Artificial Intelligence

Jargon buster: artificial intelligence

 | 

Alice Cruickshank

Futurist Ray Kurzweil predicts machines will be as smart as humans by 2029, and even smarter than us by 2045. That’s a forecast that can’t be ignored.

AI offers great potential for every industry, not least property. From chatbots to ‘smart’ digital estate agents, if you’re not currently using AI you soon will be.

As the UK Government announces major investment in this technology, do you know deep learning from machine vision? We’ve got you covered.

AI

Artificial intelligence may be a scary concept, glamorised by sci-fi films, but ultimately AI is just very complex maths. The term refers to computers that can mimic human skills of problem solving and ‘learning’. Scientists have been studying the field of AI since the 1950s, so what seems to be a thoroughly modern concept has in fact been around for a long time. However, recent developments in Big Data have seen AI come to the fore once more.

Machine learning

This is when a computer adapts its activity using data, rather than being programmed to do so. This could be by using previous labelled data to predict future events, known as supervised machine learning algorithms, or by using unlabelled data when the machine calculates its own connections, know as unsupervised machine learning algorithms.

Deep learning

Used to evaluate Big Data, deep learning is a self-adaptive algorithm that improves as it analyses more data. This is the strand of AI that most resembles the human brain, as it uses artificial neural networks with multiple hidden layers to process complicated data in a non-linear way. A programme that uses image recognition to identify properties similar to what your client is seeking would be an example of deep learning.

Machine vision

As it sounds, this refers to a machine’s ability to view and comprehend visual images. This could be as simple as identifying an object, to obtaining information from abstract pictures.

Cognitive computing

This is hardware or software that mimics the human brain’s cognitive function, with the aim of providing insights to assist decision-making. This software is adaptive and interactive, and may use both data and sensory inputs to gain information. A chatbot is an example of cognitive computing, as it analyses what a user has written and responds to it in a ‘human’ way.

Natural language processing

This term also ties into chatbots, as natural language processing is the way in which computer language and human language interact. Computers are able to determine meaning from variants of phrases, rather than just individual words or commands. Natural language is also useful for looking over large documents, such as contracts.

Evolutionary algorithms

These are algorithms that find solutions to complex problems through mimicking natural selection. What this means is each potential solution is ‘fitness tested’ – just like natural selection, only the ‘best’ option will ‘survive’, or be used by the computer. This could be used for processes such as image processing, real options analysis and computed-automated design.

Your comments

Read our comments policy here