Summary of Backtime: Backdoor Attacks on Multivariate Time Series Forecasting, by Xiao Lin et al.
BACKTIME: Backdoor Attacks on Multivariate Time Series Forecasting
by Xiao Lin, Zhining Liu, Dongqi Fu, Ruizhong Qiu, Hanghang Tong
First submitted to arxiv on: 3 Oct 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
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 tackles the problem of malicious attacks on Multivariate Time Series (MTS) forecasting models, a crucial concern for their reliable deployment in high-stakes scenarios. The authors propose an attack method called BackTime, which injects stealthy triggers into the MTS data to manipulate predictions. Specifically, BackTime identifies vulnerable timestamps and generates adaptive triggers using a graph neural network-based approach. Extensive experiments demonstrate the effectiveness of the attacks across multiple datasets and state-of-the-art MTS forecasting models. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about making sure that the machines we use for predicting things like traffic or weather don’t get tricked into giving us bad information. Someone might try to make a machine think something is happening when it’s not, just to mess with us. The authors of this paper came up with a way to do exactly that, and they tested it on different kinds of data and machines. They found that their method works really well and can be used in lots of different situations. |
Keywords
» Artificial intelligence » Graph neural network » Time series
