Papers with Graph Auto-Encoders
Unsupervised Embeddings with Graph Auto-Encoders for Multi-domain and Multilingual Hate Speech Detection (2022.lrec-1)
Copied to clipboard
| Challenge: | Hate speech detection is a challenging task, since hate messages are often expressed in subtle ways and with characteristics that may vary depending on the author. |
| Approach: | They propose an unsupervised approach to learn embeddings for hate speech detection using Graph Auto-Encoders (GAE) they represent texts as nodes of a graph and use transformer layer and convolutional layer to encode them in low-dimensional space. |
| Outcome: | The proposed method shows competitive results on small datasets. |