Advanced Computing in the Age of AI|Tuesday, August 4, 2020
  • Subscribe to EnterpriseAI Weekly Updates:  Subscribe by email

AutoML Startup dotData Supports Container-Based Models 

As data science platforms expand across enterprise applications like predictive analytics, automated machine learning vendors are steadily integrating AI models with emerging infrastructure to ease deployment and orchestration.

For example, data science automation specialist dotData this week released a container-based machine learning model aimed at real-time prediction. Applications include automated loan processing, dynamic pricing, fraud detection and industrial Internet of Things deployments such as a smart manufacturing partnership also announced this week.

The Stream platform is designed to deliver real-time prediction using dotData’s AI and machine learning models. Those models are downloaded from the company’s flagship platform via a one-click process akin to launching a Docker application container.

Deployment options include running the real-time predictor in cloud-based MLOps platforms designed to handle enterprise AI orchestration tasks. It can also run on edge servers for IoT applications, the company said Tuesday (July 7).

"We are seeing an increasing demand for real-time prediction capability," said Ryohei Fujimaki, dotData’s founder and CEO. The company, based in San Mateo, Calif., also disclosed a partnership with JFE Steel Corp. of Japan to deploy the Stream platform in its manufacturing facilities.

Along with automating AI development workflows, dotData’s MLOps tools are promoted as making machine learning models operational. The capability includes producing machine learning pipelines in production that include feature development and ML scoring used to predict future values derived from data.

The process of maintaining those pipelines is automated “as data changes over time,” the company said.

Japanese computer giant NEC Corp. spun off Fujimaki’s team to form dotData in April 2018 as enterprises began shifting away from heavy reliance of “black box” deep learning to explainable AI approaches. The spinoff’s “white box” framework seeks to provide more transparency on how predictions are made while automating the machine learning pipeline. Those tools include data collection and model training along with model selection and, now, easing the deployment of models in production via containers.

Along with its latest smart manufacturing partnership with JFE Steel, dotData’s software has been adopted by companies in the financial services, telecommunications, retail, pharmaceutical, transportation and airline industries, including Japan Airlines (OTCMKTS: JAPSY).

 

About the author: George Leopold

George Leopold has written about science and technology for more than 30 years, focusing on electronics and aerospace technology. He previously served as executive editor of Electronic Engineering Times. Leopold is the author of "Calculated Risk: The Supersonic Life and Times of Gus Grissom" (Purdue University Press, 2016).

Add a Comment

Do NOT follow this link or you will be banned from the site!
Share This