What is AI?

What is Machine Learning?

  • Supervise Learning… Input (A) to output (B) learning

What is Data?

  • Need to determine what is A (Input – Features) and what is B (Output) on the inputs
  • Acquiring data via manual labeling, observation, available downloads
  • Once you start collecting data, start feeding it to a AI team for a feedback loop to start. Start this early. 
  • You can not assume value in data, because you have a lot of it. Becareful, have an AI with to asses
  • Garbage in garbage out exists here too

The Terminilogy of AI

  • AI(ML, et al (DL/NN, et al))…Data Science cuts across all of these.

What makes an AI company?

What machine learning can and cannot do

  • Technical diligence: looking at the data, look at the input, and output A and B, and just thinking through if this is something AI can really do.
    • Can NOT do market research and write an extended market report
  • Complex interactions require lots of examples. the system would either remain vague in response, or return incoherent responses.
  • What makes an ML problem easier
    1. Learning s “simple” concept <= 1 second of human thought…Cause we have to formulate? why?
    2. Lots of data available
  • AI has a hard time of inferring intention of action in context A= hand gesture, B= stop, hitchhiker, left turn on a bike, hello, etc…

Deep Learning

  • Deep Learning and Neural Network are used interchangeably