Available Technologies

VALUE PROPOSITION
In machine learning and related applications resource reduction allows for lower economic, energetic, and data costs, which in turn (1) increase the scale of sensing and connected system deployment and (2) allow for more applications to operate on a fixed resource budget. Machine learning applications, particularly for constrained systems such as Internet of Things and mobile devices or embedded computers, are limited in deployment or not developed due to the resource costs required for operation. Many machine learning applications have over-provision requirements for data quantity and quality, whereas lesser data are often sufficient.

DESCRIPTION OF TECHNOLOGY
Rather than indiscriminately accumulating large volumes of data, and/or ultra-high-fidelity data for Machine Learning/IoT applications, there exists a strategic and efficient approach to data capture, collection, transmission, and processing. For a given application or set of applications, it is possible to identify and capture the essential subset of data which holds the maximum information required and to optimize system design and operation to capture only the required data to meet application performance targets. For Machine learning, rather than focusing on higher quantity or quality of data, there exists a way to determine the certain part or subset of data that gives sufficient output to reach the optimal balance of performance and resource allocation and utilization.  This technology achieves that allocation.

BENEFITS
•    Visualize the general trend for information density across a data feature for a specific technical application
•    Determine minimum viable data for that specific technical application by identifying the inflection point
•    Automated inflection point detection and resource optimization

APPLICATIONS
•    Selection sensor types
•    Sampling rate
•    Data processing location
•    Transmission frequency

IP STATUS
Patent Pending

LICENSING RIGHTS AVAILABLE
Full licensing rights available

DEVELOPER: Dr. Josh Siegel and Tashfain Ahmed

TECH ID: TEC2024-0098

For more information about this technology, 
contact Raymond DeVito Ph. D. at devitora@msu.edu or 1(517)884-1658