机器学习 API 介绍及实况演示

Allen Day

您是否认为自己的企业可以利用 Google 的机器学习专长来驱动和改进您的业务应用,却在构建和训练自定义模型时遇到困难?Google Cloud Platform (GCP) 提供五个 API:Google Cloud Vision API、Cloud Speech API、Cloud Natural Language API、Cloud Translation API 和 Cloud Video Intelligence API。这些 API 通过单一 API 调用访问经过预先训练的机器学习。在此会议中,我们先概要介绍了每个 API。然后,我们通过实况演示深入讲解代码内容。

机器学习是从范例和经验中学习

How to identify photo of everything? arbitary collection of pixels?

2 Ways to Benefit from ML

  • Train using your own data: Tensorflow + Cloud ML Engine
  • API

cloud.google.com/vision

Vision API

Case study: Disney

To have audience be more engaged in the making of movies.

Key Features

  • Label detection
  • Face recognition
  • Landmark detection
  • ...

Speech API

Case study: Azar

Azar: Speech + Translation + Matching

Key Feature: Timestamp

Key feature: timestamp to make content indexable

Translation API

Case study: airbnb

Key Feature: Python SDK

https://bit.ly/nyt-ai-awakening

NLP API

Case study: Wootric

Feedback collection

Key Features

  • Entity extraction
  • Sentiment Analysis

Cool Application

Integration with Google Docs, using spreadsheet along with Pivot table

Video Intelligence API

Case study: CANTEMO

From single frame to higher level features.

Auto sectioning

Structuring

Key: App Engine