I’m watching the Jeopardy episode with IBM Watson.
I’m not sure this is fair - the human contestants need to convert the sound of Alex Trebek’s voice, or the image of the words on the screen into semantic units (words/phrases/sentences) before they can parse the meaning. Watson gets its questions given as a text file, so it can begin parsing right away. To be fair, Watson should have to rely on an image of the question (which shouldn’t be too hard for it to parse), or the sound of Alex’s voice (which it could probably not parse accurately). That being said, it’s still pretty impressive that it can do what it does.
