일 | 월 | 화 | 수 | 목 | 금 | 토 |
---|---|---|---|---|---|---|
1 | ||||||
2 | 3 | 4 | 5 | 6 | 7 | 8 |
9 | 10 | 11 | 12 | 13 | 14 | 15 |
16 | 17 | 18 | 19 | 20 | 21 | 22 |
23 | 24 | 25 | 26 | 27 | 28 |
- adminbro
- 자료구조
- mistel키보드
- 42seoul
- 어셈블리어
- uuid-ossp
- 프라이빗클라우드
- 이노베이션아카데미
- raycasting
- 어셈블리
- c++
- 쿠버네티스
- schema first
- 정렬
- psql extension
- 스타트업
- 레이캐스팅
- GraphQL
- 42서울
- 동료학습
- 부동소수점
- 파이썬
- enable_if
- 스플릿키보드
- 엣지컴퓨팅
- 도커
- 창업
- SFINAE
- Cloud Spanner
- 텍스트북
- Today
- Total
written by yechoi
GCP - 머신러닝 관련 서비스 본문
Google Cloud Product Fundamentals
Google Cloud Product Fundamentals
Google 클라우드에서 제공합니다. This course builds on the Business Transformation with Google Cloud course by taking you on a journey into the ... Enroll for free.
www.coursera.org
Module 3: Transforming business with Artificial Intelligence and Machine Learning
1: Introduction
learn about
-
what is ML and why is data integral for its successful uses
-
Real-world use cases for ML
-
Google Cloud's differentiators in AI and ML
2: Machine learning
What is machine learning
"standard algorithms to get predictive insights from data to make repeated decisions"
-
Standard algorithms
-
ML 소프트웨어를 예측하려는 라벨로 분류한 데이터로 훈련시켜 만든 모델
-
ML data requirements
-
Data
-
질을 결정하는 세가지 요소
-
coverage: 가능한 모든 인풋과 아웃풋을 다룸
-
clean: 잘못 label 되지 않도록 오류 없어야
-
complete: 데이터는 모델이 세상을 보는 터널, 관련 데이터가 없으면 그 부분을 볼 수 없음
-
Predicive insights and repeated decisions
-
Predictive Insights
-
알려지지 않은 대상(미래 등)에 대한 예측
-
-
Repeated decisions
-
반복적인(<->일회성인) 판단에 적합
-
3: Real-world use cases for ML
접근 가능한 데이터가 많아지고, ML 알고리즘이 발전하고, 컴퓨팅 능력이 높아지면서 ML은 진입장벽이 낮아짐
-
ebay의 사례
-
Ocado의 사례
-
Ambra Health의 사례
4: AI and ML with GCP
-
modern ML services: 미리 훈련된 모델을 사용하거나 자신만의 모델을 만들 수 있음
-
TPU를 장착한 Compute Engine Virtual Machines
-
Google Cloud's AI hub: plug-and-play AI components
-
Vision API를 장착해 이미지 분류를 도입할 수 있음
-
AutoML Vision API는 이보다 더 정교한 이미지 분류를 가능케함
-