Machine Learning (ML) vs Deep Learning (DL)

The Best Explanation: Machine Learning vs Deep Learning

We’ve been tackling buzz words in the tech industry recently because there is a certain trend that occurs once a term is coined. Everyone uses the term without fully getting it and that causes misinformation, confusion, and sometimes fake news. In this piece we are looking at the terms machine learning and deep learning.

Every time a new tool or app is invented, a new word follows. So, let’s tackle two that have been flying around our heads for the past few years: Machine Learning (ML) and Deep Learning (DL). Techies, business gurus, and marketers love these words and throw them around whether or not they understand the differences.

Side Note: We know that this topic is old news, it’s discussed continuously. Which is why we had to write about it, clearly it’s not being fully understood because all the current content out there is either too simple or too complicated.

The 100 Word Explanation

ML and DL have one core thing in common, they both relate to Artificial Intelligence (AI). Let’s start with simple definitions:

• Artificial Intelligence: Computer system(s) that mimics and/or replicates human intelligence.
• Machine Learning: Allows computers to learn on their own.
• Deep Learning: Algorithms attempting to model high level abstractions in data to determine a high level meaning.

A simple example: If AI is being used to recognize people’s emotions in pictures, machine learning algorithms would input thousands of pictures of faces into the system. Deep learning would then help the system recognize patterns in the faces and emotions they share.

Going Deeper

The explanation above is the over simplified explanation of the three and helps those new to tech or confused with the jargon get it. In reality, it’s way more complicated than that and deep learning is by far the most confusing as it works with data, neural networks, and math.

If you’re still interested, keep reading.

“You should use a picture of Johnny 5 from Short Circuit, with the words ‘NEED INPUT’” — a Slack message I received when telling the team I was writing this piece.

Machine Learning

Machine learning analyzes data and crunches numbers, learns from it, and uses that to make a prediction/truth/determination depending on the scenario. The machine is essentially being trained, or really training itself, on how to perform a task correctly after learning from all the data it has analyzed. It’s building its’ own logic and solutions.

By the way, machine learning can be done with a bunch of different algorithms like:

• Random Forest & Decision Tree: A collection or ensemble of simple tree predictors, each capable of producing a response, like Netflix suggesting movies based off of your star ratings.
• Linear Regression: Predicts the value of a categorical outcome with limitless outcomes, like figuring out how much you can sell a car for based on the market.
• Logistic Regression: Predicts the value of a categorical outcome with a limited number of possible values, like figuring out if you can sell a car for a certain cost.
• Classification: Puts data into different groups, like filing documents or emails.
• Naive Bayes: A family of algorithms that all share a common principle, that every feature being classified is independent of the value of any other feature, like predicting happiness in photos of children.

The list goes on. There are many that all do different things, are part of algorithm families, and/or work well together or on their on.

There are also two types of machine learning algorithms, supervised learning and unsupervised learning.

Supervised learning requires a human to input the data and the solution, but allows the machine to figure out the relationship between the two. This is extremely helpful in mathematical situations.

Unsupervised is putting in random numbers/data for a certain situation and asking the computer to find a relationship and solution. It’s kind of like shooting a target in the dark, you won’t know what you hit until you put the lights on.

So, machine learning eliminates the need for someone to continuously code or analyze data themselves to solve a solution or present a logic. It cuts a huge corner and makes life a little easier.

If you want to dive more into machine learning, we recommend taking this online course from Lynda.

“MFW I think about explaining Deep Learning” — Probably most of the Internet.

Deep Learning

Deep learning crunches more data than machine learning, that is the biggest difference. So, if you have a little bit of data, machine learning is the way to go but if you’re drowning in data deep learning is your answer. Deep learning algorithms are powerful and they need a lot of data to give you the best solution/outcome, but buyer beware. Deep learning algorithms need powerful machines, machine learning algorithms don’t.

Why? Well, deep learning algorithms do complicated things, like matrix multipications, which require a graphic processing unit (GPUs). They also try to learn high-level features, so in the case of facial recognition the algorithm will get the image pretty close to the RAW version in replication whereas machine learning’s images would be blurry. Another powerful feature, it forms an end-to-end solution instead of breaking a problem and solution down into parts.

So you want the power of deep learning and you’ve got the high end machines? How much time do you have? Deep learning takes a long time to process data and find solutions and by long time, in some cases, years.

Now, you may be asking ‘what is deep learning composed of?’ Well, it is composed of the machine learning algorithms, neural networks, and AI. It is the third tier of the two and uses multi-level techniques and methodologies to build different solutions.

If you want to know more about deep learning, and there is a lot to learn, learn from the masters.

The Take-Away

Machine learning and deep learning are two different things composed of the same common core of AI. They’re also good to use in different scenarios yet one should not be used over the other unless there is an absolute need.

“[Deep learning] AI is the new electricity.”

- Andrew Yan-Tak Ng, former chief scientist at Baidu

However, when using deep learning you will use machine learning as they overlap one another. Also, according to some researchers and data scientists once we figure out deep learning beyond the guessing game that it is now, it will most likely solve many of our everyday computer, business, AI, marketing, and other problems.

As Andrew Yan-Tak Ng, former chief scientist at Baidu, where he led the company’s Artificial Intelligence Team says, “[Deep learning] AI is the new electricity”. As a facial recognition company, which uses deep and machine learning, we couldn’t agree more.

Want to learn more about the benefits of facial recognition for your business? Let’s talk — Contact us today.

Developer? Try out our API and SDK — It’s free to get started.

This article was originally published on Kairos.com

Kairos’ mission is to make it easy for any business to benefit from face analysis, enriching the experience between humans and machines, and being the premier partner for anything to do with facial recognition.