The Fine Line Between Artificial Intelligence and Machine Learning
Anyone who is partial to a science fiction film will revel in modern day life. Finally, we can walk into a room, say ‘lights on, please’ and the house dutifully responds. Likewise, if we fancy listening to a bit of Michael Jackson during a run, we need only ask our smartphone to play his greatest hits, using only our vocal chords.
Unfortunately, with new technology comes a myriad of ways in which things can be described and labeled. The Amazon Echo is a case in point – when Amazon’s digital assistant was launched in 2015, it brought the phrase ‘machine learning’ to the forefront of the media and, ever since, people have referred to both that and artificial intelligence as a means to explain what’s going on inside these wondrous boxes.
In this post, we’re going to attempt to draw the fine line between these two seemingly competing technologies.
Artificial intelligence doesn’t exist
Without wishing to rain on anyone’s parade, we do feel it important to point out that artificial intelligence (A.I.) doesn’t technically exist at this point. The closest we’ve come to a computer capable of independent thought is sophisticated algorithms that can crunch through the log files and databases containing our digital actions. Such algorithms can then use what is known as ‘machine learning’ to try and make sense of the requests we make and things we do.
Amazon’s Echo, for example, is rather more simple than it may appear. Described by many as a ‘command line’ device, it relies on relatively specific phrases and requests in order to get the job done. Apple’s Siri, on the other hand, is capable of understanding different inferences which is why we can be rather more conversational with it when seeking answers.
Let’s look at what these two types of technology really are.
What is artificial intelligence?
Artificial intelligence (AI) is a type of computer science whose aim it is to create intelligent machines and software. A machine that is capable of AI will usually be tasked with one or a combination of the following:
- Amassing knowledge
- Ability to interact with objects and the environment
The key aspect above is knowledge. AI machines can only perform the tasks they are designated if they have a wealth of knowledge from which to draw when deciding appropriate actions or solutions to problems.
While AI is capable of learning, it still does so within a confined set of rules, beyond which it is unable to see or comprehend. AI, therefore, is superb at performing common tasks regularly and without fail, but it simply can’t think creatively or use any form of emotional intelligence to trick anyone into thinking it is human.
What is machine learning?
You’ll no doubt have spotted the word ‘learning’ in the above description of AI, so how does machine learning differ from a technology that appears to do the same thing?
Machine learning is a concept that has been around for as long as computers. It is commonly used for predictive analytics, but can also be applied to optical part sorting, failure detection, analysis and product testing – in other words, machine learning is particularly useful in largely automated systems, where equipment is required to make its own decisions.
Essentially, it refers to computer software that doesn’t rely on rule-based programming, but instead on algorithms that can adapt and learn from the data they receive. Importantly, the software can continually improve the quality of the predictions they make as time goes on.
At a practical level, it means companies like Amazon, who use cloud-based machine learning in their warehouses, can save on time and the expense of using dedicated data analysts to find patterns and make predictions of seasonal highs and lows to improve the efficiency of their inventory.
In a way, you could say machine learning boils down to a type of AI, since it represents software that is capable of teaching itself to grow and expand its knowledge on a particular subject. Consequently, it can appear almost ‘human’.
As you might guess, the ability to read into that data and learn from it without human input requires some incredibly clever programming from the outset, which is why machine learning represents one of the most exciting new forms of development around.
Here are 4 very public examples of machine learning in action:
- Ever received an automated SMS message from your bank that suggests there may be fraudulent activity taking place on your account? You can thank machine learning.
- Looking forward to the much-hyped Google self-driving car? That vehicle represents the very essence of machine learning.
- Do you find recommendations popping up for the latest Amazon offers all over the internet? Are they surprisingly relevant and tempting? You’ll find machine learning behind this clever sales tactic.
- Run a business and rely on software to alert you whenever your brand is mentioned online? Combine linguistic rules with machine learning and that’s what you get.
As you’ll note above, machine learning already has very real and incredibly useful applications in everyday life. Some would argue far more so than AI, which is usually confined to prototype robots tripping over themselves in laboratories.
There you have it – the difference between artificial intelligence and machine learning. One is still on the storyboard, the other already has practical applications.
In our opinion, it’s important that business leaders embrace machine learning as a modern form of data analysis wherever they can to gain valuable insight into the behaviour of their clients and customers.
As for A.I…. One thing we can probably all agree on is that such technology is still incredibly nascent, and we have a very exciting time ahead watching it develop further.