It is definitely a mechanical advantage that is deciding this, and personally I don’t find that very impressive. Watson seems to have problems with the complex answers – the ones that are not worded in a basic manner. When it can pick out keywords it can find the answer and wins the button pushing race.
I’m not that impressed.
It is definitely an advance in question evaluation and answer retrieval, and it can handle certain levels of nuance in the question. I assume that there is a level of nuance in the answer selection that is also present that I can’t see because that’s all going on in the servers. That is impressive and intriguing. But the machine does have its problems.
I recorded this and watched it tonight. On to round 3.
No comments:
Post a Comment