Forensics at (Cosmic) Scale

Intro.
The advancements in archeological forensics has made some really impressive progress lately. Here are some notes based on the large amount of talks, NOVA episodes and various documentaries I've watched this year. But first two points:

1.
It is amazing what falls out from just asking the question: can I find x from y?

One of my favorite examples of this is "did Neandertals leave trash around there caves? Why don't I look at _anything else_ but the wall paintings for evidence of visitation or inhabitance. Oh wow there's all kinds of debris that carbon dates to the same period.

My second favorite example is "do chimpanzees use sign language in the wild? Why don't I observe all these subtle gestures and see if any correlations pop up. Oh wow they use sign language. Oh wow they are having conversations. It really should be no surprise to me that chimpanzees can learn sign language in captivity. I mean if it wasn't used in the wild evolution would have optimized out the capability.

2.
It is equally amazing what falls out from applying machine learning (ie classification) to large data sets. To be honest you don't even need compute power or machine learning. Until recently the very idea of collecting and combining data from a variety of sources and then parsing/counting/whatevering the data wasn't performed as systematically as it is now.

A great example of this was the work done on the symbols in ancient human cave settlements. The symbols and their count were catalogued from a variety of sources and combined together revealing that there was a pretty significant amount of overlap. Very exciting for what that might imply! Yet for a hundred years no such effort had been undertaken.

However, I do freely admit that the amount of data available in the past prevented many of these efforts.

The Table.