Delving into Deep Imbalanced Regression
This work addresses imbalanced regression for real-world tasks in domains like computer vision and healthcare, filling a gap in benchmarks and techniques, though it is incremental as it builds on existing imbalanced learning methods.
The paper tackles the problem of learning from imbalanced data with continuous targets, proposing distribution smoothing for labels and features to address missing data and improve generalization across the target range, with extensive experiments showing superior performance on curated large-scale datasets.
Real-world data often exhibit imbalanced distributions, where certain target values have significantly fewer observations. Existing techniques for dealing with imbalanced data focus on targets with categorical indices, i.e., different classes. However, many tasks involve continuous targets, where hard boundaries between classes do not exist. We define Deep Imbalanced Regression (DIR) as learning from such imbalanced data with continuous targets, dealing with potential missing data for certain target values, and generalizing to the entire target range. Motivated by the intrinsic difference between categorical and continuous label space, we propose distribution smoothing for both labels and features, which explicitly acknowledges the effects of nearby targets, and calibrates both label and learned feature distributions. We curate and benchmark large-scale DIR datasets from common real-world tasks in computer vision, natural language processing, and healthcare domains. Extensive experiments verify the superior performance of our strategies. Our work fills the gap in benchmarks and techniques for practical imbalanced regression problems. Code and data are available at https://github.com/YyzHarry/imbalanced-regression.