Discover Financial Services is making its machine learning research publicly available to peers in an effort to help the industry reach common standards with the technology.

The $133 billion financial institution (FI) has done extensive research on machine learning definitions, approaching machine learning models, bias quantification and explanation, bias mitigation, approaching computing grouped explainers, and tree-based machine learning models, Arjun Kannan, director of data science research, told Bank Automation News.

The FI “has done a lot of original research, which, different from others, we have actually put out our work in the public domain,” Kannan previously told BAN.

The research papers are posted on ArXiv, an open-access repository of electronic pre-prints owned by Cornell University, Kannan told BAN.

Discover also uses formal and informal networks to “spread the word” about the published reports through platforms including LinkedIn, conferences, working groups and through partners, Kannan said.

For example, Discover has researched machine learning, use of compliant data and explainable AI. In sharing those findings, other FIs can benefit from outcomes Discover has already found, Kannan said.

Investing in analytics

Discover’s total operating expenses in the second quarter increased 15% year over year to $1.4 billion due to increases in compensation and headcount, and higher technology and consulting costs, according to the FI’s earnings supplement released last week.

As Discover increased operating costs, they invested in advanced analytics to drive efficiencies, Chief Financial Officer John Greene said during last week’s earnings call. With efficiencies in mind, “We’ve decided that the first priority is get the right resources in to focus on the issues.”

In doing so, Discover has added 3,000 people to the team in 2023, Greene said. “When you bring on additional people … there’s costs that go along with it.”

“As we manage through this situation, I continue to believe there will be opportunities to drive efficiencies,” Greene said.

[stock_market_widget type=”inline” template=”generic” assets=”DFS” markup=”{name} ({symbol}) is trading at {price} ({change_pct}) as of {last_update}” display_currency_symbol=”true” api=”yf”]