Hey everyone, I'm 16 years old and really interested in diving into machine learning in the future. I'm at an intermediate level with Python and have some experience with libraries like Numpy and Pandas. I've even built a few games using Unity. Just recently, I tried out scikit-learn and played around with concepts like train_test_split and the number of neighbors in k-NN.
My main struggle is figuring out what exactly I should be learning and where to start. I know I should be working on projects, but I'm not sure how to tackle them without having a solid grasp of the syntax and algorithms yet. Plus, when I'm learning something new, I often question whether I know enough before moving on to the next topic.
By the way, I'm not a fan of studying math in isolation; I prefer to learn it as I need it for machine learning. So, if you have any recommendations for resources or advice on how to advance in this field, I'd really appreciate it! Thanks!
7 Answers
I also learned a bit of TensorFlow for projects in Bayesian statistics and image recognition. It's super useful, so consider checking that out!
Definitely give Kaggle a try! It's a data science platform that provides free ML courses and has loads of open-source datasets. You can even join competitions, which is great for practical experience. The courses don't cover everything in detail, but they're perfect for learning the basics and starting on your own projects.
Learning programming well is essential, especially if you want to get into machine learning. The more solid your programming skills, the easier it will be to grasp ML concepts!
It's awesome that you're interested in neural networks! That's the foundation of a lot of modern AI and machine learning. I built a Japanese character recognition project from scratch for a class, and it really helped me understand the basics. Make sure to dive deep into neural networks and start with the simpler models first!
Take a look for PyTorch projects on GitHub. Trying to build and run existing projects can be a great way to learn hands-on! You'll get to see real-world applications and improve your coding skills at the same time.
Check out this book: www.statlearning.com. I used it during my undergrad research, and it's a solid resource for understanding different types of neural networks. They have both Python and R versions, so you should find it useful. It might help to get some basic stats knowledge too. For a career in this field, brushing up on calculus and linear algebra could be beneficial; you don’t have to master them, but having a solid foundation will help you understand the algorithms better.
You might want to look into /r/learnmachinelearning. Their wiki has a bunch of great resources to kickstart your journey. I've heard fantastic things about Andrew Ng's courses listed there. It could be a nice starting point for you!
Did you really read through all 600 pages of it?