Canada's RBC files patents for AI inventions as bigtechs soar

Canada’s RBC files patents for AI inventions as bigtechs soar

Written by Ruby Hinchliffe
21st January 2020

The Royal Bank of Canada (RBC) is trying to put its commercial stamp on several of artificial intelligence (AI) inventions in what could be a move against big technology companies encroaching on the bank’s space.

Ten patent applications regarding AI inventions were submitted by RBC across 2019 – four of which became public in December, according to the Canadian Intellectual Property Office’s patent database.

RBC has its own AI research institute, Borealis AI

First highlighted by the Financial Post, the filings show Canada’s biggest bank attempting to stop its AI inventions from being made by other companies without permission from the bank itself.

RBC has its own AI research institute, Borealis AI. In 2018, RBC’s CEO Dave McKay told Bloomberg he did have concerns over big technology firms positioning themselves between the major lender and its customers.

He says: “The funnel of decision-making and life moments before the financing decision or before the banking decision is mission-critical in us taking control in that journey and not being beholden to somebody else’s platform.”

The four patent applications which became public last month covered ‘K-LSTM (long term memory loss) architecture for purchase prediction’, ‘correcting bias in supervised machine learning data’, ‘processing natural language statements’ and a ‘portfolio-based text analytics tool’. All four patents were filed as early as June last year by the bank.

The K-LTSM patent description read: “Client history can span many years, and events from the entire time span can contribute to predicting likelihood of purchase. Traditional neural network models suffer from long term memory loss. Accordingly, there is a need to predict likelihood of purchase, taking into account long term client history.”

The bank also submitted three patents in May 2019 for ‘machine learning architecture with adversarial attack defence’, ‘trade platform with reinforcement learning’, and ‘machine natural language processing’. Then in March 2019, it submitted two more patents for ‘monitoring machine learning models’ and ‘machine-learning step-size adaptation’.

The tenth patent, ‘pre-training neural networks with human demonstrations for deep reinforcement learning’, was submitted by RBC in January 2019.

Following McKay’s comments to Bloomberg in 2018, these patents could be seen as proof the bank is trying to prevent big technology firms from competing with it for data on certain purchasing journeys by customers.