Looking at the rest of the topics in the beginners section, I’ll probably sound like and ultra noob, but I’m interested in getting into ML. Rather as a hobby at this point.
Tiny bit about myself: in my thirties, a bit geeky and the “local tech guy” at home and work (I work in finance). Obviously interested in AI, I used to hack and jailbreak consoles, phones etc. have played with some Linux distros, but generally don’t know any python etc.
The question is where do I start to dig a bit deeper in this world? I was able to use the Google colab space to run a couple of demos posted here, so maybe I’m not a completely clueless chicken running around blind. Are there good courses to start for beginners on machine learning? Udemy works good for me as I have free access through my employer.
I’d be grateful if any of yas find a couple of minutes to nudge me in the right direction. I did some research online already, but would appreciate input from someone who’s been through this route already.
Depends on your interest area. I seem to have fallen into machine vision as i wanted to create AI to do image recognition. Took me 2 weeks of hard slog to find the code, build a simple machine learning environment in python, and then get it to work.
For machine vision, yolo ( ultralytics.com ) is simple and easy to use, having an objective will drive your study and research. Sometimes its as simple as " this particular problem really annoys me" so you set out to solve a problem. Start simple, get good with python installs and wrangling package versions etc. AI is mostly python.
I get annoyed with fan boy " connect to the internet to do everything" -mentality - i think if you cant get it running on a stand alone pc( after installs are done ) its too easy to use someone elses services and you’ll learn nothing.
Im building a stand alone ANPR system to unlock our garage at home , only when it sees only our car reg in the driveway. Its stand alone for security reasons ( not internet connected ), which means you have to do a complete install on a pc or nuc.
Some courses might be helpful, but hands on drives innovation and problem solving and is the best teacher. Pick a project and start learning…