We call it AI because no one would take us seriously if we called it matrix multiplication seeded with a bunch of initial values we pulled out of our asses and run on as much shitty data as we can get our grubby little paws on.
@aral
"To paraphrase provocatively, 'machine learning is statistics minus any checking of models and assumptions'."
-- Brian D. Ripley (about the difference between machine learning and statistics)
useR! 2004, Vienna (May 2004)
@aral in germany we call it KI (künstliche Intelligenz, artificial intelligence). Our politicians tell us, that Germany will soon be the leading nation for KI. They ignore (or they didn't know) that the rest of the world doesn't know what German politicians mean by KI, since everyone uses the term AI.
@aral This reminds me of some work I did in the 90's on chemistry analysis using multiple linear regression in complex spectroscopy.There were a few adages that may still hold:
Shit in, shit out,
Correlation and causation are not the same and may not be predictive,
Your dataset needs to adequately describe the variance of the population. Future prediction relies on this.
I haven't seen the studies covering this. Without addressing these, the techniques were often over confident their prediction
@leroc Odd, it was correct enough for me to write a machine learning implementation myself while playing around with the concepts. Does the one you wrote work differently?
@aral As a matter of fact, it does. I programmed neural networks and Markov chains in the nineties. Computers have become faster with more memory since then, but the underlying theory hasn't changed much.
The danger of calling it "just matrix multiplication" is that this downplays the real danger that AI poses to our society, especially when combined with capitalism.
@aral not withstanding the human 'sweatshop labor' required to filter and label all the scaped garbage and the psychological toll on these unfortunate workers , mostly in Africa.
Add comment