K2view Launches Synthetic Data Management Solution, Merging AI and Rule-Based Methods with Business Entity Data Modeling, to Achieve Unparalleled Data Accuracy and Compliance
K2view, a global leader in operational data management, announced the market’s first end-to-end synthetic data management solution. This comprehensive offering uniquely combines generative AI and rule-based synthetic data generation methods, with a patented business-entity data model approach, to create synthetic data of unparalleled accuracy and compliance.
K2view Synthetic Data Management addresses multiple use cases, including software testing, machine learning (ML) model training, and data collaboration and sharing.
“Through the integration of 4 data generation techniques – generative AI, rule-based generation, entity cloning, and data masking – our solution can support every use case that requires structured and semi-structured synthetic data,” explained Yuval Perlov, Chief Technology Officer at K2view. “Moreover, our solution uniquely fuses business-entity data modeling into the data synthesis process, to achieve unrivaled synthetic data accuracy and ensure referential integrity of the generated data.”
Generate realistic, compliant, and complete data K2view achieves unparalleled data accuracy across any use case by leveraging three key capabilities:
- K2view uniquely generates synthetic data by business entities (such as customer, device, order, etc.), ensuring that all the required data for each business entity is consistent and contextually accurate. A business entity data model is automatically discovered and classified from the source systems, and serves as the blueprint for generating fake data, regardless of the synthetic data generation method.
- The solution integrates 4 data generation methods: (1) Generative AI GPT models to create realistic synthetic data based on a training dataset, (2) rule-based generation, (3) entity cloning, which extracts, masks and duplicates data based on business entities, and (4) data masking.
- Precision subsetting of multi-source data feeds the generative AI algorithms with the most relevant training data, to maximize accuracy and performance.
Kathy Lange, Research Director AI Software at IDC, commented “The business entity method allows large companies to build synthetic data that accurately represents complicated relationships and hierarchies across numerous data sources. It provides a comprehensive framework for producing synthetic data while adhering to high accuracy and compliance standards.”
Complete solution supporting the entire synthetic data lifecycle
K2view Synthetic Data Management streamlines every stage of the synthetic data lifecycle, including:
- Prepare: Extract and subset production data needed for model training from all sources, regardless of underlying technology, auto-discover data structures, formats and valid values, auto-classify data types, and mask data to ensure compliance.
- Generate: Apply any combination of four generation methods to generate accurate and compliant data.
- Transform: Apply necessary data transformations on the generated data.
- Validate: Ensure data verification and enforce referential integrity across source systems.
- Provision: Instantly deliver compliant synthetic data to target datastores.
- Monitor: Track the end-to-end process of preparing, generating, and provisioning the data.
Agility and automation
The solution empowers data consumers, such as testers and data scientists, with self-service tools to control and manage the data generation process, including the ability to subset the training data, set business-rule parameters, and iterate the synthetic data generation process. It also supports synthetic data reservation, dataset version management, and instant roll-back to prior versions.
Additionally, the solution seamlessly integrates into testing CI/CD and ML pipelines via standard APIs, enabling programmatic control of all phases of the synthetic data management cycle.
[To share your insights with us, please write to firstname.lastname@example.org]