The Laplace Platform 2.0

Identify Causes & Effects

For every report, we train the machine to validate, stress-test and identify new statistically significant relationships between assets and indicators

Forecast Risks & Returns

Analyzes thousands of financial signals to recognize the current market regime and likely future scenarios in order to deliver high quality forecasts

Optimize Asset Allocations

Simultaneously analyzes correlations, expected volatility and return to select & suggest asset allocations that deliver better performance at lower risk

Adapts to Any Situations

Based on published academic research, our models incorporate many types of data streams and have been validated to perform in thousands of extreme market scenarios. Implemented on Amazon Web Services (AWS) and a Neurodigital custom AI GPU server farm, our platform routinely discovers new predictive, associative and statistical relationships within the data as new information comes in every month. 

The architecture easily incorporates new data streams, models and features, and can scale to operate in many different regimes, making the platform highly adaptable to new and unpredictable situations.  

Chart Indicating Multiple Levels of Input that Laplace Insights Considers

Performs in Extreme Scenarios

Our platform has been built using using some of the most advanced time series forecasting available in AI research. Adopting engineering techniques used to deliver very low defect rates in chip design or pharmaceutical production gives us confidence that the machine will perform in a very large range of extreme market scenarios. 

We use proprietary stress- and back-testing techniques to ensure that the algorithms do not break under all imagined corner cases, and are constantly imagining new “test vectors” to ensure that the machine becomes more robust over time. 

Learns From History

Our platform analyzes over 1000 indicators using many types of models trained on over 100,000 financial events over the past 120 years.  We train our algorithms on data from a wide range of historical shocks, such as the two World Wars, the 1929 stock market crash, the Great Depression, the 1970s stagflation period, the Dot-Com Crash and the Financial Crisis, to understand patterns driving volatility and return. 

Concurrently analyzing behavioral, fundamental and statistical patterns allows us to develop new hypotheses around causal relationships, while machine learning techniques only elevate these causalities when they become statistically significant. Learning from past financial market accidents at scale gives us confidence in our ability to perform in many extreme scenarios.

In The News