Premise-based Multimodal Reasoning: Conditional Inference on Joint Textual and Visual Clues

39310

Knowledge Triples