Most of the time, we associate Artificial Intelligence with complex maths, big server computers and computer codes. What if the key to creating better AI models lies within the human body itself? A group of innovative scientists has just revealed a novel machine learning model, called Artificial Liver Classifier (ALC). And yes, it does what a human liver does!

How can a “Digital Liver” Learn?

The liver is the last resort for detoxifying the body, as it undergoes two cycles of detoxification (Phases I and II) to maintain our health. The ALC model imitates the same process for cleaning and sorting data:

·       Phase I (The Filter): The AI process is given raw input data, in this case "toxins", and the data are filtered by applying a mathematical "cofactor" matrix that removes non-essential noise.

·       Phase II (The Clean Up): The data is then cleansed through a "vitamin" matrix, making the data just the right stuff for the AI to make a very accurate final prediction. It uses an enhanced smart algorithm called IFOX, rather than the traditional dense training methods. IFOX serves as an integrated guide, automatically adjusting the AI's settings.

David vs. Goliath: To beat the Tech Giants, we need a David.

To find out just how effective this biological technique is, the researchers put ALC (our underdog "David") up against the "Goliaths" of the tech industry: industry standard models such as XGBoost, Support Vector Machines (SVM) and Logistic Regression (LR).