Case Study : End-Point Training and MicroAI™ ATOM | ONETech
20812
post-template-default,single,single-post,postid-20812,single-format-standard,bridge-core-1.0.4,mega-menu-top-navigation,ajax_fade,page_not_loaded,,qode_grid_1400,qode-content-sidebar-responsive,wpb-js-composer js-comp-ver-5.7,vc_responsive
 
Case Studies

End-Point Training and MicroAI AtomML™

End-Point Training and MicroAI™ ATOM

End-Point Training and MicroAI AtomML™

The Industry Challenge

As IoT technology develops, and AI-based analytics become more prevalent, companies will have to balance the benefits of AI Solutions with the cost of data that traditional AI approaches require. High data costs are incurred because the standard training process for AI requires powerful hardware that is typically cloud based.

A potential work-around to this problem is edge-based AI. To train AI on the edge still requires a powerful, expensive, server to host the training process. While this solution eliminates the cost of sending massive amounts of data to the cloud, it has additional hardware costs.

The Solution

To avoid the high-priced hardware requirements of edge AI and the data costs of cloud-based AI, MicroAI has developed MicroAI AtomML™.  MicroAI AtomML™ is an end-point-based AI solution. This means that the AI lives entirely on the IoT device collecting the data. These devices are typically less expensive and are generally needed for both edge and cloud-based solutions.

Utilizing MicroAI AtomML™ for AI-enabled end-point training of IoT devices requires a few basic steps:https://www.onetech.ai/en/ourproducts/microai-atom

  • Select a MicroAI AtomML™-supported IoT device for end-point data collection
  • Ingest data to train the AI model
  • Activate the AI model
  • Send the reduced output data to the recipients and locations specified by your specific workflows

A typical MicroAI™ ATOM process flow would consist of the following steps:

micro-ai-diagram

The Impact

MicroAI’s MicroAI AtomML™ allows companies to employ AI solutions without the additional cost of high-volume data transfer to the cloud or expensive hardware. Removing these costs reduces the barriers to entry for companies attempting to implement cost-effective AI solutions. Additionally, end-point devices are ideal locations for feedback applications based on the results provided by MicroAI AtomML™. As the end-point device monitors the company asset, the device can signal the asset to change behavior based on the data provided by MicroAI AtomML™  Ultimately, MicroAI AtomML™ is a decentralized AI solution that is capable of living on small, non-expensive devices.