Rethinking Class-Balanced Methods for Long-Tailed Visual Recognition From a Domain Adaptation Perspective
Year: 2020
Type: conference-paper
Abstract: Object frequency in the real world often follows a power law, leading to a mismatch between datasets with long-tailed class distributions seen by a machine learning model and our expectation of the model to perform well on all classes. We analyze this mismatch from a domain adaptation point of view. First of all, we connect existing class-balanced methods for long-tailed classification to target shift, a well-studied scenario in domain adaptation. The connection reveals that these methods implicitly assume that the training data and test data share the same class-conditioned distribution, which does not hold in general and especially for the tail classes. While a head class could contain abundant and diverse training examples that well represent the expected data at inference time, the tail classes are often short of representative training data. To this end, we propose to augment the classic class-balanced learning by explicitly estimating the differences between the class-conditioned distributions with a meta-learning approach. We validate our approach with six benchmark datasets and three loss functions.
Authors: Muhammad Abdullah Jamal, Matthew Brown, Ming-Hsuan Yang, Liqiang Wang, Boqing Gong
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
Funder: National Science Foundation
Awards: ICE-T:RI: Towards End-to-End Resource Optimization for Time-Critical Computing Using Reinforcement Learning and Program Analysis, BIGDATA: IA: Distributed Semi-Supervised Training of Deep Models and Its Applications in Video Understanding, CAREER: Visual Tracking with Online and Prior Learning, CRII: RI: Multi-Source Domain Generalization Approaches to Visual Attribute Detection
OpenAlex ID: W3035054804
