To truly appreciate a project like SU, you need to understand what makes it tick. Yes, we’ve made SU intelligently artificial, but what does that mean?
We’re training the underlying code of SU to understand human language and context, and respond accordingly. It’s not an easy task and many mistakes are made along the way.
The point of SU is not to interpret sentences word-for-word but to extract the intent behind the message. Very soon, SU will be able to keep context and have the ability to remember and combine intents across several sentences in a conversation.
SU is still a baby and unable to process and execute complex, clinical requests. But it is smart enough to try and find the right answer from it’s growing brain of advice and guidance. We’re also using ‘Machine Learning’ to process the unknown requests and gradually map them back to answers the system might already have, or prompt US to create new ones.
Whenever SU is unsure of what to say, you get a message apologising and that’s when the unknown is passed into the system. It’s those unknowns that will eventually allow SU to know more and more. So that next time SU encounters a similar message, it will give the correct response. As you can see, machine learning is a matter of trial and error.