*nobody* knows the #AI and #ML answers, but here is my proposal for research direction, per @kchonyc’s query, and per The Algebraic Mind; more consistent w @pmdomingos and @egrefen and Josh Tenenbaum etc than pure deep learning. (also perhaps in line with @mpshanahan?)https://twitter.com/garymarcus/status/1200636202744868864 …
-
I don't buy the common sense db idea. I think that common sense is a living thing, generated on the fly, not knowledge stored in a data base. (Then again, even Minsky told me it's a good idea, and who am I to disagree with Minsky. I also admit that I have little common sense.)
it’s got to be some of both. we can’t generate it all on the fly, and we have to generate some of it on the fly kudos to you by the way in taking the high road here & actually engaging in the substance of the proposed changed in direction; that’s exactly what the field needs.
-
-
Wait what would be common sense that you can generate on the fly and the common sense that you can't? I agree with Tenenbaum and Minsky in that its all stored db and new common sense is generated from the existing base
-
Common sense is incredibly rich and contains many highly improbable inferences. While one in every handful of concepts tends to be grounded in some kind of experience, the relational tapestry seems to grow like a fetid forest from multitudes of real-time inferential explorations.
- 1 more reply
New conversation -
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.