Superior Studying Algorithms

About this Course

Within the second course of the Machine Studying Specialization, you’ll:

• Construct and prepare a neural community with TensorFlow to carry out multi-class classification • Apply finest practices for machine studying improvement in order that your fashions generalize to information and duties in the true world • Construct and use choice bushes and tree ensemble strategies, together with random forests and boosted bushes The Machine Studying Specialization is a foundational on-line program created in collaboration between DeepLearning.AI and Stanford On-line. On this beginner-friendly program, you’ll be taught the basics of machine studying and easy methods to use these methods to construct real-world AI functions. This Specialization is taught by Andrew Ng, an AI visionary who has led crucial analysis at Stanford College and groundbreaking work at Google Mind, Baidu, and Touchdown.AI to advance the AI subject. This 3-course Specialization is an up to date and expanded model of Andrew’s pioneering Machine Studying course, rated 4.9 out of 5 and brought by over 4.8 million learners because it launched in 2012. It gives a broad introduction to fashionable machine studying, together with supervised studying (a number of linear regression, logistic regression, neural networks, and choice bushes), unsupervised studying (clustering, dimensionality discount, recommender techniques), and among the finest practices utilized in Silicon Valley for synthetic intelligence and machine studying innovation (evaluating and tuning fashions, taking a data-centric strategy to bettering efficiency, and extra.) By the top of this Specialization, you should have mastered key theoretical ideas and gained the sensible know-how to rapidly and powerfully apply machine studying to difficult real-world issues. For those who’re seeking to break into AI or construct a profession in machine studying, the brand new Machine Studying Specialization is one of the best place to begin.

WHAT YOU WILL LEARN

• Construct and prepare a neural community with TensorFlow to carry out multi-class classification

• Apply finest practices for machine studying improvement in order that your fashions generalize to information and duties in the true world

• Construct and use choice bushes and tree ensemble strategies, together with random forests and boosted bushes

SKILLS YOU WILL GAIN

• Synthetic Neural Community
• Xgboost
• Tensorflow
• Tree Ensembles
• Recommendation for Mannequin Growth

Syllabus – What you’ll be taught from this course

WEEK

1

8 hours to finish

Neural Networks

This week, you’ll find out about neural networks and easy methods to use it for classification duties. You’ll use the TensorFlow framework to construct a neural community with just some strains of code. Then, dive deeper by studying easy methods to code up your individual neural community in Python, “from scratch”. Optionally, you’ll be able to be taught extra about how neural community computations are applied effectively use parallel processing (vectorization).

WEEK

2

9 hours to finish

Neural community coaching

This week, you’ll discover ways to prepare your mannequin in TensorFlow, and in addition find out about different essential activation features (apart from the sigmoid perform), and the place to make use of every kind in a neural community. You’ll additionally discover ways to transcend binary classification to multiclass classification (3 or extra classes). Multiclass classification will introduce you to a brand new activation perform and a brand new loss perform. Optionally, you can even be taught concerning the distinction between multiclass classification and multi-label classification. You’ll be taught concerning the Adam optimizer, and why it’s an enchancment upon common gradient descent for neural community coaching. Lastly, get a quick introduction to different layer sorts apart from the one you’ve seen so far.

WEEK

3

7 hours to finish

Recommendation for making use of machine studying

This week you’ll be taught finest practices for coaching and evaluating your studying algorithms to enhance efficiency. This can cowl a variety of helpful recommendation concerning the machine studying lifecycle, tuning your mannequin, and in addition bettering your coaching information.

WEEK

4

6 hours to finish

Resolution bushes

This week, you’ll find out about a sensible and really generally used studying algorithm the choice tree. You’ll additionally find out about variations of the choice tree, together with random forests and boosted bushes (XGBoost).