Re:Yann LeCun's comment on AlphaGo and true AI
by astazangasta on Monday March 14, 2016 @12:54PM
(#11283892)
That IS unsupervised learning - you are maximizing an optimization function, you are not training using a set of "truth" values.
[ Parent
| on HN ]
Re:Yann LeCun's comment on AlphaGo and true AI
by fallingfrog on Monday March 14, 2016 @01:08PM
(#11283965)
Hmm, I don't think so. From wikipedia: "Unsupervised learning is the machine learning task of inferring a function to describe hidden structure from unlabeled data. Since the examples given to the learner are unlabeled, there is no error or reward signal to evaluate a potential solution." (emphasis mine)
Re:Yann LeCun's comment on AlphaGo and true AI
by astazangasta on Monday March 14, 2016 @01:15PM
(#11284019)
Yes, "unlabeled" means that the examples are not marked as "true" or "false", and the "inferring" you are doing is making your function fit the data (e.g. minimizing some error term). Life doesn't come with pre-packaged answers; humans use unsupervised learning.
Re:Yann LeCun's comment on AlphaGo and true AI
by fallingfrog on Monday March 14, 2016 @01:39PM
(#11284185)
Right, but we do have a reward function that says pleasurable/painful and novel/boring and probably other stuff too. So that can be viewed as a labeling on the data. Earlier data can be associated with the reward labeling through induction; that's how a recurrent neural net works. Doubtless that's an oversimplification though.