Summary of Uncertainty Measurement Of Deep Learning System Based on the Convex Hull Of Training Sets, by Hyekyoung Hwang et al.
Uncertainty Measurement of Deep Learning System based on the Convex Hull of Training Sets
by Hyekyoung Hwang, Jitae Shin
First submitted to arxiv on: 25 May 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 A novel approach to deep learning uncertainty quantification has been proposed, aiming to reduce accidents and losses due to misjudgment by DNNs. The method, dubbed To-hull Uncertainty and Closure Ratio, measures uncertainty based on the convex hull of training data. This allows for the observation of positional relationships between learned data and unseen samples, enabling the inference of extrapolation from the convex hull. Compared to state-of-the-art test selection metrics, the proposed approach was found to be effective in identifying samples with unusual patterns, such as adversarial attacks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper proposes a new way to understand how uncertain Deep Neural Networks (DNNs) are. The goal is to prevent accidents and losses caused by DNN mistakes. One problem with current methods is that they don’t tell us how much data is outside what the model has learned. This new method, To-hull Uncertainty, looks at the shape of the training data and compares it to new, unseen samples. It’s like trying to see if a new sample is inside or outside the “convex hull” of the trained data. The researchers tested their approach on popular datasets and DNN models, and it worked better than other methods in finding unusual patterns. |
Keywords
» Artificial intelligence » Deep learning » Inference
