search

cesc-park / attend2u

🖼️ Attend to You: Personalized Image Captioning with Context Sequence Memory Networks. In CVPR, 2017. Expanded : Towards Personalized Image Captioning via Multimodal Memory Networks. In IEEE TPAMI, 2018.
MIT License
209 stars 57 forks source link

about superscript a and c #6

Closed bidongqinxian closed 6 years ago

bidongqinxian commented 6 years ago

In your paper，I‘ve read the CSMN model.I have trouble with the understanding of superscript a and c，would you mind telling me what the porpose of add superscript a and c to the image memory vector and the user context memory vector?Thanks a lot!

cesc-park commented 6 years ago

References to the "End-To-End Memory Networks, Sukhbaatar et al" paper will help you understand. As can be seen in the paper, the difference between a and c is a subscript that identifies each matrix that projects into input memory and output memory.