Summary of Sset: Swapping-sliding Explanation For Time Series Classifiers in Affect Detection, by Nazanin Fouladgar et al.
SSET: Swapping-Sliding Explanation for Time Series Classifiers in Affect Detection
by Nazanin Fouladgar, Marjan Alirezaie, Kary Främling
First submitted to arxiv on: 16 Oct 2024
Categories
- Main: Machine Learning (cs.LG)
- 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 The proposed paper introduces a novel explanation method for multivariate time series classifiers called SSET (Swapping–Sliding Decision Explanation). This approach aims to provide accurate explanations for ML models in real-world applications and non-differentiable settings. The method consists of two stages: swapping, which detects important variables by replacing the series of interest with close train data from target classes, and sliding, which explores salient observations by moving a window over each time step. This technique is applied to affect detection domain using a deep convolutional classifier called CN-Waterfall, achieving superior performance compared to prior models. The SSET method outperforms benchmarks such as LIME, integrated gradients, and Dynamask on two real-world physiological time series datasets, WESAD and MAHNOB-HCI. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper proposes a new way to explain why machine learning models make certain decisions when working with data that changes over time. The approach is called SSET and it helps us understand which parts of the data are most important for making predictions. This can be very useful in real-life applications where we need to know what factors contribute to a particular outcome. The method uses two steps: swapping and sliding, which help identify the key variables that cause changes in the predictions. This technique is applied to detect human emotions and shows better results than other methods. |
Keywords
» Artificial intelligence » Machine learning » Time series
