Borrowing Human Senses: Comment-Aware Self-Training for Social Media Multimodal Classification
This work addresses the problem of automating multimodal understanding for social media content, which is incremental as it builds on existing methods by incorporating comment features and self-training.
The paper tackles the challenge of implicit image-text relations in social media multimodal classification by leveraging user comments and self-training, achieving improved performance over previous state-of-the-art models on four benchmarks.
Social media is daily creating massive multimedia content with paired image and text, presenting the pressing need to automate the vision and language understanding for various multimodal classification tasks. Compared to the commonly researched visual-lingual data, social media posts tend to exhibit more implicit image-text relations. To better glue the cross-modal semantics therein, we capture hinting features from user comments, which are retrieved via jointly leveraging visual and lingual similarity. Afterwards, the classification tasks are explored via self-training in a teacher-student framework, motivated by the usually limited labeled data scales in existing benchmarks. Substantial experiments are conducted on four multimodal social media benchmarks for image text relation classification, sarcasm detection, sentiment classification, and hate speech detection. The results show that our method further advances the performance of previous state-of-the-art models, which do not employ comment modeling or self-training.