1. 사례로 파악하자
## 유투브 / 논문
1. data sources
- video corpus
- video feature
- user history
- context
+ other candidate sources
2. model
- candidate generation
- ranking
https://static.googleusercontent.com/media/research.google.com/ko//pubs/archive/45530.pdf
## 틱톡 / 리버스 엔지니어링
1. data source
2. biz logic
- Screening
- Traffic allocation
3. model
- Candidate?
Why TikTok made its user so obsessive? The AI Algorithm that got you hooked.
Tick Tok is taking the world by storm. The short video app was downloaded 2 billion times, so what is the magic behind it ?
towardsdatascience.com
2. 비디오 임베팅 벡터는 어떻게 만들어?
- 비디오는 Sequencial image frame인데 어떤 식으로 벡터를 만들까? RCNN?
- 이게 도움이 될지 모르겠어 http://vision.stanford.edu/pdf/RamanathanICCV2015.pdf
'DataMining > Recommendation system' 카테고리의 다른 글
| 추천시스템 - Metrics and DEVIEW 2017,2020 (0) | 2022.05.11 |
|---|