Papers with AREL

1 papers
No Metrics Are Perfect: Adversarial Reward Learning for Visual Storytelling (P18-1)

Copied to clipboard

Challenge: Visual captioning is aimed at depicting the concrete content of images, but its capability of performing human-like understanding is still restrictive.
Approach: They propose an Adversarial REward Learning framework to learn an implicit reward function from human demonstrations and optimize policy search with the learned reward function.
Outcome: The proposed framework improves performance over state-of-the-art (SOTA) methods in cloning expert behaviors, but human evaluation shows that it achieves significant improvement in generating more human-like stories than SOTA systems.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations