Summary of Orderbkd: Textual Backdoor Attack Through Repositioning, by Irina Alekseevskaia and Konstantin Arkhipenko
OrderBkd: Textual backdoor attack through repositioning
by Irina Alekseevskaia, Konstantin Arkhipenko
First submitted to arxiv on: 12 Feb 2024
Categories
- Main: Computation and Language (cs.CL)
- 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 AI research paper proposes a novel type of attack on NLP systems that utilize third-party datasets and pre-trained machine learning models. The authors demonstrate the possibility of hidden backdoor attacks by repositioning specific words in sentences, which maintains high success rates on SST-2 and AG classification datasets while outperforming existing attacks in terms of perplexity and semantic similarity. Furthermore, the paper shows the robustness of this attack to the ONION defense method. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine a computer system that can understand human language, but what if someone could secretly alter its understanding? This is exactly what a new type of cyberattack on artificial intelligence (AI) systems does. It’s called a “backdoor” and it allows hackers to manipulate AI models without being detected. The good news is that researchers have discovered this vulnerability and are working on ways to prevent it. |
Keywords
» Artificial intelligence » Classification » Machine learning » Nlp » Perplexity
