Artificial Intelligence (AI) and Machine Learning (ML) are the buzzwords of the moment, but does anyone really know what they mean? Or what the difference is?
Do not get caught up in the hype. The team at ebb3 are here to help by explaining the main differences between AI and ML:
Perhaps the best-known definition of AI comes from Andrew Moore, Former Dean of the School of Computer Science at Carnegie Mellon University. He defines AI as “the science and engineering of making computers behave in ways that, until recently, we thought required human intelligence.”
This definition sums up the main idea behind AI, but also reveals the scope and breadth of the field, with the words “until recently” being important. For example, fifty years ago a chess-playing computer programme would be considered AI under that definition, but nowadays this would seem antiquated as it is part of every computer’s operating system (OS).
Good clarification of the definition of AI has been provided by Zachary Lipton, Assistant Professor and Researcher at Carnegie Mellon University. He says that the term AI “is aspirational, a moving target based on those capabilities that humans possess but machines do not.”
AI is basically a term that covers numerous technological advances that have become part of our daily lives, with ML being one of them. Our favourite streaming services, such as Amazon and Netflix, use machine learning to power their video prediction systems. Google Home and Alexa are human-AI interaction gadgets which enhance our abilities as humans and foster our productivity.
There is no one definition of AI that is “right”, as the definition evolves as technology advances. What we consider cutting edge technology today may be considered simple and lacklustre in the future.
Tom M. Mitchell, Computer Scientist and machine learning pioneer, defines Machine Learning (ML) as “the study of computer algorithms that allow computer programs to automatically improve through experience.”
ML examines and compares datasets of all sizes in order to find common patterns and explore nuances. So, ML is one of the ways that we expect to be able to achieve AI and, therefore, is usually described as a branch of AI.
There are three main types of ML: “supervised learning”, “unsupervised learning” and “reinforcement learning.” Supervised learning features algorithms which try to model dependencies and relationships between the target prediction output and the input features so that the machine can predict the output values for new data based on what it has learned.
For example, if you load an ML programme with a large dataset of x-ray pictures along with their description then it should have the capacity to automate the data analysis of any x-ray pictures you feed in afterwards. It will look at the existing pictures and find common patterns through the label and indications. It will also compare its parameters with the examples it already has, to disclose how likely it is that any of the pictures contain the previously analysed indications.
Unsupervised learning algorithms, on the other hand, do not have labels on the data or output categories – and tend to be used in descriptive modelling and pattern detection. Reinforcement learning uses observations the machine has learned from its interaction with the environment to take actions that will minimise the risk. In this case, the machine is constantly learning from its environment through the use of iterations – a good example of this is computers beating humans on computer games.
Ml is a fascinating subject- especially as it is concerned with neural networks and deep learning – which seem similar to the way the human brain works (although there are essential differences).
Both AI and ML are fuelling the growth of immersive technologies – something which ebb3 use in design to help businesses realise massive productivity gains alongside significant cost savings when using our VDI/Workspace multi-purpose platforms. To find out more about how we can help you free your team, call us on 0203 818 1000 or email us at email@example.com