Bo Xu, Dongyu Zhang, Liang Yang, Junyu Lu, Xiaokun Zhang, Hongbo Wang, Haohao Zhu, Hongfei Lin
We lifted 2 functions out of this paper's own repositories and ran 1 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.
| Repository | Role | Ran |
|---|---|---|
| dut-lujunyu/toxicn_mm | canonical | 1 of 2 |
| Function | Status | Where it lives |
|---|---|---|
| QKVAttention | Ran | dut-lujunyu/toxicn_mm/model/MHKE.py pointer only (licence: NONE) · get_code("95032268d7670902") |
| MHKE | Not yet run | dut-lujunyu/toxicn_mm/model/MHKE.py pointer only (licence: NONE) · get_code("4663bcae8a86b94f") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
Harmful memes have proliferated on the Chinese Internet, while research on detecting Chinese harmful memes significantly lags behind due to the absence of reliable datasets and effective detectors. To this end, we focus on the comprehensive detection of Chinese harmful memes. We construct TOXICN MM, the first Chinese harmful meme dataset, which consists of 12,000 samples with finegrained annotations for various meme types. Additionally, we propose a baseline detector, Multimodal Knowledge Enhancement (MKE), incorporating contextual information of meme content generated by the LLM to enhance the understanding of Chinese memes. During the evaluation phase, we conduct extensive quantitative experiments and qualitative analyses on multiple baselines, including LLMs and our MKE. The experimental results indicate that detecting Chinese harmful memes is challenging for existing models while demonstrating the effectiveness of MKE. 1 Disclaimer: The samples presented by this paper may be considered profane, offensive, or vulgar.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2410.02378")
get_code_for_paper("2410.02378")
have("2410.02378")
Connect an agent — have() is free.