QUALIFIER: Question-Guided Self-Attentive Multimodal Fusion Network for Audio Visual Scene-Aware Dialog

Keyphrases

• Audio-Video 100%
• Video Scene 100%
• Self-attentive 100%
• Feature Fusion Network 100%
• Scene-aware 100%
• Audio-visual Scene-aware Dialog 100%
• Feature Fusion 75%
• Feature Extraction 50%
• Multimodal Fusion 50%
• Visual Question Answering 50%
• First Use 25%
• Question Generation 25%
• Human Evaluation 25%
• Network Architecture 25%
• Feature Extraction Methods 25%
• Natural Language 25%
• Concatenated 25%
• Ablation Study 25%
• Correlated Information 25%
• DSTC7 25%
• Fused Features 25%
• Comprehensive Feature 25%
• Better Performance 25%
• High Complexity 25%
• Question Answering Task 25%

Computer Science

• Feature Fusion 100%
• Feature Extraction 100%
• Visual Question Answering 100%
• Audiovisual Scene 100%
• Experimental Result 50%
• Network Architecture 50%
• Baseline Method 50%
• Self-Attention Mechanism 50%