Session 1.2
Automated Revit Documentation Via Machine Learning
Synopsis:
This class will look at an internal project GHD recently undertook to establish the viability of using machine learning for the automatic placement of tags using several machine learning algorithms. We will look into 2 different algorithms, convolutional neural network approach using tagged images, and a tabular approach for the semantic data of the tags. We’ll step through the problem state, how we captured the data and then the outcomes/accuracy of the algorithms that we tested.
Learning Objectives:
1. Learn different machine learning approaches
2. Strategies to capture Revit data for training ML algorithms
3. Comparison and analysis of different algorithms
Body:
This class will look at an internal project GHD recently undertook to establish the viability of using machine learning for the automatic placement of tags using several machine learning algorithms. We will look into 2 different algorithms, convolutional neural network approach using tagged images, and a tabular approach for the semantic data of the tags. We’ll step through the problem state, how we captured the data and then the outcomes/accuracy of the algorithms that we tested.