Try to follow the below steps, if you want to how to get started with the practical aspects of Machine Learning and Artificial Intelligence:
- Go to Kaggle and a make an account if you don't already have one.
- Go to the datasets tab and download a dataset: Make sure the dataset is below 5GB. The reason being, any dataset over this threshold is bound to be overwhelming to work on. Learning will require you tweaking the code over and over again and if you have a large dataset, it will take a lot of time. Even simple procedures like sorting, joining, dropping, or searching will take a lot of time. You don't need that.
- A further suggestion I often give is to download a dataset that has to do with computer vision. A lot of parameters and insights will come to light like how to make the bounding boxes of training set fit on the image for visualisation purposes, image preprocessing and reshaping, how to handle duplicity etc.
- After downloading the dataset of your choice keeping in mind the guidelines above, forget that you’re on a clock. Take your time with that dataset. Even if takes you 10 days just to preprocess the data, I would say that is incredible achievement.
- The reason I suggested Kaggle, is because for any given dataset you download, Kagglers are kind to post “Kernels” with some starter code or complete code in case you are stuck. I highly recommend NOT looking at that for the first 5 days. If you feel stuck after that, take a look at the starter code.
- Another practical aspect of ML/AI is deployment and scalability which no one ever talks about. After you’ve got the dataset preprocessed and the model gives you good enough results, attempt to implement your algorithm all over again but this time with parallel processing. If you’re not confident about parallel processing, you can take a Udemy or Coursera course for parallel processing architectures like Apache Spark.
- Compare the algorithm run times for both the implementations and you’ll see the wonders of parallel processing. Note those times down and document your entire effort nicely. This will be very useful in the future when you apply for a job. There’s also a potential preliminary research paper in there. You could use different frameworks for parallel processing and note down the algorithm run times and that comparison would make a solid preliminary research paper. If you know someone who knows a different language or if you yourself do, maybe you can use that in the comparisons too.
- If you can pull this off, it will give you a confidence boost like no other, you will have a potential research paper of your own, and you will be job-ready.
Join the machine learning online course fast, to learn machine learning concepts in detail and get certified.
Watch this video tutorial, if you want to know more information about ML