This totally self-serving piece is so egregiously disconnected from reality that it doesn't deserve a response. Past exchanges have shown that writing a response would be a waste of time. We have better things to do with our time. Like actually doing research on AI.
“The deepest problem with deep learning” & why it’s not good to confuse a snowplow & a school bus. Some reflections too long for Twitter on recent debate with @ylecun and others, w. additional discussion of @anh_ng8 lab’s important recent results.https://link.medium.com/dHSKKYBsiS
-
-
-
True to form
- 1 more reply
New conversation -
-
-
Gary, I am surprised by the amount of energy you spend in this, which increasingly looks more like a personal quest to me. Why concentrate so much on negativity?
-
i write fast. it’s a spare time project, but important to me, and i am dismayed by what i see, particularly the politics.
End of conversation
New conversation -
-
-
Lots of nice progress on incorporating symbolic reasoning recently, along these lines, e.g., http://papers.nips.cc/paper/7381-neural-symbolic-vqa-disentangling-reasoning-from-vision-and-language-understanding … http://papers.nips.cc/paper/8203-learning-to-reason-with-third-order-tensor-products …
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I would be interested in your opinion of our work on simulation-based internal models: http://alanwinfield.blogspot.com/2018/09/experiments-in-artificial-theory-of.html … The approach is (1) entirely algorithmic - although we're now adding a learning component using a Bayesian network... 1/
-
(2) embeds causality since the robots' simulation (of itself and others) allows it to model and predict the likely consequences if its (and their) actions, and 2/
- 1 more reply
New conversation -
-
-
If you want to recognize objects in strange positions, you have to include examples of objects in strange positions in your training data, not just ImageNet.
-
This Tweet is unavailable.
New conversation -
-
-
This has been considered many times before - Neural-Symbolic Cognitive Reasoning - https://www.springer.com/us/book/9783540732457 … but it may be worth bringing up again.
-
Also by Minsky - Symbolic vs. Connectionist http://web.media.mit.edu/~minsky/papers/SymbolicVs.Connectionist.html …pic.twitter.com/DOO55KrAiX
- 2 more replies
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.