Research map / Vision-language and multimodal / Visual question answering
Affective Visual Dialog: A Large-Scale Benchmark for Emotional Reasoning Based on Visually Grounded Conversations
Kilichbek Haydarov, Xiaoqian Shen, Avinash Madasu, Mahmoud Salem, Li-Jia Li, Gamaleldin F Elsayed, Mohamed Elhoseiny
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add dialog D to the input of explanation generation models (ablation) | T5-Large · AffectVisDial Questioner explanation | BLEU: 0.05 -> 0.26 | Improved |
| add dialog D to the input (ablation) | BART-Large · AffectVisDial Questioner explanation | Emo-F1: 2.10 -> 51.62 | Improved |
| add dialog D to input of zero-shot GPT-4 (ablation) | GPT-4 zero-shot · AffectVisDial | Emo-F1: 24.28 -> 29.79 | Improved |
| add image description I to text-only models (ablation) | T5-Large / BART (no dialog) · AffectVisDial Questioner explanation | BLEU / Emo-F1: Emo-F1 15.20 -> 38.11 (T5, no dialog) | Improved |
| add image description I when dialog is already present (ablation) | BART-Large · AffectVisDial Questioner explanation | Emo-F1: 51.62 -> 50.43 | No clear effect |
| remove emotion labels E from the input (ablation) | BART-Large · AffectVisDial Questioner explanation | Emo-F1: 51.62 -> 47.49 | Hurt |
| fine-tune MiniGPT-4-v2 on AffectVisDial instead of zero-shot | MiniGPT-4-v2 · AffectVisDial explanation | Emo-F1: 25.28 -> 46.62 | Improved |
| use discriminative instead of generative models for dialog QA | LTMI · AffectVisDial dialog QA | R@1: 0.18 -> 0.25 | Improved |
| fine-tune on AffectVisDial for G-Eval groundedness and relevance | LLaMa2-7b · G-Eval | Relevance: 2.02 -> 2.34 | Improved |
| fine-tune LLaMa2-7b on AffectVisDial for coherence and fluency | LLaMa2-7b · G-Eval | Fluency: 3.10 -> 2.83 | Hurt |
Closest papers by meaning
- Customizing Visual Emotion Evaluation for MLLMs: An Open-vocabulary, Multifaceted, and Scalable Approach (ICLR 2026 · similarity 0.89)
- AffectGPT: A New Dataset, Model, and Benchmark for Emotion Understanding with Multimodal Large Language Models (ICML 2025 · similarity 0.89)
- VidEmo: Affective-Tree Reasoning for Emotion-Centric Video Foundation Models (NeurIPS 2025 · similarity 0.88)
- EEmo-Logic: A Unified Dataset and Multi-Stage Framework for Comprehensive Image-Evoked Emotion Assessment (ICML 2026 · similarity 0.87)
- MME-Emotion: A Holistic Evaluation Benchmark for Emotional Intelligence in Multimodal Large Language Models (ICLR 2026 · similarity 0.87)
- EmpathyChat: Structured Cognitive Reasoning in Empathetic Spoken Dialogue (NeurIPS 2026 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
