Explore machine learning's essentials for software engineers, delve into supervised learning, neural networks, and deep learning, and gain skills to tackle real-world data challenges effectively.

pml_final.tar.gz

just_run

chapter06

SPA_07

SPA_gdb_frontEnd

J12-1

J12-2

SPA-server_image

Handson-10

Machine learning is the future for the next generation of software professionals.

This course serves as a guide to machine learning for software engineers. You’ll be introduced to three of the most relevant components of the AI/ML discipline; supervised learning, neural networks, and deep learning. You’ll grasp the differences between traditional programming and machine learning by hands-on development in supervised learning before building out complex distributed applications with neural networks. You’ll go even further by layering networks to create deep learning systems. You’ll work with complex real-world datasets to explore machine behavior from scratch at each phase.

By the end of this course, you’ll have a working knowledge of modern machine learning techniques. Using software engineering, you’ll be prepared to build complex neural networks and wrangle real-world data challenges.

Fundamentals of Machine Learning for Software Engineers

# Objective
We’ll get an introduction to supervised learning in this chapter. Within a couple of chapters, we’ll code our first machine learning (or simply “ML”) system. Then we’ll evolve this system, one small step at a time, until it becomes powerful enough to read handwritten digits.

In the future, we are unlikely to ever write machine learning algorithms from scratch. But doing it once to grasp the fundamentals is priceless. We’ll understand every line in the final program. Machine learning will never look like magic again.

Author’s note on the magic of machine learning

Software developers like to share war stories. As soon as a few of us sit down in a pub, somebody asks: “What project are you working on?” Then we nod our heads off as we listen to each other’s amusing, and sometimes horrible, tales.

In the mid-90s, during one of those evenings of bantering, a friend told me about the impossible mission she was on. Her managers wanted a program that would analyze X-ray scans and identify diseases, such as pneumonia.

My friend had warned management that the task was hopeless, but they refused to believe her. If a radiologist could do it, they reasoned, then why not a Visual Basic program? They even paired my friend with a professional radiologist so that she could learn the job and turn it into code. That experience only reinforced her opinion that radiology required human judgment and intelligence.

We laughed at the futility of the task. A few months later, the project was canceled.

Fast-forward to more recent times: in late 2017, a research team at Stanford University published an algorithm to diagnose pneumonia from X-ray scans. The algorithm wasn’t just okay—it was more accurate than professional radiologists. That was supposed to be impossible! How the heck could they write that code?

The answer is that they didn’t. Instead of writing code, they cracked the problem with machine learning. Let’s see what that means.

# Comparison between ML and conventional programming

A program can be built to play video games by storing some rules which define how the game works. For example, some rules might define how enemies behave or how the user fires the weapon. These rules can then be encoded into a program that executes them if:

enemy = get_nearest_enemy()
​​if​ enemy.distance() < 100:
    decelerate()
​​    if​ enemy.is_shooting():
​        raise_shield()
​    else​:
​        ​if​ health() > 0.25:
​ 	     shoot()
​ 	 ​else​:
​ 	    rotate_away_from(enemy)
​else​:
    //# ...a lot more code​

Most of the code would be a big collection of `if..else` statements, mixed with imperative commands such as `shoot()`.

Granted, modern languages give us the means to replace those ugly nested ifs with more pleasant constructs such as polymorphism, pattern matching, or event-driven calls. However, the core idea of programming stays the same. We tell the computer what to look for, and we tell it what to do. We must list every condition and define every action.

This approach has served us well, but it has a few flaws:
 
First, one must be exhaustive. We can probably imagine dozens or hundreds of specific situations that we would have to cover in that video game–playing program. What happens if the enemy is approaching, but there is a power-up between us and the enemy, and the power-up is shielding us from enemy fire? A human player would quickly notice the situation and take advantage of it. If we coded for that special case, then our program will deal with it. However,  we know how hard it is to cover all special cases, even in structured domains like accounting. Good luck listing each and every possible special case in complex domains like playing video games, driving a truck, or recognizing an image!

Even if we could list all those decisions, we would have to know how to take them in the first place. That’s a second limitation of programming, and a showstopper in some domains. For example, take a computer vision task like our original problem (mentioned in [The magic of machine learning](https://www.educative.io/courses/fundamentals-of-machine-learning-for-software-engineers/programming-versus-machine-learning#Objective)): identifying pneumonia in chest scans.

We don’t really know how a human radiologist recognizes pneumonia. Yes, we have a high-level idea of it, like: “the radiologist looks for opaque areas.” However, we don’t know how the radiologist’s brain recognizes and evaluates an opaque area. In some cases, the expert herself cannot tell us how they came to a diagnosis, except for a rather vague one. Since we don’t know how those decisions happen, we can’t instruct a computer to take them. That’s a problem shared by all typically human tasks, such as tasting a drink or understanding a sentence.

Machine learning, on the other hand, turns traditional programming on its head. Instead of giving instructions to the computer, ML is about giving data to the computer and asking it to figure out what to do:


Distinguish between conventional programming and machine learning.

How Machine Learning Works

Our First Learning Program

Walking the Gradient

Hyperspace

A Discern Machine

Get Real

The Final Challenge

The Perceptron

Designing the Network

Building the Network

Training the Network

How Classifiers Work

Batchin’ Up

The Zen of Testing

Let’s Do Development

A Deeper Kind of Network

Diabetes Prediction Using Keras

Defeating Overfitting

Taming Deep Networks

Beyond Vanilla Networks

Into the Deep

Recognize Handwritten Digits Using a Deep Neural Network

Programming versus Machine Learning

Objective