BBVA collaborates with MIT to enhance card fraud detection

Maria Nikolova

The aim is to reduce the level of false positives when detecting fraudulent card transactions.

BBVA has partnered with researchers from the Massachusetts Institute of Technology (MIT) in an effort to improve the efficiency of card fraud detection. In particular, the aim is to minimize the level of false positives in fraudulent card transactions.

For years, banks have been using machine learning techniques to detect card fraud. Algorithms are trained to learn to recognize behavioral patterns typical of fraudulent transactions and block subsequent ‘suspicious’ operations. However, this results in false positives. When customers pay an unusual amount of money or use their cards far from their usual locations, their cards often get blocked because the system erroneously qualifies these anomalies as “suspicious”. Let’s recall that, in March last year, for instance, there was a flood of complaints in Russia about bank accounts being blocked following checks for AML compliance conducted by automated systems.

The reason for this is that the automated systems using machine learning are subject to a series of limitations when applied to card fraud. As their ability to detect actual fraudulent transactions grows, so does the number of false positives they derive; put otherwise, the number of legitimate transactions they erroneously label as fraudulent also rises.

Aiming to improve the situation, BBVA’s Corporate Security and Engineering Risk (CS&ER) team started collaborating with MIT’s Computer Science and Artificial Intelligence Lab (CSAIL) in 2015. The MIT and BBVA research team used the Deep Feature Synthesis method, a model developed by two MIT researchers, Kalyan Veeramachaneni and James Max Kanter, which makes it possible to derive highly detailed features from datasets in an automated manner.

This new approach allowed the research team to derive over 200 new additional features from every transaction. These features can be used to describe the behavior of card transactions in greater detail and improve the results yielded by the detection engines. Some of these new features used to fine tune true positive results indicate, for example, whether the user was physically present while the transaction was taking place, or the average time spent in certain stores. This helps to more accurately identify when a particular card user’s transaction veers from the norm.

So far, tests are yielding promising results. False positives have fallen markedly, while keeping the same level of fraud detection, compared to other existing products and systems. Specifically, applying the model to a dataset of 1.8 million BBVA transactions cut the number of false positives by 54% over traditional models, saving the bank approximately €190,000.

Applying the results of this research to BBVA’s fraud detection systems has proven that they have the potential to improve customer satisfaction in card transactions without compromising current security standards.

Read this next


US and South Korea seek extradition of Luna founder Do Kwon

Both US and South Korean officials are seeking to extradite Terraform Labs CEO Do Kwon, just hours after he was arrested in Montenegro.

Retail FX

Pepperstone UK doubles profit, client assets in 2022

The London-based entity of Australian FX broker Pepperstone has reported its financials for the fiscal year ending June 30, 2022. The group had outperformed the last year’s flat performance, having doubled revenues and boosted the broker’s bottom line and shareholders’ fortunes.

Institutional FX

CLS FX volume rises to just shy of $2 trillion in February

Foreign exchange settlement provider, CLS Group saw strong volumes in February 2023 as the banking crisis continues to weigh on a world economy that’s yet to fully recover from the Russia-Ukraine war’s shocks.

Digital Assets

Binance restores trading after 2-hour outage

Binance suffered a breakdown on its trading engine that lasted for about two hours, but the premier cryptocurrency exchange finally managed to restore normal operations at around 14:00 UTC.

Digital Assets

Tether earns $700 million in Q1, taking excess reserves to $1.6 billion

Tether chief technology officer Paolo Ardoino said the world’s largest stablecoin issuer expects to earn more than $700 million in the January-Mach quarter, which will be added to the reserve backing its stablecoin (USDT).

Digital Assets

Narwhal Finance Secures $1M in Seed Funding Led by Animoca Ventures

Narwhal Finance received strong support from Animoca Ventures and angel investors in a $1 million seed funding round, reinforcing the company’s vision of providing an accessible platform to all.


SteelEye tries ChatGPT for market surveillance

This capability can be used as a starting point for initiating a surveillance investigation and to standardize workflow processes to boost the throughput and consistency of cases. It is also useful when analyzing communications in foreign languages, as the system returns the above insights in English regardless of the languages being used.

Industry News

SEC charges ex-Morgan Stanley advisor of NBA players after $13m fraud

Darryl Matthew Cohen was arrested this week and is facing three different federal counts of fraud, which could amount to 20 years in prison if convicted, besides the SEC complaint. 

Industry News

AWS FinTech Africa Accelerator launched, applications until April 27, 2023

Founders will be offered tech resources, expert guidance, and a global network of industry leaders, technologists, entrepreneurs, investors, associations, and partners, in order to build their fintech products.