Summary of Vcc-infuse: Towards Accurate and Efficient Selection Of Unlabeled Examples in Semi-supervised Learning, by Shijie Fang et al.
VCC-INFUSE: Towards Accurate and Efficient Selection of Unlabeled Examples in Semi-supervised Learning
by Shijie Fang, Qianhan Feng, Tong Lin
First submitted to arxiv on: 18 Apr 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computer Vision and Pattern Recognition (cs.CV)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
Summary difficulty | Written by | Summary |
---|---|---|
High | Paper authors | High Difficulty Summary Read the original abstract here |
Medium | GrooveSquid.com (original content) | Medium Difficulty Summary A novel Semi-supervised Learning (SSL) framework is proposed to effectively utilize unlabeled data for improved classification performance and reduced training time. The existing pseudo-label-based methods are inadequate as they rely on inaccurate confidence scores from the classifier, leading to poor results. To address these limitations, the authors introduce two techniques: Variational Confidence Calibration (VCC) and Influence-Function-based Unlabeled Sample Elimination (INFUSE). VCC is a confidence calibration method that utilizes a variational autoencoder to select more accurate pseudo labels based on three types of consistency scores. INFUSE is a data pruning method that constructs a core dataset of unlabeled examples under SSL, eliminating unnecessary samples and reducing training time. The proposed methods demonstrate effectiveness in multiple datasets and settings, achieving a 1.08% reduction in error rate for FlexMatch on the CIFAR-100 dataset while saving nearly half of the training time. |
Low | GrooveSquid.com (original content) | Low Difficulty Summary A new way to use extra data without labels is being developed. Right now, most methods pick examples based on how confident they are that they’re correct, but this isn’t very accurate. The proposed method, called Variational Confidence Calibration (VCC), uses a special kind of computer model to pick better examples. Another part of the method, called Influence-Function-based Unlabeled Sample Elimination (INFUSE), helps by getting rid of unnecessary data. This makes training faster and more effective. |
Keywords
» Artificial intelligence » Classification » Pruning » Semi supervised » Variational autoencoder