Powering Finetuning in Few-Shot Learning: Domain-Agnostic Bias Reduction with Selected Sampling


  • Ran Tao Carnegie Mellon University
  • Han Zhang Carnegie Mellon University
  • Yutong Zheng Carnegie Mellon University
  • Marios Savvides Carnegie Mellon University




Machine Learning (ML), Computer Vision (CV)


In recent works, utilizing a deep network trained on meta-training set serves as a strong baseline in few-shot learning. In this paper, we move forward to refine novel-class features by finetuning a trained deep network. Finetuning is designed to focus on reducing biases in novel-class feature distributions, which we define as two aspects: class-agnostic and class-specific biases. Class-agnostic bias is defined as the distribution shifting introduced by domain difference, which we propose Distribution Calibration Module(DCM) to reduce. DCM owes good property of eliminating domain difference and fast feature adaptation during optimization. Class-specific bias is defined as the biased estimation using a few samples in novel classes, which we propose Selected Sampling(SS) to reduce. Without inferring the actual class distribution, SS is designed by running sampling using proposal distributions around support-set samples. By powering finetuning with DCM and SS, we achieve state-of-the-art results on Meta-Dataset with consistent performance boosts over ten datasets from different domains. We believe our simple yet effective method demonstrates its possibility to be applied on practical few-shot applications.




How to Cite

Tao, R., Zhang, H., Zheng, Y., & Savvides, M. (2022). Powering Finetuning in Few-Shot Learning: Domain-Agnostic Bias Reduction with Selected Sampling. Proceedings of the AAAI Conference on Artificial Intelligence, 36(8), 8467-8475. https://doi.org/10.1609/aaai.v36i8.20823



AAAI Technical Track on Machine Learning III