Summary of Validation, Robustness, and Accuracy Of Perturbation-based Sensitivity Analysis Methods For Time-series Deep Learning Models, by Zhengguang Wang
Validation, Robustness, and Accuracy of Perturbation-Based Sensitivity Analysis Methods for Time-Series Deep Learning Models
by Zhengguang Wang
First submitted to arxiv on: 29 Jan 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 This paper evaluates various interpretability methods for time-series deep learning, focusing on perturbation-based sensitivity analysis. The authors investigate how different post-hoc interpretation methods, such as back-propagation and approximation, perform when applied to modern Transformer models. Specifically, they answer three research questions: whether different sensitivity analysis methods yield similar outputs and importance rankings; whether different deep learning models impact the results of sensitivity analysis; and how well the results align with ground truth. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper looks at ways to understand why AI models make certain decisions when working with time-series data. The researchers compare different techniques that try to figure out how inputs affect outputs, including one called perturbation-based sensitivity analysis. They use this method on special AI models called Transformers and ask three key questions: Do the various methods give similar answers? Do different AI models change the results? And do these results match what we know is true? |
Keywords
* Artificial intelligence * Deep learning * Time series * Transformer
