Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

NUBIA, EvalNLGEval'20 #367

Open
AkihikoWatanabe opened this issue Jun 2, 2021 · 2 comments
Open

NUBIA, EvalNLGEval'20 #367

AkihikoWatanabe opened this issue Jun 2, 2021 · 2 comments

Comments

@AkihikoWatanabe
Copy link
Owner

NeUral Based Interchangeability Assessor. A new SoTA evaluation metric for text generation.
https://wl-research.github.io/blog/

@AkihikoWatanabe
Copy link
Owner Author

TextGenerationに関するSoTAの性能指標。BLEU, ROUGE等と比較して、人間との相関が高い。
image

image
pretrainedされたlanguage model(GPT-2=sentence legibility, RoBERTa_MNLI=logical inference, RoBERTa_STS=semantic similarity)を使い、Fully Connected Layerを利用してquality スコアを算出する。算出したスコアは最終的にcalibrationで0~1の値域に収まるように補正される。

@AkihikoWatanabe
Copy link
Owner Author

意味的に同等の内容を述べた文間でのexample
image
BLEU, ROUGE, BERTのスコアは低いが、NUBIAでは非常に高いスコアを出せている。

@AkihikoWatanabe AkihikoWatanabe changed the title NUBIA NUBIA, EvalNLGEval'20 Aug 14, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

1 participant