Though I am highly interested in Machine Learning, I am very hesitant to make the dive and jump into it. Having been trained as a physicist and currently a PhD student in synthetic and systems biology, I am very drawn to natural systems and the underlying universal truths contained within them. Every time I begin to skim the surface of ML, I quickly get lost in between "the amazing new algorithm X" and "the statistically relevant approximation method Y" being applied to "extremely specific problem Z".
Can an experienced person in this field try to appeal to my theoretical physicist instincts and explain how ML is important and relevant on a universally fundamental scale? I think I have a project which could benefit from a ML treatment, but I'd really like to shake this naive feeling that I'm wading in a sea of overly-specific algorithm optimization races. I know a good start would be to realize that most ML is based purely on statistics, the field whose application with the right minds birthed statistical mechanics, thermodynamics and quantum mechanics all of which I feel are universally fundamental.
Anyone else feel like this? I would love review articles or any references to help persuade me. Thanks so much.
[link][12 comments]