MUMC: Minimizing uncertainty of mixture of cues

作者：

Highlights：

• Propose a method for obtaining posterior distributions by considering the advice of experts.

• Propose a MoE method to combine contributions of multiple cues embedding.

• Propose a moderator to capture cue embeddings & minimise uncertainty for the generation task.

摘要

Highlights•Propose a method for obtaining posterior distributions by considering the advice of experts.•Propose a MoE method to combine contributions of multiple cues embedding.•Propose a moderator to capture cue embeddings & minimise uncertainty for the generation task.

论文关键词：Uncertainty estimation,Mixture of cues,Visual Question Answering,Paraphrase,Visual Question Generation,LSTM,CNN,Encoder-decoder

论文评审过程：Received 6 July 2020, Revised 17 July 2021, Accepted 10 August 2021, Available online 24 August 2021, Version of Record 8 September 2021.

论文官网地址：https://doi.org/10.1016/j.imavis.2021.104280