Skip to content
New issue

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

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

Already on GitHub? # to your account

[20210704] Weekly AI ArXiv 만담 #15

Closed
jungwoo-ha opened this issue Jul 3, 2021 · 5 comments
Closed

[20210704] Weekly AI ArXiv 만담 #15

jungwoo-ha opened this issue Jul 3, 2021 · 5 comments

Comments

@jungwoo-ha
Copy link
Owner

jungwoo-ha commented Jul 3, 2021

@ghlee3401
Copy link

ghlee3401 commented Jul 3, 2021

@Kyung-Min
Copy link

@hollobit
Copy link

hollobit commented Jul 4, 2021

WHO, 보건분야 AI의 6가지 원칙 담은 보고서 발간

제12차 혁신성장 BIG3 추진회의 개최 - ②-3 소프트웨어 의료기기 육성을 위한 제도개선 방안

딥마인드 “강화학습이 범용AI 달성하는 시대 곧 온다”

How on-device machine learning has changed the way we use our phones

Shenzhen, China’s Silicon Valley, plans to turbocharge local AI development with first-of-a-kind regulations

Study finds that few major AI research papers consider negative impacts

@jiyounglee-0523
Copy link

Unsupervised Vision-and-Language Pre-training Without Parallel Images and Captions

image

  • unparallel한 텍스트와 이미지에 대한 pretraining 기법
  • RoI 기법을 사용하여 이미지에서 text 추출

@jwlee-ml
Copy link

jwlee-ml commented Jul 4, 2021

FAIR 에서 나온 논문입니다
Early Convolutions Help Transformers See Better

image

6월 초에는 google에서 이런 논문도 있었습니다
CoAtNet: Marrying Convolution and Attention for All Data Sizes

image

결론은 앞에는 conv 뒤에는 transformer

# for free to join this conversation on GitHub. Already have an account? # to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

6 participants