MELD
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Emotion Recognition in Conversation | MELD | Weighted Avg F170.44 | 180 | |
| Emotion Recognition in Conversation | MELD (test) | Weighted F170.44 | 152 | |
| Emotion Recognition | MELD (test) | Weighted F170.12 | 89 | |
| Multimodal Emotion Recognition in Conversation | MELD standard (test) | WF167.6 | 53 | |
| Multimodal Emotion Recognition in Conversation | MELD | Weighted Avg F1 Score69 | 36 | |
| Emotion Detection | MELD (test) | Weighted-F10.699 | 32 | |
| Speech Emotion Recognition | MELD | Accuracy63.5 | 30 | |
| Emotion Recognition in Conversation | MELD Standard (test) | Weighted F169.15 | 19 | |
| Emotional Speech Synthesis | MELD | BERT Score0.55 | 18 | |
| Emotion Recognition in Conversation | MELD 1.0 (test) | Weighted F165.61 | 17 | |
| Emotion Recognition in Conversation | MELD | F1 Score67.29 | 16 | |
| Basic emotion recognition | MELD (official) | Hitrate71.9 | 15 | |
| Cross-scenario Multimodal Emotion Recognition in Conversations | MELD -> IEMOCAP noise rate 40% (test) | Joy Accuracy41.46 | 15 | |
| Sentiment Classification | MELD (test) | Accuracy68.5 | 15 | |
| Basic Emotion Recognition | MELD | Accuracy66.71 | 14 | |
| Emotion Recognition in Conversation | MELD | Weighted F167.12 | 14 | |
| Speech Emotion Recognition | MELD In-Domain v1 (test) | Accuracy54.06 | 14 | |
| Multimodal Empathetic Response Generation | MELD | Distinct-1 (Dist-1)28.14 | 12 | |
| Emotion Recognition | MELD | UACC64.34 | 12 | |
| Conversational Emotion Recognition | MELD (test) | Macro F1 Score61.9 | 12 | |
| LLM-generated text detection | MELD Overall (eval) | TPR @ 1% FPR99.9 | 10 | |
| LLM-generated text detection | MELD Qwen-3.6-Plus (eval) | TPR @ 1% FPR99.9 | 10 | |
| LLM-generated text detection | MELD-eval Claude-Haiku-4.5 | TPR @ 1% FPR100 | 10 | |
| LLM-generated text detection | MELD GPT-5.4-Mini (eval) | TPR @ 1% FPR100 | 10 | |
| AI-text detection | MELD (eval) | AUROC99.99 | 10 |
Showing 25 of 51 rows