Our Technology

alphaAI is a technology-focused company. We believe in making the best possible technology available to our clients. Most of our efforts are focused on technology research and development. Our strategies and solutions are powered by an innovative and proprietary AI system. That being said, there are many misconceptions out there about what AI is. In this piece, we will dispel some common myths and shed light on the topic.

‍

BTW: This article is an intro to AI. If you already have a basic grasp of AI, dive deeper here. And for the highly technical readers, learn more about machine learning for stock trading in our depth series: Part 1 and Part 2.

‍

What is AI?

At its core, AI is simply machine learning (ML). ML is a branch of mathematics focused on the development of models that can learn patterns from data. The idea is simple: models are first trained and then used to predict. During the training process, a model is fed some dataset where it learns patterns in the data. This learning usually occurs through the minimization of prediction error. The lower the prediction error, the better the model has learned patterns in the data. After the training process, the model is now ready for prediction. When fed new, unseen data, a trained model can make predictions on that data by drawing from the patterns it previously learned.

This may sound complicated, but have you ever heard of linear regression before? A linear regression model is a very simple and primitive type of ML model. That’s right, you may have unknowingly created an AI model in your statistics class! Hopefully, this helps illustrate that AI isn’t some black box, but rather a collection of mathematical methods that you may or may not have already worked with before.

‍

How do we use AI?

Of course, modern-day ML models have improved significantly from linear regression, but the idea remains the same. At alphaAI, our proprietary models are based on industry-leading concepts but heavily modified to fit our use case. Our AI system is primarily comprised of two parts: predictive models and portfolio management systems.

Our predictive models are trained on multiple decades of data for over 10,000 global stocks. On average, each model is trained on more than 10 billion data points. Each model is trained to perform a unique predictive capability, and multiple models work together to make trading decisions. You can think of it as a team of investment professionals working together. Each individual has their own unique viewpoint and set of skills. As a whole, the collective experience of the entire team enables them to make better decisions than any single individual could. In the same way, each one of our models is unique in its predictive capabilities. The probability of making the best decisions greatly increases when multiple models are used together as they can cover each other’s weaknesses.

Our portfolio management systems use a rules-based approach to decide what to do with the predictions that our models generate. This includes making trades and managing risk. These systems also include numerous failsafe protocols to ensure that all the actions that are taken do not exceed the limits that we set.

‍

Why use AI for investing?

You may be surprised to hear that more than 90% of stock market trades are currently being made by machines. That’s because AI-driven trading offers numerous benefits over human-driven trading. Machines are simply better equipped to deal with numbers than humans are. Machines can process more data, faster. And they make emotionless decisions. When it comes to generating consistent, risk-adjusted returns over a long period of time, it’s extremely hard to beat systematic, AI-driven strategies.

‍

Closing Thoughts

Hopefully, you now have a better understanding of what AI is and how we use it. Don't worry, AI doesn’t have sentience, and there is no chance of it going off and making its own decisions. AI is another word for machine learning, and machine learning simply consists of a collection of predictive methods and models that can learn patterns from data.