Balanced VQA against language priors

How this idea evolved

Suggested from citations. No curator has recorded what this idea builds on yet. These ideas from the same theme come from papers that VQA v2 cites, directly or one step removed. Citation is a fact; the connection between the ideas is not verified.

• Open-ended visual question answeringVQA · 2015cited directly

  Answer free-form natural-language questions about an image.

• End-to-end neural VQA (CNN + LSTM)Ask Your Neurons · 2015cited directly

  Encode the image with a CNN and the question with an LSTM, and train everything jointly.

• Generating QA pairs from captionsImage QA models & data · 2015cited directly

  Turn existing image descriptions into question–answer training data automatically.