every blog needs subheader text Structuring Machine Learning Projects

Table of contents.

This course covers how to think about and improve machine learning systems.

You will learn how to build a successful machine learning project. If you aspire to be a technical leader in AI, and know how to set direction for your team's work, this course will show you how.

Course Resources

Week 1: ML Strategy

There are lots of ways to improve a deep learning system, so its important to sse quick and effective ways to figure out the most promising things to try and and improve.


Setting up your goal

Single number evaluation metric

Classifier Precision Recall
A 95% 90%
B 98% 85%

Satisficing and Optimizing metric

#### Train/dev/test distributions

Size of the dev and test sets

When to change dev/test sets and metrics

Comparing to human-level performance

Why human-level performance?

Avoidable bias

Bias Example Variance Example
Humans 1% 7.5%
Training error 8% 8%
Dev Error 10% 10%

Understanding human-level performance

Error Bias Example Variance Example
Human/Bayes 1% 1%
Training 5% 1%
Dev 6% 5%

Surpassing human-level performance

Improving your model performance

Andrej Karpathy interview

Week 2: More ML Strategy

Error Analysis

Carrying out error analysis

Cleaning up incorrectly labeled data

Build your first system quickly, then iterate

Mismatched training and dev/test data

Training and testing on different distributions

Bias and Variance with mismatched data distributions

Variance Problem data mismatch Problem
Training Error 1% 1%
Training-dev error 9% 1.5%
Dev Error 10% 10%

Addressing data mismatch

Learning from multiple tasks

Transfer learning

Multi-task learning

End-to-end deep learning

What is end-to-end deep learning

Whether to use end-to-end deep learning

Ruslan Salakhutdinov interview

course done and dusted:

posted , updated
tagged: courses View on: github