Summary of Timeinf: Time Series Data Contribution Via Influence Functions, by Yizi Zhang et al.
TimeInf: Time Series Data Contribution via Influence Functions
by Yizi Zhang, Jingyan Shen, Xiaoxue Xiong, Yongchan Kwon
First submitted to arxiv on: 21 Jul 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Machine Learning (stat.ML)
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 TimeInf, a novel method for estimating the contribution of individual data points to a model’s prediction in time-series datasets. This is critical for interpreting model predictions and improving model performance. Existing methods have focused on independent and identically distributed (i.i.d.) settings, but not on handling temporal dependencies inherent in time-series data. TimeInf uses influence functions to attribute model predictions to individual time points while preserving temporal structures. The results demonstrate that TimeInf outperforms state-of-the-art methods in identifying harmful anomalies and helpful time points for forecasting. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps us better understand how models make predictions from time-series data. Right now, we can’t easily tell which parts of the data are most important for making accurate predictions. The authors introduce a new method called TimeInf that helps us figure out which pieces of data contribute most to a model’s prediction. This is useful because it lets us identify patterns in the data that might be helping or hurting our predictions. By using this method, we can get a better understanding of how our models work and make them more accurate. |
Keywords
* Artificial intelligence * Time series
