A definition of "machine learning", Drumroll please.
A computer program is said to learn from experience with respect to a class of tasks and a performance measure
P, if its performance at tasks as measured by
P improves with experience beyond a baseline accuracy derived by: "Simply predicting the most frequently occurring outcome.
3Blue1Brown shows how trained neural networks simulate human abilities: https://www.youtube.com/watch?v=aircAruvnKk
CGP Grey shows how the genetic algorithm can simulate human abilities: https://www.youtube.com/watch?v=R9OHn5ZF4Uo
Machine learning is just computer programming. The theory behind machine learning has been well-known since the 1960's, the only difference is $2500 worth of Disk/Mem/CPU/GPU has enough horsepower over the last 15 years to begin outperforming your
1*10^18 signal-processing neurons, and roughly 90 yottabyte recall-memory.
To do machine learning you have to excel in these classes: All Mathematics and computer science themed classes,
Discrete and continuous mathematics,
Calculus I through III. The one concept in Calculus you absolutely must understand both backwards and forward is the derivative and integral of multiple variables, because that 'finding slope of a point of a tangent on a curve' is the operating principle behind machine learning: knowing which way to tease a
y=mx+b to reduce error. https://www.youtube.com/watch?v=EkZGBdY0vlg
Tesla could have Level 5 autonomy now on some roads, if they simply geofenced their software to the few roads that the model was specifically trained to traverse. https://www.youtube.com/watch?v=tlThdr3O5Qo
If you think something is correct, prove it with a test. If you think something is fast, time it with a test. If you think one algorithm is better than another, compare them with a timed test and let the
> sign have the last word.
Member for 6 years, 11 months
203 profile views
Last seen Dec 6 '17 at 13:03