Using generative adversarial network to improve the accuracy of detecting AI-generated tweets

Yang Hui · 2024

Read Using generative adversarial network to improve the accuracy of detecting AI-generated tweets by Yang Hui (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-11599750.txt.

Description

Yang Hui presents a 2024 paper on spotting tweets written by AI. As machine text floods social media, telling it from human writing matters. The method has four steps. Text is first cleaned: noise is cut, words are split, common filler words are dropped, and rare words are handled. Each text then becomes a grid that stores word links and weights.

A generative adversarial network, a pair of rival models, pulls strong features from those grids. Its judging half becomes the feature extractor. A weighted random forest then makes the final call. The pipeline reached about 99.6 percent average accuracy. That beat rival methods by about 1.5 points. The author says the rival-model features capture the gap between human and machine style.

Catalog record

CREATORYang Hui
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE59 KB
VERSION1.0
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC11599750

Browse the full library · About the Tower of Babel