Transprecise Computing

The following are current and recent funded projects on the topic of transprecise computing.

SWEET: Sustainable Wearable Edge Intelligence
Building on some of the ideas of the DiPET project, we turn to the problem of serving machine learned models in distributed environments, including wearable devices, in particular considering variability of resources such as power and network bandwidth or strength.

DiPET: Distributed Stream Processing on Fog and Edge via Transprecise Computing
We investigate models and techniques that enable distributed stream processing applications to seamlessly span and redistribute across fog and edge computing systems. The goal is to utilise devices dispersed through the network that are geographically closer to users to reduce network latency and to increase the available network bandwidth.

Transprecise Multi-Physics Simulation
A PhD project in the context of the Doctoral Training Programme on Secure Connected Intelligent Design and Manufacturing (SCIDM)

RAPID: ReAl-time Process ModellIng and Diagnostics: Powering Digital Factories
Transprecise analytics for real-time analytics in manufacturing.

SoftNum: Software-Defined Number Formats
Revisiting computer arithmetic to maximise the benefits of transprecision for applications in machine learning, cryptography and data analytics.