0:0030:57

But what is Machine Learning? : A complete starter pack

Mayuri Madhur 30:57 2021

Deep learning is not replacing machine learning, it lives inside it

A car's resale price is easy to guess from its age alone, until condition and odometer readings join the table and the arithmetic in your head collapses. That is the gap machine learning fills, says Flipkart software development engineer Mayuri Madhur, and newcomers often get lost in the jargon around it. So where do classical models, ensembles, reinforcement learning and neural networks actually sit, and when does each one win?

Get a plain-language map of machine learning types and when to choose each one

Watch on YouTube

In short

Mayuri Madhur, a Software Development Engineer at Flipkart, gives beginners a structured map of machine learning: what it is, why it differs from traditional programming and where it sits within AI. She walks through classical supervised and unsupervised learning, reinforcement learning, ensembles and neural networks, and argues that choosing a model depends on the problem rather than on hype about deep learning.

“In the first approach, we do the thinking. In the second approach, machine does the thinking.”
“Hence, data is the most important thing for making our machines learn. The more diverse the data, the better is the result.”

What you will learn

  • Distinguish traditional rule-based programming from machine learning, where the machine derives the rules from data and outputs.
  • Place machine learning within AI, and neural networks and deep learning within machine learning.
  • Tell supervised from unsupervised learning, and classification from regression, using everyday examples.
  • Explain how stacking, bagging and boosting combine weak learners into a stronger model.
  • Choose between an ensemble model and a neural network based on real-time needs and problem complexity.

Key takeaways from the speaker

  • Clearer picture on Machine Learning. Ideal for people who are looking for a point to get started with Machine Learning.
  • This session will help you in clearing some of the misconceptions you might have about machine learning and data science
  • It will be a very basic session on Machine learning which will help you understand what is Machine learning and where does it stand in AI

Details

Level
Beginner
Format
Tech talk
Technical depth
Introductory technical
For
Student, Early career, Career changer
Length
30:57
Published
July 2021

Questions this talk answers

How is machine learning different from traditional programming?

In traditional programming people write the rules and the machine produces outputs; in machine learning the machine receives data and outputs and works out the rules itself. Those learned rules can act differently on different kinds of data.

Are neural networks and deep learning the same as machine learning?

No. Machine learning is one part of the wider field of AI, neural networks are one type of machine learning, and deep learning is the modern way of building, training and using neural networks.

What is the difference between classification and regression?

Classification puts data into categories, such as whether a student passed or failed, while regression forecasts a number, such as the percentage the student scored.

When should you use reinforcement learning?

When you have no data but do have a simulation environment, so the system can learn from rewards and penalties for its own actions, as self-driving cars do.

Should you always pick a neural network over a classical or ensemble model?

No. Madhur compares it to choosing between a sports car and a loaded truck: ensembles such as random forest suit real-time needs, while neural networks suit very complex, unpredictable problems.

About this talk

The main topics to be discussed: Machine Learning : What is it and why it matter? Learning vs Intelligence Main types of Machine Learning : i) Classical ML ii) Reinforcement Learning iii) Ensembles iv) Neural Networks The complete picture

Topics

Related talks

Browse the library