Summary of Robust Semi-supervised Learning by Wisely Leveraging Open-set Data, By Yang Yang et al.
Robust Semi-supervised Learning by Wisely Leveraging Open-set Data
by Yang Yang, Nan Jiang, Yi Xu, De-Chuan Zhan
First submitted to arxiv on: 11 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
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 The proposed Wise Open-set Semi-supervised Learning (WiseOpen) framework is designed to handle open-set semi-supervised learning (OSSL) scenarios where unlabeled data may come from classes unseen in the labeled set. To tackle this issue, WiseOpen employs a gradient-variance-based selection mechanism that selectively leverages a friendly subset of the open-set data for training, rather than using the entire dataset. This approach is shown to enhance the model’s capability of in-distribution (ID) classification while reducing computational expense. Two practical variants are also proposed, which adopt low-frequency updates and loss-based selection respectively. The effectiveness of WiseOpen is demonstrated through extensive experiments compared to state-of-the-art approaches. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Wise Open-set Semi-supervised Learning (OSSL) is a way for machines to learn from some labeled data and a lot of unlabeled data. But what if the unlabeled data comes from classes that we didn’t see before? This can cause problems for machines learning from this data. To solve this issue, researchers developed a new approach called WiseOpen. It selects only the useful parts of the unlabeled data to help the machine learn better. This makes the machine more accurate at classifying things it has seen before. Two ways to make WiseOpen work even faster are also proposed. The results show that WiseOpen is better than other approaches at doing this task. |
Keywords
» Artificial intelligence » Classification » Semi supervised
