What is Machine Learning and Why is it Relevant to Construction?

As business roles such as Head of Innovation or Digital Transformation increase exponentially, some feel confused about the myriad of technologies now available to us.

Machine learning is one of the many terms that gets thrown around, along with artificial intelligence (AI) and other previously mythical concepts. So let's make sure we know the definition of machine learning to start with... here's an amusing one from a Google search: "Machine learning studies computer algorithms for learning to do stuff" (thanks, Robert Schapire). So basically, it's AI that can learn from experience, without the need to be programmed (or "taught") by humans.

Now we've got that cleared up, let's look at machine learning through a construction lens...

What are the use cases for machine learning in construction?

Machine learning and AI have already impacted other industries to improve customer experience and drive profit. Of course the technology industry is the first to come to mind. For example, Google's image recognition lets you search more easily through your pictures. Facebook's machine learning algorithm makes your news feed more relevant to you (in theory!). But don't be tricked into thinking machine learning is exclusive to the tech world - companies we use every day are rapidly developing this technology. Self-driving cars are closer to becoming available to everyone. Larger retailers use AI online to recommend products we might like. There's even (disturbingly) a Barbie doll that uses machine learning to respond to a child according to what the child tells it...

AI and machine learning sometimes get a bad reputation, but that's usually because people fear what they don't understand. It seems like magic because there's so much behind it that we can't see how it works. Once you find out more about the humans behind the machine learning, the concept becomes easier to accept.

In architecture, engineering & construction (AEC), technology has been surprisingly slow to take hold - but even this is starting to change now. At Digital Construction Week, we heard fascinating examples of how machine learning is is already making a difference in the AEC industry. For example, Rania Kaadan, Technical Consultant at Invennt, discussed how AI in architecture could help people rebuild old cities destroyed by war.

Does construction need machine learning?

Well you could say for 2018 (and this blog was written in December 2018): "no". But to remain competitive in a fast-evolving world, we believe the answer will soon be "yes". While much of the industry still asks, "what is machine learning?", more progressive companies are already using it in areas such as:
- Project planning (where Kreo comes in)
- Training and education on site
- Monitoring on site

Potentially the biggest opportunity for machine learning to take hold is tied up with Building Information Modelling (BIM). We at Kreo are already speaking to many organisations that are excited to start using our advanced technology. As with all technologies, after a few companies can boast about the time and money they have saved from changing their process, others will follow. More importantly, as soon as enough companies start undercutting their competition because of machine learning time savings, everyone will soon rely on it.

What is machine learning's future in construction?

Machine learning will soon go beyond the grasp of the early adopters - it will, like other technologies, become a necessity. Of course for us, the easiest example to use is Kreo. Our experts have trained machines over time to recognise BIM elements automatically. It identifies anything that is labelled correctly in a BIM model. Where the software is unsure of the information, it will classify according to the most likely option. The software will ensure that the human agrees with its assumption and can be taught to learn any new information if the assumption is incorrect. Eventually you'll have a tailor-made product to create instant accurate cost and schedule estimates.

Another example is how machine learning will impact the property tech market - again the useful example comes from Kreo. Learning from thousands of other building designs, the machine learning algorithm can generate a BIM model from scratch according to the users requirements. This could save companies weeks of time in waiting for initial design concepts. More importantly, it will help them get their cost and schedule predicted correctly from the start. This will stop companies who unrealistically bid for projects who later find out that the real cost and schedule will be much higher. Going over time and budget will be a thing of the past.

"If it ain't broken, don't fix it"

When we hear people say that they want to go ahead with their old ways of working, we know that the problem is systemic. We at Kreo say, "It is broken!", and it's all of our duty to fix it.

If we risk continuing with the status quo, we will continue seeing companies going into administration and people losing their jobs en masse. We will continue seeing a lack of young people wanting to join the industry. We will continue to see a lack of diversity. We will continue to accept mediocre profit margins across the AEC industry if we do not all work together and accept responsibility for the industry's success.

How can we influence change?

By now you'll likely agree that it's time the AEC industry caught up with the rest of the world and embraced modern technology. But with many professionals in the business saying, "This is the way we've always done it...", how do you influence change? If your business is not even planning for digital transformation, how will you even start to talk these issues?

Unfortunately, this magic seems harder to bring about than machine learning itself. We suggest that it's up to the BIM believers to work together on the solution. There needs to be more collaboration even between those who are already forward-thinkers. We must keep reminding all levels in all sectors of the risks involved in lagging behind and the opportunities to be had in getting ahead.

What is machine learning going to impact?

Perhaps surprisingly, given the subject matter of this blog, we believe machine learning will not impact poor pre-construction planning for most companies in 2018. All AEC companies must agree to collaborate, stop blaming others for poor performance and start new ways of working. Companies that get that far (and we have started talking to a few) will speed ahead of their competition as they rapidly become more productive.

After this, the question will be: "What is machine learning not going to impact?". AI in pre-construction planning will positively impact productivity, the entire project lifecycle and the industry's reputation as a whole.

Kreo is an AI-powered platform for architecture, engineering and construction to automate Building Information Modelling. This will enable professionals to plan projects, design buildings and bid for work faster and more accurately than ever before. There are three main products within Kreo and they interact seamlessly to provide one Common Data Environment for everyone involved in a construction project throughout the entire pre-construction phase.

Kreo Software Ltd
3 Waterhouse Square, 138 Holborn, London, EC1N2SW
For all inquiries including press, contact:
Julia Valentine, Head of Marketing, jvalentine@kreo.net, 02086105010

This release was published on openPR.