Jobin Reji, King’s College London
We have looked at AI in our previous articles and most recently, about the convenience it offers to traders as they execute deals. As a branch of computer science, artificial intelligence attempts to build machines capable of intelligent behaviour. Today we are to look at the role Artificial Intelligence has on Machine Learning, the difference between the two, and finally what the future would look like for machine learning as AI becomes more advanced and capable.
Firstly, we can classify AI into two categories: Applied AI and General AI. Applied AI is when the system is given a specific task to do, for example, the technology that allows autonomous driving in vehicles or self-service checkouts at supermarkets. These are designed for only a single task hence their capabilities are limited. On the other hand, general AI is made to think like humans, and the category Machine Learning could be identified in. Unlike applied, general AI can carry out various tasks it is faced with. Stanford University defines machine learning as the science of getting computers to act without being explicitly programmed. The greatest breakthrough in machine learning came in 1959 when Arthur Samuel came to the realization that instead of teaching the computer how to perform the task, teach the computer how to learn to carry out the task so they learn for themselves. This is the basis of much of the machine learning equipment that is used today. In the late 20th century, engineers designed computers to be coded like humans, this allowed the computers to access the abundance of the internet and use this to fit their needs and requirements. An example of this is Google’s search engine. It learns from a user’s typos and uses it to re-correct them the next time they search for the same.
“AI is basically the intelligence – how we make machines intelligent, while machine learning is the implementation of the computer’s methods that support it. The way I think of it is; AI is the science and machine learning is the algorithms that make the machines smarter.”
Nidhi Chapel, Head of Intel’s Machine Learning
Machine Learning has certainly been seized as an opportunity by marketers. After AI has been around for so long, it’s possible that it started to be seen as outdated even before its potential has ever truly been achieved. It is true that we have more knowledge about technology in general than we used to just 30 years ago, however, there hasn’t been enough developments that have been above and beyond our reckoning. Is this due to the monotonous pace AI is being developed, or that our scientists still haven’t been given or discovered the tools for these advances in AI? From a medicinal standpoint to climate change, AI can make a difference, but we haven’t been seeing advances that are all worthy additions to public scrutiny of general AI capabilities.
In order for AI to progress, machine learning must make undertake great strides in terms of performance, and this is merely possible in the traditional computing world, where problems are well-defined. Machine learning algorithms still have vast improvements, and that’s the reason many of the large technology companies are making it a central focus of their strategy and working tirelessly to make machines more intelligent.