Reading the article below my first reaction was "so it isn't just me",
as, many years ago, when a buddy at Motorola Research was describing his
study of neural networks relating to making smart car technology, the
"inside" of the network was a black box and in the decades since I
haven't found any explanation of how they work that doesn't boil down to
another figurative black box as far as the mappings of inputs to outputs
for a particular pattern recognition use case after training sessions.
So I was relieved to find the inner workings of machine learning (ML)
systems can be just as impenetrable to others while also being
distressed to read about a cat being declared guacamole after (human-)
indiscernible changes to image files as described in the article.
The two major topics that are explored, intellectual and technical debt,
seem like the kind TriEmbed folks may find interesting both generally
and in regards to ML.
https://medium.com/@zittrain/from-technical-debt-to-intellectual-debt-in-ai-e05ac56a502c
-Pete
PS Yesterday Dan Stipe and I were marveling at a technical book offer by
Amazon for the low low price of $4800 plus $3.99 shipping. Now I
understand where this may have come from!!
_______________________________________________
Triangle, NC Embedded Computing mailing list
To post message: [email protected]
List info: http://mail.triembed.org/mailman/listinfo/triembed_triembed.org
TriEmbed web site: http://TriEmbed.org
To unsubscribe, click link and send a blank message:
mailto:[email protected]?subject=unsubscribe
