Contents Covered
[editorskit display=”wordcount” before=”Reading Time: ” after=” min”]
Art And Science Of Machine Learning:
The final course in Machine Learning with TensorFlow on Google Cloud Platform Specialization, Machine Learning Courses by Google offered through Coursera is art and science of machine learning. In this Machine Learning Courses by Google, you will learn ML intuition, good judgment and fine-tune experimentation and the skills needed to optimize your ML models for best performance.
In this course you will learn many rules and controls involved in training a model. First you need to manually adjust them to see their effects on model performance. Once familiar with knobs and controls, otherwise known as hyper-parameters, you will learn to tune them automatically using the cloud machine learning engine on the Google Cloud Platform. The Art and Science of Machine Learning, Machine Learning with TensorFlow on Google Cloud Platform Specialization’s course number 5 can be completed within 3 weeks in about 9 hours of study.
The skills that you can gain after completing The Art and Science of Machine Learning, a specialization of Machine Learning Courses by Google are:
- Tensorflow,
- Machine Learning,
- Cloud Computing,
- Estimator
PLATFORM: | COURSERA |
PROVIDER: | GOOGLE CLOUD TEAM |
COST: | FREE (AUDIT) |
DIFFICULTY: | INTERMEDIATE |
LANGUAGE: | ENGLISH |
CERTIFICATE: | AVAILABLE (PAID) |
EFFORT: | 9 HOURS |
START DATE: | ON GOING |
DURATION: | 3 WEEKS |
Syllabus – Art and Science of Machine Learning
Week 1: Introduction
This course overview highlights key objectives and modules in this course. First, you shall learn about those aspects of machine learning that require some intuition, good judgment, and experimentation. We can call it the Art of Machine Learning. You will then proceed in learning many knobs and levers involved in training a model. You will adjust them manually to see their effects on model performance.
- Introduction – (Video) 3 min
The Art of Machine Learning
In this course of Machine Learning you get to learn about The Art of Machine Learning. We will review those aspects of machine learning that require intuition, judgment, and experimentation to find the right balance and which is never quite right.
- Introduction – (Video) 1 min
- Regularization – (Video) 4 min
- L1 & L2 Regularizations – (Video) 4 min
- Lab Intro: Regularization – (Video) 12 sec
- Lab: Regularization – (Video) 2 min
- Learning rate and batch size – (Video) 5 min
- Optimization – (Video) 1 min
- Practicing with Tensorflow code – (Video) 1 min
- Lab Intro: Hand-Tuning ML Models – (Video) 18 sec
- Lab Solution: Hand-Tuning ML Models – (Video) 7 min
- Art of ML – (Practice Exercise) 2 min
- Learning Rate and Batch Size – (Practice Exercise) 10 min
Hyperparameter Tuning
In this module you can learn how to differentiate between parameters and hyper-parameters. Then there is a discussion on the traditional grid search approach and also learn how to think beyond it with smarter algorithms. Finally you will find out how convenient the Cloud Machine Learning engine makes automating hyper-parameter tuning.
- Introduction – (Video) 52 sec
- Parameters vs Hyperparameters – (Video) 2 min
- Think Beyond Grid Search – (Video) 3 min
- Lab Intro: Improve model accuracy by Hyperparameter Tuning with Cloud AI Platform – (Video) 23 sec
- Lab Solution: Improve model accuracy by Hyperparameter Tuning with Cloud AI Platform – (Video) 30 sec
- Hyperparameter Tuning – (Video) 8 min
Week 2: A pinch of Science
In this module, you will learn to associate science with the art of machine learning. First you are going to learn about how to regularize for sparsity so that we can have simpler, more concise models. Then you shall learn about logistic regression and learn to determine performance.
- Introduction – (Video) 55 sec
- Regularization for sparsity – (Video) 5 min
- Lab: L1 Regularization – (Video) 3 min
- Lab Solution: L1 Regularization – (Video) 51 sec
- Logistic Regression – (Video) 17 min
- L1 Regularization – (Practice Exercise) 4 min
- Logistic Regression – (Practice Exercise) 2 min
The science of neural networks
This module is focused on diving deep into the science of Machine Learning, specifically with neural networks.
- Introduction – (Video) 58 sec
- Neural Networks – (Video) 18 min
- Lab: Neural Networks Playground – (Video) 12 min
- Training Neural Networks – (Video) 14 min
- Lab: Using Neural Networks to build a ML model – (Video) 11 min
- Multi-class Neural Networks – (Video) 10 min
- Training Neural Networks – (Practice Exercise) 8 min
- Multi-class Neural Networks – (Practice Exercise) 4 min
Week: Embeddings
In this module, you will learn how to use embeddings to manage sparse data and to build machine learning models that use sparse data which consume less memory and train faster. Embeddings is also a way to reduce dimensionality, and in that way, it make the model simpler and more general.
- Intro to Embeddings – (Video) 2 min
- Review of Embeddings – (Video) 5 min
- Recommendations – (Video) 4 min
- Data-driven Embeddings – (Video) 3 min
- Sparse Tensors – (Video) 4 min
- Train an Embedding – (Video) 4 min
- Similarity Property – (Video) 7 min
- Embeddings – (Video) 4 min
Custom Estimator
In this module you will go beyond from using canned estimators by writing a custom estimator. By writing a custom estimator, you will be able to gain more control over the model function itself.
- Custom Estimator – (Video) 4 min
- Model Function – (Video) 6 min
- Lab: Implementing a Custom Estimator – (Video) 11 min
- Keras Models – (Video) 4 min
- Demo: Keras Models + Estimator – (Video) 2 min
- Custom Estimator – (Practice Exercise) 6 min
Summary: Art and Science of Machine Learning
All the key concepts that are covered in the Art and Science of Machine Learning course are reviewed.
- Summary – (Video) 1 min
- Specialization Summary – (Video) 2 min