Seeq expands machine learning support

Seeq Corporation has announced the expansion of its efforts to integrate machine learning algorithms into Seeq applications. The company says these improvements will enable organisations to operationalise their data science investments, and their open-source and third-party machine learning algorithms, for easy access by frontline employees.

Seeq’s strategy for enabling machine learning innovation is designed to provide end-user access to algorithms from a variety of sources, rather than forcing users to rely on a single machine-learning vendor or platform. This addresses the diversity and types of algorithms available to organisations, including:

• Open-sources algorithms and other public resources. For example, Seeq is publishing two Seeq add-ons to GitHub, including algorithms and workflows, for correlation and clustering analytics, which users can modify and improve based on their needs.
• Customer-developed algorithms in Seeq Data Lab — or machine-learning operations platforms such as Microsoft Azure Machine Learning, Amazon SageMaker, Anaconda and others — as part of data science or digital transformation initiatives.
• Third-party algorithms provided by software vendors, partners and academic institutions. AWS Lookout for Equipment, Microsoft Azure AutoML, BKO Services’ Pump Prediction and Brigham Young University’s open-source offerings are examples of the emerging marketplace for industry and vertical market-specific algorithms.
   

The Seeq initiative also address the critical ‘last mile’ challenge of scaling and deploying algorithms in a manufacturing organisation by putting data science innovation in the hands of plant employees in easy-to-use applications: Seeq Workbench for advanced analytics, Organizer for publishing insights and Seeq Data Lab for ad hoc Python scripting.

This is in addition to Seeq support for the foundational elements of success with machine learning. This includes access to all manufacturing data sources — historian, contextual and manufacturing applications — for data cleansing and modelling, support for employee collaboration and knowledge capture, quick iteration, and performance-based continuous improvement workflows.

“Data science innovation in manufacturing organisations has the potential to deliver a step change in plant sustainability, productivity and availability metrics,” said Kevin Prouty, VP Industrials, IDC Corporation. “But to land this opportunity, companies must be able to deploy data science innovation to frontline engineers with the expertise, data and plant context to make decisions on insights provided by these new algorithms.”

Examples of customers using Seeq applications to access and integrate data science innovation include an oil and gas company deploying a deep-learning-based emissions prediction algorithm, a pharmaceutical company using an unsupervised learning algorithm to proactively detect sensor drift in sensitive batch processes, and a chemical company using pattern learning to identify root causes of process instability and extend cycle time.

“Seeq provides a bridge between data science teams and their algorithms to frontline employees in hundreds of plants around the world,” said Brian Parsonnet, CTO at Seeq Corporation. “Deploying algorithms is now as simple as registering them in Seeq, and then defining which employees have access to each algorithm in their Seeq applications.”

Image credit: ©stock.adobe.com/au/zapp2photo