Summary of Exploring Beyond Logits: Hierarchical Dynamic Labeling Based on Embeddings For Semi-supervised Classification, by Yanbiao Ma et al.
Exploring Beyond Logits: Hierarchical Dynamic Labeling Based on Embeddings for Semi-Supervised Classification
by Yanbiao Ma, Licheng Jiao, Fang Liu, Lingling Li, Shuyuan Yang, Xu Liu
First submitted to arxiv on: 26 Apr 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
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 This paper proposes a novel method for generating pseudo-labels in semi-supervised learning, called Hierarchical Dynamic Labeling (HDL). Unlike existing methods that rely on model predictions, HDL uses image embeddings to generate sample labels. The authors also introduce an adaptive hyperparameter selection method to enhance the versatility of HDL. Additionally, they demonstrate how HDL can be combined with general image encoders like CLIP to serve as a fundamental data processing module. The proposed approach is evaluated through experiments on datasets with class-balanced and long-tailed distributions, showing improved model performance. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps us understand how we can make machines learn better without needing lots of labeled data. Right now, we have methods that use the machine’s own guesses to figure out what things are, but these methods aren’t very good when the data is noisy or biased. This new approach uses a different way to look at images and generate labels, which makes it more reliable and adaptable. By combining this method with other tools, we can create better machines that learn from lots of different types of data. |
Keywords
» Artificial intelligence » Hyperparameter » Semi supervised
