ACADEMIA
ATB Financial partners with UAlberta scientists in projects to create a framework for ethical design of AI
From artificial intelligence (AI) to data science to mathematical finance, ATB Financial has its eyes on the future for its clients, Albertans and the tech sector.
Following a red hot 2017, with multiple announcements of collaborations and partnerships, University of Alberta scientists have just received another green stamp of approval from the financial sector with the announcement of a new collaboration with ATB Financial.
In a four-year partnership that will see an infusion of nearly a million dollars ($940,000) into various research projects and student internships, ATB's data scientists are working with University of Alberta scientists, collaborating to create customer-centric experiences that redefine the value a financial institution can deliver.
Three initial projects are underway. They are focused on real-time fraud detection, predictive analytics--to identify key moments in customer or business journey--that are focused on customer support and a tailored recommendation system similar to those ubiquitous with Amazon. All of the projects are designed to enhance human capabilities, allowing ATB to function more effectively, deepening their efforts in maintaining customer relationships.
"Partnering with the University of Alberta was a priority for ATB," says Wellington Holbrook, Chief Transformation Officer. "Both our organizations share a deep commitment to elevating Alberta as a leader in AI and machine learning (ML), making meaningful contributions to the global conversation. We are deeply committed to serving Albertans by leveraging the incredible intellectual capital that's in our own backyard."
AI for the win
"Data scientists at ATB are interested in reimagining machine learning to understand ways in which ATB can improve the financial confidence of their clients," said Dustin Morrill, a University of Alberta PhD student who is no stranger to creating best positions for big wins. Morrill worked with his supervisor Michael Bowling, best known as one of the lead scientists with DeepMind Alberta, who already enjoys an international reputation for his work developing artificial intelligence. (Morrill and Bowling were two of the co-authors on 2017's landmark Science paper for DeepStack, which defeated human professional poker players in heads-up no-limit Texas hold'em.)
Morrill will be applying similar techniques with ATB as part of his PhD in order to make robust decisions in real-world problems where there is recognition of uncertainty in models. In this case, it's not cards on the poker table but helping the financial health of Albertans that is the aim of the game. "We are working on a recommendation system, like those used by Amazon and Netflix, for example, but with the aim of improving the financial well-being of ATB clients."
Data analytics platform
Another project deals with the use of data science for trend analysis and fraud detection, useful for modeling the past to predict the future of user behavior. And for that, it's over to computing scientist and assistant professor Omid Ardakanian. "We are developing the ability to do real-time analytics at scale, that is to analyze billions of data points in a fraction of a second," said Ardakanian, who will be working with assistant professor Hamzeh Khazaei to build a data analytics platform which enables real-time fraud detection along with many other applications.
Data science meets risk management
Christoph Frei, associate professor in mathematical finance, will be working at the intersection of data science and risk management, with his two graduate students to create early notification systems for ATB. Specifically, the researchers will be diving into early prediction of customer events, such as encountering financial difficulties, which is quite possibly indicative of other issues and an opportunity where ATB can deliver unparalleled service to help customers get back on track.
"These things may not just happen randomly or suddenly, and most often individuals hesitate to reach out for help because they feel vulnerable. Through our work, we can help ATB use this opportunity to reframe financial conversations and create a more positive outcome for the customer, a scenario that is good for the customer and ATB," said Frei.
Not only does Frei like the aspect of helping consumers and businesses, he is taking advantage of the opportunity for graduate students. "It's very beneficial to students when they are exposed to real-world problems so that they can connect what they are learning in theory with practical applications."
It is also expected that researchers on all three projects will publish their results, meaning benefits to the entire financial and tech sector.
Interdisciplinary solutions
"The University of Alberta has one of the top three academic artificial intelligence and machine learning (AI/ML) research groups in the world. Our skills and technologies can be used to help companies better inform their decision-making processes," said the Dean of the University of Alberta Faculty of Science, Jonathan Schaeffer.
"We are excited to work with visionary companies on interesting problems, especially those that have a direct benefit to Albertans. ATB wants to be a leader in applying AI/ML to challenges in the financial sector. Their aspirations are a perfect match for our ambitions."
In addition to the research collaboration with three projects already underway, ATB has expanded its commitment in artificial intelligence space through the establishment of an Alberta-centric Research & Affiliate Network, led by UAlberta computing science professor and artificial intelligence expert Randy Goebel that will connect top experts and practitioners in the field to support future minds and work in the space.
Also, the ethical application of AI has never been a more essential global conversation. As such, the AI, People & Society initiative kickstarted by ATB and led by UAlberta humanities computing professor Geoffrey Rockwell, is an open, community driven platform that will bring together diverse thinkers and experts from across humanities, design, social sciences, technology and beyond to help shape a vision of the future that benefits all and to create a framework for ethical design of AI.