GTC 2020: Accelerating ML at the Edge, Real-Time Well Engineering
Vanessa Kemajou, Halliburton | Joseph Winston, Halliburton
Real-time machine learning enables quantitative analytics at the edge. To calculate the results in a satisfactory time window, both the hardware and software should be up to the task. Recent improvements in hardware, such as the NVIDIA Jetson Nano, allow the cost-effective deployment of field accelerators. Likewise, software, such as RAPIDS and TensorRT, helps with the extract, transform, and load process and optimizes the model, respectively. This methodology will be applied to Real-Time Well Engineering, a DecisionSpace 365 cloud application that processes and displays real-time drilling data for monitoring and predicting drilling conditions.