Summary of Hyper Evidential Deep Learning to Quantify Composite Classification Uncertainty, by Changbin Li et al.
Hyper Evidential Deep Learning to Quantify Composite Classification Uncertainty
by Changbin Li, Kangshuo Li, Yuzhe Ou, Lance M. Kaplan, Audun Jøsang, Jin-Hee Cho, Dong Hyun Jeong, Feng Chen
First submitted to arxiv on: 17 Apr 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Machine Learning (cs.LG)
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 Hyper-Evidential Neural Network (HENN) framework explicitly models predictive uncertainty due to composite class labels in training data, leveraging Subjective Logic’s belief theory. This novel approach treats predictions as parameters of hyper-subjective opinions and learns the network that collects both single and composite evidence leading to these hyper-opinions using a deterministic DNN from data. The results demonstrate HENN outperforms state-of-the-art counterparts on four image datasets. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper introduces a new way for deep neural networks (DNNs) to handle tasks where classes have similar visual features. When human annotators find it hard to tell classes apart, they use composite class labels. The researchers created a special kind of neural network called Hyper-Evidential Neural Network (HENN). HENN helps DNNs understand when predictions are unsure due to these composite class labels. It works by treating predictions as opinions and learning from data what kind of evidence supports these opinions. In the end, HENN does better than other methods on image datasets. |
Keywords
» Artificial intelligence » Neural network
