Jun 03 2020
3:30 pm - 4:30 pm
AEC – Advanced Approaches – AI & ML Approaches
Track Name: Wednesday 3:30 PM - 4:30 PM
Session Date: Jun 3 2020 3:30 pm - 4:30 pm
Achievable Artificial Intelligence for AECO - A Roadmap to Unlocking the Power of your BIM Data
3:30 PM - 3:45 PM
When most of us hear the term 'Artificial intelligence' we tend to think of the kind of science fiction AI that can perform complex calculations and feats of logic. In design and construction is still seen as years away from being a reality in enhancing or disrupting current technologies and processes even with recent advances in processing power. There is a common misconception that AI is a system, but it is not a system. AI is implemented in the system. There are also different types of AI and subsets of AI such as machine learning and computer vision, which we'll expand on. In this session we will focus on Artificial Narrow Intelligence (ANI) in existence today and is something most of us interact with on a daily basis. Think of Google Assistant, Google Translate, Siri or Alexa. They are all machine intelligence that use Natural Language Processing (NLP). Now, without data there is no AI and even with data you need to mine it, classify and structure it to make it meaningful, so how does this apply to AEC? Think about it, architects and engineers already use past construction, design, and building data to tackle new projects. The ability to utilize huge amounts of previous data in a second to enhance the design process will be a game changer. AI doesn't just automate, rather, it analyzes data and generates new ideas or solutions. Through several real case studies of recently completed pilot projects, we'll demonstrate techniques for data collection and mining from tools like Revit using its Journals, Dynamo and the Revit DB link tool and what useful features were used in the machine learning pipeline including cleaning and normalizing the data and performing prediction using different ML models including neural networks. You will see applications for evaluating and measuring efficiency and productivity at a user and project level and be able to predict behavior. Also, we'll explore different approaches to using your existing BIM data for creating predictive tools using the Revit API to streamline common modeling and documentation tasks. The possibilities of applying machine learning to BIM are countless. Classification algorithms, anomaly detection, and even time series analysis can be used with BIM. Both repeatable and predictable activities can be replaced by machine learning tools. We'll look at a few examples of how AI can benefit construction in areas such as commissioning, safety and construction scheduling and finally look at operations and asset management with practical uses of smart sensors and IoT platforms to drive a prescriptive maintenance approach reducing human intervention. To conclude, we'll demonstrate a class of machine learning for mapping architectural and engineering knowledge in assisting us create viable design options through exploring the concept of generative adversarial networks (GAN) and how it improves on the drawbacks of parametric and rules based computational design. Overall, get a breakdown on developing a roadmap and strategy to make sense of your BIM data to deploy a measure of real artificial intelligence in your workflow that yields tangible benefits and get a sneak peek at what's ahead.
Get Rich: The Right Data for a Successful AI Strategy
3:45 PM - 4:00 PM
Construction industry market trends are pointing to increased use of data, analytics and artificial intelligence-based solutions to automate many processes in the construction life cycle. The AEC community, as well as owners, are increasingly looking at AI to improve their competitive edge and margins. But the promise of AI is entirely dependent upon data. And when most of that data is distributed across multiple systems, with manual or non-existent workflows and siloed in systems that don't "talk" to each other, the data becomes stale and disjointed. In order to deliver results with AI, the data strategy requires better architecture and augmenting internal data with enriched, continuously updated data. Additionally, any artificial intelligence involved in making decisions generally made by humans, will need to understand construction "language". Attend this session to discover how a combination of deep learning, natural language processing, data science, construction vocabulary, and, more importantly, "rich" data, and artificial intelligence can deliver on the promise of the future!
The Rapidly Emerging Role of Artificial Intelligence in Infrastructure Solutions
4:00 PM - 4:15 PM
The industry is leaning on emerging technologies such artificial intelligence and most specfically machine learning for a range of aplications bringing greater insights and effiiencies to our infrastructure solutions. We are also finding new and exciting use cases that help support rapid analysis and knowledge capture. This presentation will highlight a few State DOT case studies from significant transportation initiatives leveraging AI/ML for many benefits.
Audience Q & A Panel Discussion
4:15 PM - 4:30 PM