VQA as a factoid question answering problem: A novel approach for knowledge-aware and explainable visual question answering
作者：
Highlights：
• The proposed model is a free form, open ended and knowledge aware VQA model.
• VQA modeled as an explainable, end to end factoid question answering problem.
• Model capable of leveraging granular details, correlate inter-related details in scenes.
• Model capable of leveraging external world knowledge to answer questions.
• Model capable of predicting likely explanations to justify the predicted answers.
摘要
Highlights•The proposed model is a free form, open ended and knowledge aware VQA model.•VQA modeled as an explainable, end to end factoid question answering problem.•Model capable of leveraging granular details, correlate inter-related details in scenes.•Model capable of leveraging external world knowledge to answer questions.•Model capable of predicting likely explanations to justify the predicted answers.
论文关键词：Visual question answering,Factoid question answering,Knowledge based reasoning,Explainable VQA
论文评审过程：Received 2 September 2020, Revised 6 October 2021, Accepted 7 October 2021, Available online 24 October 2021, Version of Record 1 November 2021.
论文官网地址：https://doi.org/10.1016/j.imavis.2021.104328
