Self-driving cars a great example, because in this case there are two competing approaches -- the symbolic one, mostly consisting of handcrafted software encoding human abstractions, and the deep learning one, learned end-to-end. One will get to L4--even L5, the other never will.https://twitter.com/GaryMarcus/status/1003314562261712896 …
-
Show this thread
-
Replying to @fchollet
@GaryMarcus Tho hesitating to comment w/ two eminences, isn't pragmatic answer likely to lie in combination of (i) symbolic &#ML approaches as part of single SDV system & (ii) constraining domain - in case of roads & driving, limiting to certain routes, pre-mapped, etc.?2 replies 0 retweets 1 like
Replying to @kennethanderson @fchollet
absolutely we should seek a merger (that was point of my reply, and the thrust of what i have advocated throughout my career). but rene your second point, in driving over-constraint can lead to outlier cases in a complex world. (eg if maps are wrong because of a detour)
0 replies
1 retweet
3 likes
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.