DC-COMP Overview

DC Compression helps to produce models which are significantly smaller in both memory and computations requirement and run faster with minimal impact on accuracy

  • Highly accurate machine learning models rely on ‘millions’ to ‘billions’ of internal model parameters and must be trained over very long periods of time.
  • These qualities enable a much wider deployment of machine learning models including:
  • Real-time applications, like high-frequency trading algorithms credit decisions and fraud detection where speed is important
  • Deployment on smaller devices with limited computational resources.

Your algorithm can be right 100% of the time, but, if the result is received past the action point, then it was useless

“Your algorithm can be right 100% of the time, but, if the result is received past the action point, then it was useless.”

“DC Compression helps to produce models which are significantly smaller in both memory and computations requirement and run faster with minimal impact on accuracy.”

DC-COMP Platform

“DC Compression helps to produce models which are significantly smaller in both memory and computations requirement and run faster with minimal impact on accuracy”

DC-COMP Use Cases

DC-COMP Resource Library

Contact Us

Please leave your contact details below or email us at capitalmarketsdelivery@deltacapita.com