Celebrations and Commemorations In An Algorithmic World
In September 2047 it will be the 50th anniversary of the death of Princess Diana and the 25th anniversary of Queen Elizabeth II and the 46th anniversary of 9/11, as well as commemorations of events that will occur as 10th and 20th anniversaries.
I had thought that YouTube's algorithm results would have featured the annual 9/11 commemoration, but it didn't. This is why organic memory is so important and is one of the good things about moderated and curated media: It reports on things that are more universal and traditional instead of relying on what algorithms perceived were important to me. We shouldn't have to go hunting go hunting for our own news. We know what the concept of a Zebra is.
From the recent excellent book, How to Stay Smart in a Smart World: Why Human Intelligence Still Beats Algorithms
"A deep neural network has no concept of a zebra but searches for typical patterns and images, such a stripes. As a consequence of a picture of a zebra with two additional legs (and thus more stripes), would startle a human but may make a deep neural network for confident that the zebra is indeed a zebra. Why not simply teach a network that zebras cannot have more than four legs? Fixing this specific error will not eliminate the underlying problem. The network doesn't know that a picture represents something in the real world, it has no concept of things. Its intelligence is confined to detecting patterns in colors, texture, and other features. Human intelligence in contrast is more about representing the world." (60)
Comments