Machine Learning (ML) is a flavor of Artificial Intelligence (AI). This news release from Seeq illuminates a bit of the mystery surrounding much discussion of the topic.
Seeq Corporation released R52 with new features to support the use of machine learning innovation in process manufacturing organizations. These features enable organizations to deploy their own or third-party machine learning algorithms into the advanced analytics applications used by front line process engineers and subject matter experts, thus scaling the efforts of a single data scientist to many front-line OT employees.
New Seeq capabilities include Add-on Tools, Display Panes, and User-defined Functions, each of which extend Seeq’s predictive, diagnostic, and descriptive analytics. The result is faster development and deployment of easy-to-use algorithms and visualizations for process engineers. With R52, end users will also be able to schedule Seeq Data Lab notebooks to run in the background, fulfilling a top customer request.
Seeq customers include companies in the oil and gas, pharmaceutical, chemical, energy, mining, food and beverage, and other process industries. Investors in Seeq—which has raised over $100M to date—include Insight Ventures, Saudi Aramco Energy Ventures, Altira Group, Chevron Technology Ventures, Cisco Investments, and Next47, the venture group for Siemens.
As a compliment to the new extensibility features, Seeq data scientists are working with customers to develop and deploy machine learning algorithms tailored to the industrial process domain. Current areas of focus include automatically detecting performance changes in monitored assets, identifying causal relationships among process variables, and improved diagnostics by identifying and labeling patterns within a data set. For example, a super-major oil & gas company is using Seeq extensibility features to enable easy access by process engineers to a neural-network algorithm created by their data science team, helping reduce greenhouse gas emissions.
“Analytics software for manufacturing organizations is an area overdue for innovation,” says Steve Sliwa, CEO and Co-Founder of Seeq. “Spreadsheets replaced pen and paper 30 years ago for analytics and haven’t changed much since. By leveraging big data, machine learning and computer science innovations, Seeq is enabling a new generation of software-led insights.”
Seeq first shipped easy to use machine learning-enabled features in 2017 in Seeq Workbench, and then in 2020 introduced Seeq Data Lab for Python scripting and access to any machine learning algorithm. This support for multiple audiences—with no code/low code features for process engineers and a scripting environment for data scientists engaged in feature engineering and data reduction efforts—democratized access to machine learning innovation.
Seeq’s approach to integrating machine learning features in its applications addresses many of the reasons data science initiative fail in manufacturing organizations.
- Seeq connects to all underlying data sources—historian, contextual, manufacturing applications, or other data sources—for data cleansing and modeling.
- Seeq supports the connected, two-way, interaction of plant data and process engineering expertise in OT departments with the data science and algorithm expertise in IT departments.
- Seeq provides a complete solution for algorithm development, updating and improving algorithms over time, employee collaboration and knowledge capture, and publishing insights for faster decision making.
In addition to Seeq Data Lab support for machine learning code and libraries, Seeq also enables access to the Seeq/Python library by third-party machine learning solutions including Microsoft Azure Machine Learning, Amazon SageMaker, and open source offerings such as Apache Anaconda. For example, a manufacturer using Amazon SageMaker is evaluating their machine learning insights with Seeq to create work orders in their SAP system.