AIForEveryone_AI project

What you can do with data

Sales data:

• Data science can help to find the way to optimize sales funnel
• Machine learning automate prioritizing customers based on possibility

Manufacturing line data:

• DS can optimize manufacturing line
• ML can automate visual inspection

Marketing data:

• DS can get insights from A/B testing
• ML can do customized product recommandation

Agriculture:

• DS can do crop anlytics
• ML do precision weed killing

How to choose AI project

Brainstorming framework

• think about automating tasks rather than automating jobs. ex) call center routing
• What are the main drivers of business value?
• What are the main pain points in our business?
• You can make progress even without big data(even though more data is good)

Choosing AI project subject:

• Technical diligence
  • can AI system meet desired performance?
  • How much data is required?
  • Engineering timeline(time, resource)?
• Business diligence
  • lower costs?
  • increase revenue?
  • help launching new product or business?

Implementing by build or buy?:

• ML -> in-house | outsource
• DS -> commonly in-house
• don't try to make wheel. Don't build industry standard, just use it

Hard to get 100% arcuracy due to:

• limitations of ML
• insufficient data
• mislabeled data
• ambiguous labels

AI technical tools

• opensource ML frameworks:
  • TensorFlow
  • PyTorch
  • Keras
  • MXNet
  • CNTK
  • Caffe
  • PaddlePaddle
  • Scikit-learn
  • R
  • Weka
• Research publications: Arxiv
• repository : GitHub