https://bdtechtalks.com/2021/03/15/machine-learning-causality/?utm_source=rss&utm_medium=rss&utm_campaign=machine-learning-causality Technology solving problems... and creating new ones Tue, 15 Mar 2022 01:29:56 +0000 hourly 1 https://bdtechtalks.com/2021/03/15/machine-learning-causality/comment-page-1/#comment-29722 Tue, 15 Mar 2022 01:29:56 +0000 https://bdtechtalks.com/?p=9814#comment-29722 This is a very interesting read. What do you think about Bayesian Neural Networks (BNNs)? To me, they appear to be the closest merge between the two fields.

]]> https://bdtechtalks.com/2021/03/15/machine-learning-causality/comment-page-1/#comment-17496 Fri, 09 Apr 2021 21:14:38 +0000 https://bdtechtalks.com/?p=9814#comment-17496 Perhaps, the answer lies in complexity science -https://necsi.edu/fixing-science-using-a-new-science-of-science

]]> https://bdtechtalks.com/2021/03/15/machine-learning-causality/comment-page-1/#comment-17119 Mon, 22 Mar 2021 12:23:44 +0000 https://bdtechtalks.com/?p=9814#comment-17119 What about reasoning backwards from the assumption that every event is caused? Then use the data to rule out discrete theories. Instead of looking for correlations you would look for interruptions in statistical patterns.

]]> https://bdtechtalks.com/2021/03/15/machine-learning-causality/comment-page-1/#comment-16842 Tue, 16 Mar 2021 06:55:05 +0000 https://bdtechtalks.com/?p=9814#comment-16842 Idea – Since the hidden fully connected layer has output that looks random. The last layer has not much error correct capability as if had been a photo or an image. This creates a classification problem. Create a noise free next to last image that can be error corrected?

]]>