ScaleOut Enhances Digital Twin Intelligence With Generative AI and ML

Source: Shutterstock

As organizations juggle the complexity of real-time systems, they’re under increasing pressure to stay ahead by identifying issues and responding to them before they can disrupt operations. 

However, traditional monitoring tools often fall short, especially for systems that generate vast amounts of streaming data from various data sources. The real-time monitoring inefficiencies lead to delayed anomaly detection, high manual workload, and static models. 

ScaleOut Software, a company specializing in in-memory computing solutions for enhanced operational intelligence, aims to overcome some of these challenges by adding GenAI and automatic machine learning (ML) retraining capabilities to its platform. The newly released Version 4 of the ScaleOut Digital Twins platform allows operators to use GenAI and ML to quickly identify and address emergency issues while reducing their workload. 

Digital twins refer to virtual replicas of real-world systems that use real-time data to monitor, analyze, and optimize operations in real time. The new version of the platform, with advanced AI and ML features, makes these digital twins smarter and more helpful. 

Retraining ML models dynamically improves accuracy without disrupting operations. ScaleOut’s Version 4 adds automatic retraining for ML algorithms running within digital twins, continuously improving their monitoring capabilities as they process new telemetry data. The platform can now identify spikes, trends, and unusual patterns across historical data streams. 

According to ScaleOut, integrating AI technologies enables organizations to monitor and respond to complex system dynamics and uncover insights that might otherwise go unnoticed. 

“ScaleOut Digital Twins Version 4 marks a pivotal step in harnessing AI and machine learning for real-time operational intelligence,” said Dr. William Bain, CEO and founder of ScaleOut Software.

“By integrating these technologies, we’re transforming how organizations monitor and respond to complex system dynamics — making it faster and easier to uncover insights that would otherwise go unnoticed. This release is about more than just new features; it’s about redefining what’s possible in large-scale, real-time monitoring and predictive modeling.”

The new capabilities are a step forward toward autonomous operations. It pushes real-time monitoring to a level where these systems can analyze data, detect anomalies, and take proactive actions with minimal human intervention. 

Large and complex systems exist in several industries, and ScaleOut’s Version 4 might be able to better handle the requirements of such systems. Potential use cases include security systems, transportation networks, power grids, military asset tracking, and smart cities, according to the company. 

Along with automatic anomaly detection with GenAI, Version 4 also features natural language data exploration. Instead of writing complex queries, users can interact with the plant in plain language. This is particularly valuable for non-technical team members who need access to data insights. 

The platform now works with both TensorFlow and ML.NET, giving users more options for running machine learning models. ScaleOut claims the platform can handle large-scale tasks, processing over 100,000 messages per second across millions of digital twins. Additionally, faster data sharing through an in-memory grid makes it easier for digital twins to work together. 

Source: ScaleOut

ScaleOut’s open-source APIs allow developers to create digital twin models for real-time monitoring and simulation on the ScaleOut Digital Twins platform. To simplify development, the platform includes an open-source workbench where applications can be tested before deploying them at scale. 

Dr. Bain shared with BigDataWire that “the combination of digital twins, ML, and GenAI helps make real-time monitoring more reliable and autonomous. This technology improves the odds that problems are detected and addressed effectively”.

Elaborating on the core technology behind ScaleOut’s platform, Dr. Bain explained that “the platform uses a technology called in-memory computing that enables it to process incoming messages within a few milliseconds and aggregate data every few seconds, while analyzing thousands or even millions of data streams. This allows it to monitor very large systems with many data sources generating continuous telemetry”.

If ScaleOut can effectively utilize its AI and ML advancements, it can help organizations monitor and manage complex systems and reduce some of the more persistent operational challenges. However, ScaleOut faces key challenges in ensuring GenAI remains accurate and grounded in real-time data while integrating with continuous ML retraining. Dr. Bain shared that to overcome this challenge, ScaleOut ensures “ that responses are factually based on real-time digital twin data and constrains them using structured data outputs.”

Source: ScaleOut

Dr. Bain emphasized that processing vast amounts of telemetry data directly with GenAI is impractical. “To address this, we aggregate data to extract key insights while maintaining accuracy,” he explained. “We’ve also been focused on designing and refining prompts to ensure generative AI effectively detects anomalies in the aggregated data.” 

He further highlighted the importance of real-time validation mechanisms in the continuous retraining of ML algorithms. “These mechanisms allow us to evaluate ML responses in real-time, generating high-quality supplemental training data while preventing issues like model drift or degraded performance.”

Related Items 

Dataiku Launches GenAI Real-Time Cost Monitoring Solution

Real-Time Analytics to Conquer the AdTech Data Deluge

To Improve Data Availability, Think ‘Right-Time’ Not ‘Real-Time’