Hello to all, how was your day two?? Comment down fast, first. With my #7daysofml day2.
I hope it was great. Here’s how mine went.
Contents
Goals For The Day
Today I thought of raising the bar a bit. That’s why the list is longer.
- Types of Machine Learning
- Main Challenges of Machine Learning
- Testing and Validating
- Data Mismatch
- Hyperparameter Tuning and Model Selection
Challenges For The Day
Now, before diving into what I did I would like to share what went wrong:
First, I don’t know why my internet refused to work well on the earth(And we still can’t go to Mars😭). Only I know how I made this 👇 reach to you.
Then all the difficult terms I came across especially the Hyperparameter thing. I couldn’t understand that in the first few reads.
Next, the thing called time. I have never worked this fast with this slow internet.
And my brain, just when I started writing and editing things, it sort of froze.
I actually couldn’t think of anything to write about for like half an hour.
Some Interesting Things I Learnt On My Day 2
There are more than 2 types of machine learning. I always thought it was just supervised and unsupervised.
And the most interesting thing was for some reason a large amount of good data makes a better machine-learning model. I mean data is more important than the algorithm.
There’s some research on this you can read it here.
Another interesting thing was the “no free lunch theorem“. It said that all the algorithms are no better than the others if we don’t make assumptions.
All in all machine learning turned out to be everything I didn’t think. I thought it was all about fancy algorithms.
#7daysofml Day 2 Notes
But as I read I got to know it is much more than that. Much more data and resource expensive task.
Instead of sharing the whole thing as an article and attaching the resources, I thought why not a PDF of my notes?
(I know my handwriting sucks but the pointers are great for revision)
And I would anyways share articles on each topic after the 7 days end.
And yes if I ever forget to let you know all your suggestions are always welcomed(Badly needed).
See you tomorrow.👋
