Artificial Intuition | “Artificial Intuition” of Artificial Intelligence Imitating Human Intuition

Original article was published by Jarvis+ on Artificial Intelligence on Medium

Artificial Intuition | “Artificial Intuition” of Artificial Intelligence Imitating Human Intuition

Artificial Intelligence (AI) is one of the most powerful technologies ever developed, but it is not as new as you think. In fact, it has undergone several evolutions since its inception in the 1950s. The first generation of artificial intelligence was “descriptive analysis”, which answered the question “What happened?” The second generation of “diagnostic analysis” points out: “Why happened?” The third generation is “predictive analysis”, which answers the following questions: “Based on what has already happened, what will happen in the future?”

Although predictive analysis can be very useful and save time for data scientists, predictive analysis still relies entirely on historical data. Therefore, in the face of new unknown scenarios, data scientists will be helpless. In order to have true “artificial intelligence”, we need machines that can “think” on their own, especially when faced with unfamiliar situations. The AI ​​we need can not only analyze the displayed data, but also show “intuition” without adding up. In short, we need artificial intelligence that can mimic human intuition.

Artificial Intuition

The fourth generation of AI is “artificial intuition”. Artificial intuition can accurately analyze the unknown without any historical background. It uses complex algorithms to identify any correlations or abnormalities between data, and then make judgments. Just like human intuition allows us to make decisions without special instructions. This is similar to an experienced detective who can enter the crime scene and immediately know that something seems to be wrong, or an experienced investor can spot trends before others.

Artificial intuition is one of the development trends of artificial intelligence in the future. In fact, the concept of artificial intuition was considered impossible five years ago. But now, companies like Google, Amazon, and IBM are working hard to develop solutions, and some companies have already let them run.

How artificial intuition works?

How can artificial intuition accurately analyze unknown data to point it in the right direction without any historical background? The answer lies in the data itself. Once the current data set is displayed, sophisticated artificial intuition algorithms can identify any correlations or anomalies between the data points.

First, human intuition is not to build quantitative models for processing data, but to apply qualitative models. It analyzes the data set and develops a contextual language that represents the overall configuration it observes. This language uses various mathematical models such as matrices, Euclidean spaces and multidimensional spaces, linear equations and eigenvalues ​​to represent the “big picture.” If you think of the big picture as a huge puzzle, you can intuitively see the complete puzzle from the beginning, and then backtrack according to the relationship between feature vectors to fill in the gaps.
Conceptually, this provides a guide for visualizing exception identifiers. Then, mark any feature vectors that cannot fit into the large image as suspicious.

How to use artificial intuition?

Artificial intuition can be applied to almost any industry, and great progress has been made in the field of financial services. Large banks around the world use it to detect complex new financial cybercrime schemes, including money laundering, fraud and ATM hacking. Suspicious financial activities are usually hidden in thousands of transactions with their own set of connection parameters.

By uncovering these hidden relationships between seemingly innocent transactions, human intuition is able to detect the bank and send it an “unknown unknown” (previously invisible and therefore an unexpected attack) and alarm. Not only that, but it can also interpret the data in a traceable and recorded manner, allowing bank analysts to prepare executable suspicious activity reports for the Financial Crime Enforcement Network (FinCEN).
Artificial intuition does not happen automatically. Instead, it analyzes the data set by applying a qualitative model and develops a contextual language that represents the overall configuration it observes. By using extremely complex mathematical algorithms, human intuition can quickly identify the five most influential parameters and present them to analysts.