AI & Machine Learning
What is AI and Machine Learning in Construction
Artificial Intelligence (AI) is no longer a futuristic fantasy. It’s happening now, and many in the construction industry are already reaping its benefits. Forward-thinking GCs have the opportunity to get ahead of the trend now and use AI with tools that are currently on the market to increase performance, maximize profits, and improve safety. So, what is machine learning in construction? Here’s what you need to know.
What Is Machine Learning in Construction?:
AI, also called machine learning, has been portrayed in movies and other media so often that it’s gathered quite a mythology about it. From fears that it will take over the world, to curiosity about whether an AI can be sentient, it’s easy to get caught up in a fantasy land of imagination.
To understand how AI impacts construction, we must first get a good understanding of what AI actually is–not the science fiction, but the reality. In general, AI is a broad topic that involves computer science, psychology, and even philosophy and linguistics. However, in regard to how it impacts construction, when we talk about AI we primarily refer to two specific areas: Machine learning and deep learning.
A visual representation of the differences between A.I., Machine Learning, and Deep Learning
Machine learning involves algorithms that allow computers to learn from data without being explicitly programmed. For instance, an AI algorithm can be “trained” to identify spam by exposing it to large quantities of emails that have been manually tagged as spam or not-spam. The algorithm “learns” to identify patterns that help it to identify spam “intelligently.”
Deep learning is a specialized form of machine learning that is based on neural networks. It’s a more recent development that has allowed breakthroughs to be made in image and language processing, opening the door for advanced applications.
AI and Machine Learning Applications in Construction:
The potential applications of AI and Machine Learning in construction are vast. It’s beyond the scope of any single article to make an exhaustive list of all the possible uses. To understand just how vast the field is, consider that a few years ago you probably received many spam emails in your inbox every day, and today you probably receive very few. That is because spam filters now use machine learning to identify patterns and keep spam out, and they’re very good at it.
While that isn’t an application specific to construction, it does impact all of us, making us more productive and able to focus on our work-related communications. And that’s just one tiny, well-established use.
A typical construction project can have thousands of open issues, hundreds of RFIs, and numerous change orders that are open on any given day. Imagine a smart assistant who can analyze this mountain of project data and alert you about the top 10 critical things that need your attention today? Machine learning is that smart assistant, helping teams identify the most critical risk factors from a construction safety and quality perspective that need immediate attention.
Machine learning apps, like Smartvid.io, can grab an image of a worker stepping off a ladder and add related tags.
Superintendents and Project Engineers often characterize their job as putting out fires. It is very reactive. Machine learning is fast becoming an assistive tool that proactively identifies risks and help them make decisions before it impacts their project.
A second application that is poised to have a huge impact on construction is image tagging and analysis. We already see powerful AI at work in social media, where algorithms identify facial patterns to automatically tag individual people with astonishing accuracy.
The same AI technology, with new training, can be used to identify and analyze safety hazards, categorize and tag site photography, and send notifications when PPE is not being properly used on the job site. It can even be used to identify who is violating safety standards, and tag them and/or their supervisors to address the problem.
Other applications may include sorting notifications, identifying potential issues such as conflicts or missing materials, tagging and organizing documents, and even piloting drones, running machinery, and assisting in design. As you can see, the possibilities are wide-ranging.
Current Uses of Machine Learning in Construction:
Machine learning is already being used in a variety of ways, from mundane spam filtering to advanced safety monitoring. Technologies already exist and are in use by innovative companies to tag visual data and analyze it for safety violations, potential hazards, and to mitigate all kinds of risks. Among other applications, current machine learning technology can be used to:
- Predict and mitigate risks before they impact project margins
- Identify high risk issues and automatically classify them into actionable categories
- Identify high risk subcontractors based on real-time data as well as past performance and other factors
- Identify and prioritize potential safety concerns across the project lifecycle
- Tag existing safety hazards based on visual data coming out of the job site
How to Benefit from Machine Learning on Your Projects:
As we’ve seen, in some ways you are already benefiting from machine learning, in the form of spam filters and other technologies that are already running inside the programs and technologies you use every day. But in order to stay ahead of the curve and gain competitive advantage from machine learning, construction companies must be proactive in understanding and implementing it on their job sites.
Algorithms in machine learning applications can differentiate and assess objects in an image
As with any new technology, it’s important to apply basic good project management to any implementation. When it comes to machine learning, there are a few additional best practices to keep in mind.
- Digitize workflows and documentation. That’s step one, for tools like machine learning to provide value it needs data to act on. If you are not capturing your quality inspections and safety observations in a digital form, then you are missing out on learning from those observations and making targeted improvements.
- Start with clean data. When it comes to introducing machine learning on your job site, remember: “bad input = bad output.” Make sure that your construction documents, issues, and visual data is organized and clean so you don’t teach bad habits to your intelligent technology.
- Choose the right data platform. There are a number of technology vendors providing solutions to manage data, but very often they are not compatible with each other. Machine learning and AI are most effective when they have access to all of the data across all of your platforms, and that’s only possible if the platforms integrate with each other. Look for a core platform like BIM 360 that allows you to manage all your construction data in one place, and that permits third party integrations for things like ERP and project management.
- Piloting for early insights. Start small, find a few ideal projects and project teams, define what success looks like and pilot machine learning based tools. This will help you and teams gain benefit on a day to basis as well develop best practices before rolling out more broadly.
Many of today’s leading construction companies already employ a number of construction-specific machine learning applications on the job site. Meanwhile, many of the industry’s leading software companies are developing new and valuable applications to add to the many current uses. If you’d like to know more about how to take advantage of the machine learning revolution on your job sites, we invite you to join our free webinar on The Rise of AI in Construction.
SOURCE: Manu Venugopal