How I Think About Machine Learning

I'm obviously working on the Pro Version of my AutoML software, Black Tree AutoML, and most of the work is simply restating what I've done in a manner that interacts with the GUI I've developed (pictured below). This in turn causes me to reconsider work I've already done, and so I thought it worthwhile to present again my basic views on Machine Learning, nearly six years after I started working on A.I. and Physics full-time, as they've plainly evolved quite a bit, and have been distilled into what is, as far as I know, humanity's fastest Deep Learning software. In a series of Lemmas (see, Analyzing Dataset Consistency [1]), I proved that under certain reasonable assumptions, you can classify and cluster datasets with literally perfect accuracy (see Lemma 1.1). Of course, real world datasets don't perfectly conform to the assumptions, but my work nonetheless shows, that worst-case polynomial runtime algorithms can produce astonishingly high accuracies: This informs my work generally, which seeks to make maximum use of data compression, and parallel computing, taking worst-case polynomial runtime algorithms, producing, at times, best-case constant runtime algorithms, that also, at times, run on a small fraction of the input data. Even when running on consumer devices, Black Tree's runtimes are simply incomparable to typical Deep Learning techniques, such as Neural Networks, and the charts below show the runtimes (in seconds) of Black Tree's fully vectorized "Delta Clustering" algorithm, running on a MacBook Air 1.3 GHz Intel Core i5, as a function of the number of rows, given datasets with 10 columns (left) and 15 columns (right), respectively.