Summary of Ospc: Detecting Harmful Memes with Large Language Model As a Catalyst, by Jingtao Cao et al.
OSPC: Detecting Harmful Memes with Large Language Model as a Catalyst
by Jingtao Cao, Zheng Zhang, Hongru Wang, Bin Liang, Hao Wang, Kam-Fai Wong
First submitted to arxiv on: 14 Jun 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
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 approach to detecting harmful memes integrates image captioning, Optical Character Recognition (OCR), and Large Language Model (LLM) analysis to comprehensively understand and classify harmful memes in multicultural and multilingual contexts like Singapore. The system utilizes the BLIP model for image captioning, PP-OCR and TrOCR for text recognition across multiple languages, and the Qwen LLM for nuanced language understanding. By fine-tuning the approach with GPT-4V labeled data, the framework achieves top performance on the Online Safety Prize Challenge hosted by AI Singapore, outperforming previous benchmarks from FLAVA and VisualBERT. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study helps us better understand and detect harmful memes spreading across the internet. The research team developed a special computer system that can recognize and classify these negative messages in different languages, like English, Chinese, Malay, and Tamil. This system uses advanced AI techniques to analyze images and text, making it more accurate than previous methods. By fine-tuning their approach with extra training data, they were able to make their system even better at identifying harmful memes. |
Keywords
» Artificial intelligence » Fine tuning » Gpt » Image captioning » Language understanding » Large language model
