by Craig Roth | November 17, 2016 | Comments Off on Humans Training Bots
In my previous posting “Humans Masquerading as Smart Machines to Avoid the Stupid Valley” I commented on an article about the process behind the scenes at a smart personal assistant chatbot company. I asked “Do I want a bot taught by “helicopter trainers?”.
After that blog post I was contacted by the x.ai folks, who gave me a demo and some insight into what the humans are doing over there. They validate that the translation from the free-form text people send to the form fields needed to schedule a meeting are correct. Us humans are rather sloppy and could use some help being understood. Once that “annotation” is correct, the AI takes it from there.
In the words of x.ai:
The role of the AI trainer (as we’ve written about here and here) is to verify the machine’s annotations of scheduling-related data that the machine has extracted from scheduling conversations. We verify the vast majority of ANNOTATIONS because that is how you train a machine to schedule meetings nearly flawlessly. Same concept as Google Self-Driving Car.
There are two issues at play here. First is natural language recognition, and the second is replicating what a human assistant would do once the instructions are understood. It’s instructive to see that the first issue is the one we haven’t fully conquered yet.
Comments or opinions expressed on this blog are those of the individual contributors only, and do not necessarily represent the views of Gartner, Inc. or its management. Readers may copy and redistribute blog postings on other blogs, or otherwise for private, non-commercial or journalistic purposes, with attribution to Gartner. This content may not be used for any other purposes in any other formats or media. The content on this blog is provided on an "as-is" basis. Gartner shall not be liable for any damages whatsoever arising out of the content or use of this blog.