Born in the Mathematics Department of Princeton, FMI Technologies began as an attempt to provide a much more robust and transparent alternative to AI for time series forecasting and classification, anomaly detection, and correlation analysis. Today we aid hedge funds, demand forecasting and supply chain firms, as well as any one interested in time series and other structured datasets. Having partnered with Princeton, IAS, UC Berkeley and Stanford, we have used the sharpest minds at our disposal to combine tools from chaos theory, AI, field theory and control theory to create what is now known as Field Machine Intelligence (FMI). Our models train quickly and satisfy the accuracy bound. They are efficient at arbitrarily nonlinear, non-stationary and non-IID datasets and are built to account for sudden regime shifts.
We provide custom built APIs implementing FMI for quant-focused funds and discretionary traders alike. Ideally you would schedule a consultation where you would specify your input and desired output. We have provided to our clients, trading signals, strength and stability scores for nonlinear correlations, as well as fully fledged trading strategies. A sample output from one of our custom built APIs is displayed above- plotted are the strengths of various correlations/trading signals derived from a given dataset as they evolve in time. Clearly, the impact of regime shifts or market shocks on the couplings is immediately seen as a destabilization. If there is a market anomaly in your dataset given a fixed resolution and time horizon, our mathematicians have proved you will discover it. On the other hand, if it weakens or disappears, you will know almost immediately!
We are working on an automated API to allow the user interactive access to field theoretic tools. This is ideal for discretionary traders with minimal exposure to programming or statistics, wherein they can test out various hunches about the market on our interface. Email us at email@example.com with the subject line 'API' to stay posted about this. Meanwhile we would be glad to build a custom-made tool for you.
The above plot represents a three dimensional slice of a potentially infinite dimensional space of all trading strategies offered to our clients. Each straight line represents a trading strategy, on which the blue and yellow points represent the upper and lower. bounds of the returns to be earned from each strategy at a 90 percent confidence interval. the midpoint quantifies the average returns and the 'spread' quantifies the risk. As we place bounds on the risk and return of our desired strategy, our infinite dimensional space thins out and we are left with the choicest of market anomalies/trading strategies that provide the highest returns at the lowest risk.