Day 2 of the ML: Teach by Doing Project

In Day 2, I dived into what exactly is machine learning. I looked at the entire history of machine learning.

I also learnt about the differences between artificial intelligence (AI), machine learning (ML) and deep learning (DL).

Here are my learnings from Day 2:

(1) ML is everywhere. Right from Alpha Go to Netflix/Youtube recommendation systems to Cancer diagnosis to Voice assistants and Face detection.

(2) We live in a unique age where tremendous computer power, combined with brilliant algorithm research, combined with availability of large amounts of data has made AI extremely powerful. This was not the case 20 years ago.

(3) There is a difference between AI, ML and DL. AI is the biggest umbrella. ML is its subset. DL is the subset of ML. An example of a machine which is AI but not ML: Rule based chatbots.

(4) ML is not just a magic box which takes in data and gives awesome results. Humans have a very big role to play in making ML systems work.

(5) Here are the 6 steps to make any ML application work:

• Step 1: Get data
• Step 2: Think of the space of possible solutions
• Step 3: Characterise what makes a good solution
• Step 4: Find algorithm
• Step 5: Run algorithm
• Step 6: Validate result

(6) Step 5 is the only part which the computer does for us. Users have to do all other steps.

(7) ML focuses on generalisation more than estimation. Generalisation problems are very hard. Example: weather/climate prediction.

Above is the Youtube video on Day 2 of the ML: Teach by Doing Project.

My Lecture Notes and links I go through in the video have been added in the Youtube video information.

See you in Day 3, where I will be looking at Types of ML models.