“MASTER” : An intuitive green computing active learning system on appliances
“MASTER” is a software suite running on our Sm_All appliances to allow industries with limited bandwidth or without a proper cloud connectivity to run the active learning system locally.
All the AI detections, AI classifications, AI segmentations happen on green computing appliances with AI capabilities and benefit from a continuous automatic active learning system.
To make it possible, Sm_All works automatically in combination with IBM Maximo Visual Inspection®.
All the process takes place automatically except the annotations process, to enhance drastically the accuracy of the embedded AI.
Suitable for : Industry 4.0, E-Health, Transports, Smart Cities, Public Cleanliness, Security, …
HIGHLIGHTS
Allow people who do not master AI-based technologies to perform visual inspection with active learning in a virtuous process on appliances
Identify potential quality and performance issues in real time on appliances and configure to notify key stakeholders immediately
Integrate with maintenance and quality workflows for a fast and prescriptive response at an appliance perspective
Enable an intuitive and positive feedback loop that improves model accuracy
A continously improved accuracy of your detections thanks to an easy active learning process
The power of the cloud on a green computing circuit board.
The IBM Maximo Visual Inspection® annotation and training features are directly available on our devices.
The annotator (a person without any specific skills) from your company has access to a very simple tool to let know to the AI what is needed to be detected via labels editing.
The system knows automatically once the annotations are done then Sm_All launches the AI (re)training on IBM Maximo Visual Inspection®.
IBM Maximo Visual Inspection® annotation example
The active learning process with “MASTER”
-
Take pictures from the builtin camera or the camera available (IP) with unknown or undetected objects on it
* on Sm_All -
Collect as much pictures as required (settings on the device)
* on Sm_All -
Wait for the right moment (settings on the device) to upload the batch of pictures to the cloud
* on Sm_All -
Send an email to the right person (settings on the device) with a link on the device to annotate pictures on IBM Maximo
* on Sm_All -
Annotate pictures on IBM Maximo
** on IBM Maximo -
AI training on IBM Maximo once the annotations are done
-
Sm_All makes sure the most accurate model is pulled from the IBM Maximo or keep the same model available locally.
*on Sm_All
-
Given the brand new model, make a better inference with new improved detections and avoid false detections.
*on Sm_All
“MASTER” will be released officially at CES LAS VEGAS in January 2023 and showcased on February 14th at SIRRIS Research Center in Seraing (Belgium) on a robot.
Available through our growing network of integrators and TECH DATA (TD SYNNEX) network.