3

I'm testing couple of IBM Watson APIs like the following: enter image description here

Does Watson get smarter and learn more about my data the more I use it?

I read that Watson is getting smarter with more data it learns and processes. I'm not sure if this is only done behind the scenes by IBM Watson team, or if these API's as well allow an instance of Watson for example to be smarter with my specific application I'm developing.

Community
  • 1
  • 1
Marko
  • 570
  • 5
  • 21
  • 2
    This is a question for IBM, not SO. What does the EULA say? – jonrsharpe Jan 23 '16 at 16:11
  • 3
    I can't speak for Bluemix officially but the answer is no. Watson engagements we normally use only the customers data and public domain data (which is legally allowed to be used). One customer instance has no knowledge of another customers instance when it comes to how it understands that domain data. – Simon O'Doherty Jan 23 '16 at 16:37

2 Answers2

2

If you mean that Watson is using the data you input into your instances, then no. Watson is IBM's, but your data is always yours.

By default, instances are isolated.

By smarter, they mean they have their very own instances of APIs which they train. Also they, improve algorithms behind the scenes.

Walter Montes
  • 101
  • 5
1

It depends on your definition of learning. Is it offline learning or online learning? Do you refer to Watson learning from your corpus on the entire domain and use it later on, or just on your data.

It also depends which services you use, check out Retrieve and Rank or Natural Language Classifier for examples of services that learn from your data

Dudi
  • 2,340
  • 1
  • 24
  • 39