Mind Your Outliers! Investigating the Negative Impact of Outliers on Active Learning for Visual Question Answering
arXiv:2107.02331
Abstract
Active learning promises to alleviate the massive data needs of supervised machine learning: it has successfully improved sample efficiency by an order of magnitude on traditional tasks like topic classification and object recognition. However, we uncover a striking contrast to this promise: across 5 models and 4 datasets on the task of visual question answering, a wide variety of active learning approaches fail to outperform random selection. To understand this discrepancy, we profile 8 active learning methods on a per-example basis, and identify the problem as collective outliers -- groups of examples that active learning methods prefer to acquire but models fail to learn (e.g., questions that ask about text in images or require external knowledge). Through systematic ablation experiments and qualitative visualizations, we verify that collective outliers are a general phenomenon responsible for degrading pool-based active learning. Notably, we show that active learning sample efficiency increases significantly as the number of collective outliers in the active learning pool decreases. We conclude with a discussion and prescriptive recommendations for mitigating the effects of these outliers in future work.
arXiv commentAccepted at ACL-IJCNLP 2021. 17 pages, 16 Figures
References in corpus (12)
- What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?
- Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings
- Bayesian Active Learning for Classification and Preference Learning
- Exploring Models and Data for Image Question Answering
- Simple Baseline for Visual Question Answering
- Adversarial Filters of Dataset Biases
- A Benchmark and Comparison of Active Learning for Logistic Regression
- Listen, Attend, and Walk: Neural Mapping of Navigational Instructions to Action Sequences
- An Empirical Study of Example Forgetting during Deep Neural Network Learning
- Selection via Proxy: Efficient Data Selection for Deep Learning
- Active Learning for Visual Question Answering: An Empirical Study
- Deep Bayesian Active Learning for Multiple Correct Outputs