LMAO. and a good example how statistics ≠ understanding.pic.twitter.com/BhKnSaYKkG
Show this thread
-
Gary, try https://mosaickg.apps.allenai.org by typing "Gary stacks kindling and logs and drops some matches". Sorry I used deep learning... :)pic.twitter.com/W2c0PaEzZM
not everyone who starts a fire is mean, rude or reckless. generally those that are mean use gasoline etc, not kindling and logs, and so on. but you are certainly welcome to use your system to try to generate answers to questions like this and ones I posed earlier tonightpic.twitter.com/EaBbohhMS3
-
-
About your comment "not everyone is so and so..." that's exactly the nature of *commonsense models*: stochastic expectations on what are _likely_ to be true, not _necessarily_ true. Analogous to how *language models* are not about which word _must_ follow vs _could_ follow.
-
you could compare with actual human distributions of data. in this particular case the data seem unconvincing; the system doesn’t recognize that a campfire is different from arson. in other cases , comet does better
- 3 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.