The first principle design approach is the critical thesis that guided the protocol structure of our architecture. DeepMoneys core competence is deploying best practice AI systems which possess the innate ability to generalise intelligence and evolve through learning. 

We did not start with a performance yield target trading a particular instrument in a fixed asset class; we started with questions.

Can we build one common architecture with the capabilities to:

1. Support not just one instrument, but thousands, such as cryptos, without tuning individual algorithms?
   
   

2. Operate in different asset classes from FOREX to stocks to commodities to crypto?
   
   

3. Trade assets equally on different timeframes to capture profits?
   
   

4. Trade multiple markets whether it is Binance, Nasdaq or Interactive Brokers?

Fortunately the answers to all four questions is yes and furthermore, the implications of architecting DeepMoney AI systems in such a manner also resonate with:

• Risk-Management

• Compliance

• Regulation

Engineering excellence at DeepMoney is tightly bound to commercial reality and the requirements of individual investors and institutions. Such a universal architectural approach means the operational cost of computing is ring fenced. The benefits of this novel design is elaborated in further detail this section: Low power High value Reinforced Learning 

Directly coupled to scaling the deployment with virtually zero incremental costs is the ability to scale into new financial instruments and different asset classes. Critically the capital and risk-management policies are unified.

Diversity with the universal architecture

It may be tempting to classify DeepMoney technology under one AI pattern (use case) as ‘autonomous’, however to achieve success in that complex field for financial markets we had to invoke upon our overall architecture the capability to seamlessly integrate nearly all of the AI patterns.

AI can be decomposed into seven universal patterns (use cases): some AI systems use a single pattern for their application while others combine a few together. DeepMoney artificial intelligence systems range across the patterns in an impressive manner with solutions for most of the verticals in the financial services industry.

1. Hypersonalization

Develops a unique profile of each individual, displaying relevant content, providing personal assistance on a wide range of life issues, like entertainment, healthcare, finance.

AI Signals CoPilot understands its clients assets and can offer individual expert assistance.

2. Human Interaction

Where machines and humans communicate across a variety of methods including voice, text, and image forms. AI Signals CoPilot will interact directly with individual clients and even go so far as to track a particular trade they made and give advice on its progress.

3. Pattern Detection

Identifying patterns used in law enforcement to detect financial fraud or DeepMoney machines detecting market moves or recognising past price action.

4. Recognition

Identify and determine objects or other data points within image, video, audio, text, or other primarily unstructured data. This type of AI is used in facial recognition or healthcare skin cancer diagnoses for example.

5. Goal Driven Systems

The ability to learn through trial and error like bidding real time auctions, IBM’s. Deep Blue chess and DeepMoney machines are rewarded to maximise PnL and risk manage a client's capital allocation.

6. Predictive Analytics

A common example is weather forecasting, by using data to assess past forecasts and improve predictions over time. DeepMoneys unique approach is to use a combination of AI disciplines to predict market prices including the component of memory.

7. Autonomous

To reach a goal without human interaction ranks among the most challenging of the AI patterns. Autonomous AI systems such as vehicles or financial markets are roads well travelled without any panacea as of yet. Wealth machines by DeepMoney trading non-static environments with live market data is the apex.
Citation: https://mneguidelines.oecd.org/RBC-and-artificial-intelligence.pdf