Machine Learning for Industrial Decarbonization
What to expect?
Decarbonising heavy industries is a key part of mitigating climate change. Carbon Re is an AI for materials company on a mission to reduce carbon emissions by gigatonnes every year through the use of AI technology. Carbon Re is focused on decarbonizing cement and other foundational materials (such as steel and glass) which are responsible for more than 20% of global emissions.
In this talk, we’ll explore some of the machine learning challenges faced by Carbon Re, from grappling with irregularly sampled time series and ever-shifting distributions to dealing with unreliable sensors. We’ll see a demo of Carbon Re’s Delta Zero platform, which is empowering cement plant operators to make more informed decisions and reduce the environmental impact of cement. We’ll also explore how W&B’s Model Registry enabled the team to easily implement an automatic retraining system and streamline their process.