AFTAs 2022: Most innovative third-party technology vendor (AI and machine learning)—CompatibL

Project: Autoencoder Market Models (AEMMs)


The inclusion of multifactor Autoencoder* Market Models (AEMMs) in the latest release of the CompatibL Platform v8.0 represents a new capability in addition to the range of conventional interest rate models available within the software. Users can choose AEMMs for all calculations within the software, from valuations and XVA in a risk-neutral measure to counterparty credit risk and limits in a real-world measure.

* An autoencoder is a form of artificial neural network designed to learn data encodings in an unsupervised manner.


The philosophy driving CompatibL’s products is continuous development without losing sight of present market requirements/drivers. As part of the firm’s R&D initiative, it encourages new ideas in order to respond to current market conditions and meet its clients’ needs. This includes discovering, inventing and developing new features and services to address their growing demands and evolving expectations.

“With the unprecedented upheaval of financial markets caused by recent world events, portfolio valuation and risk require forward-looking scenarios that must cover all the possible market regimes. We believe Autoencoder Market Models can succeed in representing these diverse risk scenarios required by today’s market landscape and envision a future where reduced reliance on conventional models, as well as an increased automation of interest-rate modeling, can be adopted on a broader scale.”
Alexander Sokol, founder and executive chairman, CompatibL 

Examples of innovation

  • AEMMs: A new type of interest rate model based on autoencoders, used to optimally represent historical yield curves and volatility surface shapes.
  • Autoencoders for the visualization of market data regimes: Generated maps that illuminate the interest rate regimes covered by historical time-series for each currency.
  • Market generators: Machine-learning algorithms for generating realistic samples of market data when historical time-series have insufficient length or gaps.

Future developments

CompatibL was the first vendor to deploy autoencoders for interest rate models in a production environment, underlining its commitment to innovation. The firm is actively working to extend AEMMs to other asset classes with high-dimensional market data represented by curves and surfaces, including credit and commodities. According to Alexander Sokol, founder and executive chairman of CompatibL, this functionality will be released sometime next year.   

Why they won

Innovation is at the heart of what CompatibL does: introducing new methods of addressing complex problems for its capital markets clients. This win follows last year’s success in the best new technology introduced over the last 12 months (trading and risk) category in these awards, thanks to the release of its machine learning-based market generators underpinning the firm’s credit risk model. Given CompatibL’s propensity for pushing the AI-based risk model envelope, who would bet against it picking up another category in these awards next year?    

Only users who have a paid subscription or are part of a corporate subscription are able to print or copy content.

To access these options, along with all other subscription benefits, please contact or view our subscription options here:

You are currently unable to copy this content. Please contact to find out more.

You need to sign in to use this feature. If you don’t have a WatersTechnology account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an individual account here