Summary of Detecting Machine-generated Long-form Content with Latent-space Variables, by Yufei Tian et al.
Detecting Machine-Generated Long-Form Content with Latent-Space Variables
by Yufei Tian, Zeyu Pan, Nanyun Peng
First submitted to arxiv on: 4 Oct 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Machine Learning (cs.LG)
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 abstract proposes a new approach for detecting machine-generated text from human-written ones by incorporating abstract elements such as event transitions. Existing methods primarily focus on token-level distributions, which are vulnerable to real-world domain shifts and adversarial attacks. The proposed latent-space model is trained on sequences of events or topics derived from human-written texts and can distinguish between machine- and human-generated texts with a 31% improvement over strong baselines like DetectGPT. This method is robust in three different domains, including machine-generated texts that are originally inseparable from human texts at the token level. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper solves a big problem: how to tell if text was written by a computer or a person. Right now, it’s hard because computers can write very realistic text. The authors of this paper have come up with a new way to solve this problem that is better than what we currently use. They looked at how events are organized in text and found that machines and humans do it differently. This helps their method work well even when the text looks similar on the surface. |
Keywords
» Artificial intelligence » Latent space » Token
