The annual KDD conference is the premier interdisciplinary conference bringing together researchers and practitioners from data science, data mining, knowledge discovery, large-scale data analytics, and big data.
As part of KDD 2020’s virtual conference, we demonstrated visual tools and metrics using a pre-trained machine learning model to help others better understand how well calibrated their classification models are. During this tutorial, we also demonstrated ways to calibrate the model after training - from theory to a Keras implementation. For more information, check out our KDD website or watch the tutorial here.
A two-year joint venture between nPlan (lead), University of Cambridge and Kier.
We have teamed up to create an AI model to recommend optimised pathways for schedule execution based on data of thousands of previous projects. Using nPlan’s existing work we will be creating algorithms to predict project risk. The model will allow Project Managers of Construction Sites to identify the risks within their projects so that they can make better informed decisions on where they should focus their efforts.
A 12 month joint venture between nPlan (lead) and Atkins.
We are delivering a programme to further advance and implement our existing technology to determine the most likely duration of construction projects undertaken by Atkins. With this knowledge, we hope we can help highlight future contractual incentives.
A two-year joint venture between Skanska (lead), nPlan, BRE, Vinci and Assentian.
We are creating a user friendly supply chain management tool that can be used in the planning of construction projects. Using this tool we hope will resort to better project planning, improve collaboration across the supply chain and a cause in developing a business model for commercial use.