Welcome to the synthetic future
Hazy generates statistically controlled synthetic data that can fix class imbalance, unlock data innovation and help you predict the future.
Synthetic data to solve class imbalance
Class imbalanced data sets are a major pain point in financial data science, including areas like fraud modelling, credit risk and low frequency trading.
Hazy Balance uses a range of best-in-class techniques to optimally re-balance structured data sets. This boosts low-frequency signal so that models can be trained on better data, leading to better results.
Synthetic data for forecasting and simulation
Forecasting and simulations need to be based on a wide variety of data to be able to model and predict the future.
Hazy Transfer combines advanced transfer learning techniques with synthetic data generation to enable forecasting and simulation of future events and markets from incomplete data and previously unseen behaviour.
Privacy preserving synthetic data
Privacy and security concerns often block or slow down data provisioning due to data security, privacy, ownership and compliance issues.
Hazy Generate boils down data sets into a Hazy data generator. These can be used to generate statistically equivalent synthetic data sets that preserve utility without being subject to security and compliance concerns.
Remove personal data from unstructured text
Personal data is often found in unstructured text, like customer communications and free text fields.
Hazy Redact uses a combination of advanced AI techniques, rules based logic and contextual awareness to identify personal data (PII) in unstructured text. It can then replace the PII with consistent alternative values to generate masked or redacted data.
Hazy products are deployed as fully managed tactical tools, with command line, web service and browser-based interfaces. They can be used to quickly and automatically generate and provision data into production, test, sandpit, labs, cloud and external environments.
Next to your data, within your network
We are available to run on-premises for select enterprise customers and we are optimised for secure private cloud deployment, where our technology runs on resources that are dynamically provisioned within your cloud.
In a on-premise environment, we install and package our software onto virtual machine (VM) instances to run within your private network and data centre.
In a cloud environment like AWS and Microsoft Azure, we dynamically deploy model training and data generation virtual machine (VM) instances within your private cloud.