeler effort. We hypothesize that people are good at providing textual feedback at the concept level, a capability that existing teaching frameworks do not leverage. We propose Clarify, a novel interface and method for interactively correcting model misconceptions. Through Clarify, users need only provide a short text description of a model's consistent failure patterns. Then, in an entirely automated way, we use such descriptions to improve the training process. Clarify is the first end-to-end system for user model correction. Our user studies show that non-expert users can successfully describe model misconceptions via Clarify, leading to increased worst-case performance in two datasets. We additionally conduct a case study on a large-scale image dataset, ImageNet, using Clarify to find and rectify 31 novel hard subpopulations.
UIST 2024. Interface code available at https://github.com/yoonholee/Clarify
References in corpus (14)
- Learning Transferable Visual Models From Natural Language Supervision
- Shortcut Learning in Deep Neural Networks
- LAION-5B: An open large-scale dataset for training next generation image-text models
- Snorkel: Rapid Training Data Creation with Weak Supervision
- Everyday algorithm auditing: Understanding the power of everyday users in surfacing harmful algorithmic behaviors
- Understanding Practices, Challenges, and Opportunities for User-Engaged Algorithm Auditing in Industry Practice
- Discovering and Validating AI Errors With Crowdsourced Failure Reports
- Zeno: An Interactive Framework for Behavioral Evaluation of Machine Learning
- Concept Induction: Analyzing Unstructured Text with High-Level Concepts Using LLooM
- Model Sketching: Centering Concepts in Early-Stage Machine Learning Model Design
- Mind the Gap: Assessing Temporal Generalization in Neural Language Models
- Refining Language Models with Compositional Explanations
- Contrastive Adapters for Foundation Model Group Robustness
- Editing a classifier by rewriting its prediction rules