Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Graph Classification | Accuracy68.7 | 57 | ||
| Maximum Clique | Twitter MC instances (static) | Mean ApR1 | 38 | |
| Aspect-Based Sentiment Analysis | Accuracy79.76 | 35 | ||
| Visual Emotion Recognition | Twitter II | Accuracy91.68 | 34 | |
| Visual Emotion Recognition | Twitter I | Accuracy95.63 | 34 | |
| Feature Attribution | Comprehensiveness76 | 33 | ||
| Target-dependent sentiment classification | Twitter (test) | Accuracy78.61 | 31 | |
| Maximum Independent Set | TWITTER (test) | Approximation Ratio1 | 25 | |
| User-level depression detection | Twitter (test) | Recall84.4 | 24 | |
| Multimodal Named Entity Recognition | Twitter 2017 | F1 Score88.07 | 24 | |
| Multimodal Named Entity Recognition | Twitter 2015 | F1 Score79.33 | 23 | |
| Named Entity Recognition | Twitter NER | F1 Score78.52 | 23 | |
| Personalized user simulation | Sim Fidelity0.7864 | 21 | ||
| Multimodal Named Entity Recognition | Twitter-15 (test) | F1 Score79.21 | 21 | |
| Sentiment Analysis | Accuracy80.59 | 20 | ||
| Rumour Detection | Twitter 15 | Accuracy92.3 | 19 | |
| Aspect term-polarity pair extraction | Twitter ST (test) | F1 Score51.37 | 18 | |
| Sentiment Analysis | Twitter (test) | Accuracy86.6 | 17 | |
| Aspect-Based Sentiment Analysis | Twitter (test) | Acc77.17 | 17 | |
| Aspect Polarity Classification | F1 Score (APC)75.16 | 17 | ||
| Bot Detection | Twitter tweet-level | Precision96 | 17 | |
| Crisis response generation | Professionalism99 | 16 | ||
| Node Classification | Twitter HOMO (test) | Mean Accuracy75.92 | 15 | |
| Node Classification | Twitter HET (test) | Mean Accuracy0.7317 | 15 | |
| Multimodal Aspect Sentiment Classification | TWITTER 2015 | F1 Score78.1 | 15 |
Showing 25 of 126 rows