| Budget Amount *help |
¥17,160,000 (Direct Cost: ¥13,200,000、Indirect Cost: ¥3,960,000)
Fiscal Year 2021: ¥4,030,000 (Direct Cost: ¥3,100,000、Indirect Cost: ¥930,000)
Fiscal Year 2020: ¥4,030,000 (Direct Cost: ¥3,100,000、Indirect Cost: ¥930,000)
Fiscal Year 2019: ¥3,900,000 (Direct Cost: ¥3,000,000、Indirect Cost: ¥900,000)
Fiscal Year 2018: ¥5,200,000 (Direct Cost: ¥4,000,000、Indirect Cost: ¥1,200,000)
|
|---|
| Outline of Final Research Achievements |
Visual Question Answering (VQA) is an interdisciplinary field, lying on the vision and natural language fields, which is recently advanced drastically due to deep learning. Current techniques for VQA rely on rather a statistics approach, where the distribution of the training set solely matters. We need to go beyond this to make VQA more practical. Our core research question is: “Can VQA systems can answer questions that require inference?”, and we have been committed to building a system that uses knowledge for visual question answering (knowledge-based visual question answering; KBVQA), while also exploring an effective video representation.
|
