AI trained to detect least energy-efficient homes
- Published
- comments
Our homes use up a lot of energy - on heating, lighting and powering the objects we use like computers, cookers, fridges and kettles - but how can you tell if energy is being wasted?
Scientists at University of Cambridge might have the answer - they've trained artificial intelligence (AI) to detect when a house isn't very energy-efficient.
AI is the science of getting computers to do things that usually you'd need a human to do.
This model is the first of its kind, and can see where and how much energy might be leaking from a home.
Houses with more red are harder to decarbonise
These leaky homes are described as hard-to-decarbonise (HtD).
Decarbonisation means switching something from using fossil fuels to green energy.
Fossil fuels are things found in the Earth's crust and contain carbon and hydrogen, which can be burned for energy - but are non-renewable - they will eventually run out
Examples of fossil fuels are coal, oil and natural gas
Burning fossil fuels creates carbon dioxide that traps heat from the Sun in the Earth's atmosphere - one of the causes of global warming
So, a home that is HtD is one that is difficult to switch to cleaner energy to heat and power it.
Houses can be HtD for many reasons, such as their age, how they're built, and their location.
Solar panels can provide green, renewable energy to homes
The AI model built by data scientist Maoran Sun and his supervisor Dr Ronita Bardhan can identify HtD homes with an up to 90% accuracy rate, something they expect to increase as they gather more data.
This is important as HtD houses are responsible for more than a quarter of all emissions from housing, they say.
How does artificial intelligence work?
How can AI help homes to be better for the planet?
Dr Bardhan says governments need to identify these HtD homes in order to reach net-zero targets.
"Our model can direct them to high priority houses, saving them precious time and resources," she said.
The model can see where houses are losing most heat, such as through roofs or windows, and can also identify whether a building is old or new.
The scientists think that as more data is added to their AI model and it becomes more accurate, it will help local authorities and governments achieve green targets more quickly and easily.