Everybody seems to be talking about artificial intelligence and machine learning. Ever since Spielberg’s 2001 movie A.I. Artificial Intelligence, the abbreviation AI has been readily recognizable. News outlets have recently carried headlines such as, “AI in your car can brake faster than you”, or “Police use AI to predict crime”, or similar flashy statements. But then you may also read about machine learning algorithms that can convert a 2D image to 3D or have learned the grammar of a language.
After reading all that, you may try to explain these interesting advancements to your technically savvy friends. Only once you start describing do you realise how confusing these terms can be. Was it AI that predicted crime or was it a machine learning algorithm? Who extracts the grammatical structure of a sentence, AI or machine learning? It seems both. And why did they say in one article that AI is learning, when obviously machine learning has the magic ‘learning’ capability?
Maybe the two terms refer to the same thing, you may ponder. Is AI just the fancier form of the more technical machine learning? The term ‘artificial intelligence’ is certainly more inspiring than the somewhat dry ‘machine learning’. Are the two interchangeable, just like ‘motorized vehicle’ and ‘automobile’?
End the confusion
It’s time to end your confusion: AI and machine learning are not the same thing.
That being said, there is a lot of overlap in what machine learning does and what an AI does.
What if we told you that you can have AI that consists of zero machine learning? Also, there are many machine learning algorithms, really smart ones, that by no means constitute an AI.
To understand that, let us draw a comparison to a distinction between a car and an engine of a car. The engine is one of the most important technological components that makes it possible to build a car (wheels being perhaps the next most important), and yet the engine alone is not even close to being a car.
Critical as the engine is, many more components are needed to create a car. Wheels, a chassis, steering system and brakes are just four crucial elements to a car. Importantly, only once we put these base components together can we begin thinking about selling that product as a car.
To have a really attractive car product, we would add more components to our car. Lights for night driving, a windshield, mirrors, air conditioning/heating, seatbelts (and seats for that matter), suspension, indicators, a radio, airbags… the component list goes on. There is almost no end to how much we can improve the vehicle whilst still referring to it as the same thing: a car.
At some point in this process, we shifted from selling a collection of components to selling a car.
In a similar way, machine learning can be considered an engine that drives AI. AI is a final product that a person ignorant of the underlying technology can adequately use. It is that simple.
It’s really that simple
If a product is a machine learning software or hardware, then it’s meant for those that understand the technology and may be using it to build an AI. On the other hand, an AI would be sold to end users.
For example, physicians may use an AI diagnostic system to help them make early discoveries of diseases. By contrast, a person who writes code in Python to run deep learning models on TPUs (tensor processing units) is busy with machine learning hardware (the TPUs) and machine learning software (various libraries that can be used with Python). The latter person used machine learning to build cool things, including, but not limited to, an AI system.
To get yourself into machine learning you can download open source libraries such as scikit-learn, CNTK, TensorFlow and Keras. To get hands on with an AI, simply play with Cortana, Siri or Google Now – one of which is likely to be built into the smartphone you already own.
As we saw above, not all machine learning ends with AI, and not every AI solution runs on machine learning. Also, even if machine learning is the engine propelling a particular AI, other parts are needed for the whole thing to be deemed artificially intelligent.
To understand better how to find the exact point at which a transition takes place, we will discuss in our next instalment of this blog series, what machine learning is really about.
Discover more about AI for the enterprise: https://www.teradata.com/Insights/Artificial-Intelligence.
Danko Nikolic – Senior Data Scientist, Teradata
Danko Nikolic is a brain and mind scientist, as well as an AI practitioner and visionary. His work as a senior data scientist at Teradata focuses on helping customers with AI and data science problems. In his free time, he continues working on closing the mind-body explanatory gap, and using that knowledge to improve machine learning and artificial intelligence.
Dr. Frank Säuberlich – Director Data Science & Data Innovation, Teradata GmbH
Dr. Frank Säuberlich leads the Data Science & Data Innovation unit of Teradata Germany. Part of his responsibility is to make the latest market and technology developments available to Teradata customers. Currently, his main focus is on topics such as predictive analytics, machine learning and artificial intelligence.
Following his studies of business mathematics, Frank Säuberlich worked as a research assistant at the Institute for Decision Theory and Corporate Research at the University of Karlsruhe (TH), where he was already dealing with data mining questions.
His professional career included the positions of a senior technical consultant at SAS Germany and of a regional manager customer analytics at Urban Science International.
Frank Säuberlich has been with Teradata since 2012. He began as an expert in advanced analytics and data science in the International Data Science team. Later on, he became Director Data Science (International).