（五十九）：Bottom-Up and Top-Down Attention for Image Captioning and Visual Question Answering

（五十九）：Bottom-Up and Top-Down Attention for Image Captioning and Visual Question Answering

• Abstract
• 1. Introduction
• 2. Related Work
• 3. Approach
• • 3.1. Bottom-Up Attention Model
  • 3.2. Captioning Model
  • • 3.2.1 Top-Down Attention LSTM
    • 3.2.2 Language LSTM
    • 3.2.3 Objective
  • 3.3. VQA Model
• 4. Evaluation
• • 4.1. Datasets
  • • 4.1.1 Visual Genome