TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Our weird behavior during the pandemic is messing with AI models

45 pointsby giladabout 5 years ago

10 comments

sn41about 5 years ago
The title is backwards. It should be that AI models are wrong. Not that a given AI model is a gold standard that human behavior is failing to live up to.
评论 #23200889 未加载
评论 #23200295 未加载
BIackSwanabout 5 years ago
We disabled all ML enabled predictive features when the pandemic began.<p>This has prevented us from the pain of having “bad” unreliable predictions which are worse than having “no” predictions. This saved us a lot of pain in maintainence and debugging.
评论 #23202008 未加载
asquabventuredabout 5 years ago
So it turns out that humans are more complex than what the black box AI models have hyper optimized for.
评论 #23200762 未加载
评论 #23219131 未加载
IIAOPSWabout 5 years ago
Most models (&quot;AI&quot; or otherwise) behave very poorly in the situation known a &quot;structural break&quot; [1]. In extreme instances you can even end up with something like Simpsons paradox where the predicted correlation is the opposite of the true correlation [2]. I&#x27;ve had to deal with this exact problem before. The trick is to have a meta model that sits on top and tries to detect when the recent data is highly unlikely to have been produced if the existing model is true. When that happens you flush the older data. This might be what our brains do [3].<p>[1] <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Structural_break" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Structural_break</a> [2] <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Simpson%27s_paradox" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Simpson%27s_paradox</a> [3] <a href="https:&#x2F;&#x2F;slatestarcodex.com&#x2F;2017&#x2F;09&#x2F;06&#x2F;predictive-processing-and-perceptual-control&#x2F;" rel="nofollow">https:&#x2F;&#x2F;slatestarcodex.com&#x2F;2017&#x2F;09&#x2F;06&#x2F;predictive-processing-...</a>
评论 #23203183 未加载
评论 #23200692 未加载
sircastorabout 5 years ago
As others have noted, the models don&#x27;t represent the way things are anymore. And bring frank,I don&#x27;t they&#x27;ll represent things ever again. I very much doubt that the models will be a good match for behaviors going forward as people are likely going to have a hard, long transition towards normalcy
评论 #23200714 未加载
im3w1labout 5 years ago
It seems a fundamental nature of the universe that there is a tension between continuum and discrete. You see it everywhere:<p>Wave vs particle, gradual heating vs phase change, but also soup with little chunks of meat (themselves smaller continuum), gas clouds coalescing into planets. It goes on and on.<p>In this case model was built for the continuum and in came a big discrete shock. Something it wasn&#x27;t built for. Shocks require a different modality of thought. More reasoning, logic-y than regression.
SomeoneFromCAabout 5 years ago
Our neighborhood cats were seriously confused and disturbed by the looks of our faces in masks, no wonder AI is confused too.
lostmsuabout 5 years ago
To be fair, the predictive ability of humans also reduced quite a lot.
2019-nCoVabout 5 years ago
Is anti-fragile AI possible?
评论 #23200878 未加载
评论 #23207240 未加载
评论 #23201260 未加载
sbmthakurabout 5 years ago
It&#x27;s high those AI models catch up with humanity.