Summary of Counterfactual Explanations For Multivariate Time-series Without Training Datasets, by Xiangyu Sun et al.
Counterfactual Explanations for Multivariate Time-Series without Training Datasets
by Xiangyu Sun, Raquel Aoki, Kevin H. Wilson
First submitted to arxiv on: 28 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Methodology (stat.ME)
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 presents CFWoT, a novel reinforcement-learning-based counterfactual explanation (CFE) method that generates CFEs when training datasets are unavailable. The proposed approach is model-agnostic and suitable for both static and multivariate time-series datasets with continuous and discrete features. Key features of CFWoT include the ability to specify non-actionable, immutable, and preferred features, as well as causal constraints which the method guarantees will be respected. The authors demonstrate the performance of CFWoT against four baselines on several datasets, finding that despite not having access to a training dataset, CFWoT finds CFEs that make significantly fewer and smaller changes to the input time-series. This property makes CFEs more actionable, as the magnitude of change required to alter an outcome is vastly reduced. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper introduces a new way for machines to explain their decisions without needing all the data they were trained on. It’s called counterfactual explanations (CFEs), and it helps people understand how small changes to inputs can affect outcomes. The problem with current CFE methods is that they often need access to the training data, which isn’t always available. This new method, called CFWoT, can create CFEs without needing the original data. It’s flexible and can handle different types of data and constraints. The authors tested their approach on several datasets and found it performed well compared to other methods. |
Keywords
» Artificial intelligence » Reinforcement learning » Time series
