Summary of Swe2: Subword Enriched and Significant Word Emphasized Framework For Hate Speech Detection, by Guanyi Mou et al.
SWE2: SubWord Enriched and Significant Word Emphasized Framework for Hate Speech Detection
by Guanyi Mou, Pengyi Ye, Kyumin Lee
First submitted to arxiv on: 25 Sep 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 proposed SWE2 framework is a novel hate speech detection method that only relies on the content of online social network messages to identify hate speech. This approach exploits both word-level semantic information and sub-word knowledge, making it intuitive and effective. The model outperforms 7 state-of-the-art baselines with an accuracy of 0.975 and macro F1 score of 0.953 under no adversarial attack. Furthermore, SWE2 remains robust and performs well even when subjected to extreme character-level adversarial attacks that manipulate up to 50% of the messages, achieving an accuracy of 0.967 and macro F1 score of 0.934. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Online social networks are dealing with a big problem: hate speech. This is when people use mean or hurtful language online, which can make others feel sad, scared, or even bullied. Some smart people are working on ways to detect and stop this kind of behavior. One new approach they’re trying is called SWE2. It’s a computer program that looks at the words in social media messages to figure out if they’re hateful or not. The program works pretty well, correctly identifying hate speech most of the time. |
Keywords
* Artificial intelligence * F1 score
