Nand Kishor is the Product Manager of House of Bots. After finishing his studies in computer science, he ideated & re-launched Real Estate Business Intelligence Tool, where he created one of the leading Business Intelligence Tool for property price analysis in 2012. He also writes, research and sharing knowledge about Artificial Intelligence (AI), Machine Learning (ML), Data Science, Big Data, Python Language etc... ...Full Bio
Nand Kishor is the Product Manager of House of Bots. After finishing his studies in computer science, he ideated & re-launched Real Estate Business Intelligence Tool, where he created one of the leading Business Intelligence Tool for property price analysis in 2012. He also writes, research and sharing knowledge about Artificial Intelligence (AI), Machine Learning (ML), Data Science, Big Data, Python Language etc...
Data science is the big draw in business schools
1092 days ago
7 Effective Methods for Fitting a Liner
1102 days ago
3 Thoughts on Why Deep Learning Works So Well
1102 days ago
3 million at risk from the rise of robots
1102 days ago
Top 10 Hot Artificial Intelligence (AI) Technologies
Demystifying machine learning
In 1959, Arthur Samuel defined machine learning as a "field of study that gives computers the ability to learn without being explicitly programmed." No, wait - Did I just type "1959?" Yes, true!
Machine learning, which is a branch of artificial intelligence (AI), has been around a while, but the technology has made enormous strides since 1959. The ability to learn without explicitly being programmed sounds like a mantra taken from a child-rearing manual from the 1950s. However, this concept has even greater potential today when we apply the technology to fraud prevention.
Machine learning will still require the presence of humans to calibrate and confirm fraud, and to assist the model in learning from historical data. But the promise of a faster and more accurate solution for fraud prevention has CO-OP very excited about the possibilities - which is why our company is becoming an early adopter of this technology.
According to a study recently published on finextra.com, global financial services firms could save $12 billion annually - or more - by optimizing adaptive, machine learning-based fraud management technology. This declaration alone is important: We have to do a better job of fighting fraud - and of reducing the money left on the table for fraudsters to steal. We also need to do this with fewer false positives and a higher degree of accuracy, benefits that will drive our members back to us time and time again as their preferred payment provider.
It's time to jiggle the high stakes game of fraud prevention in our favor for once. And I think we can accomplish this quite handily if we better understand both our challenges and where the benefits of machine learning can alleviate our struggles.
The Struggle Is Real
Every day in every state across North America, credit unions struggle with the need to prevent fraud while also providing a superior experience for their members. The three pillars of success are generally defined as follows:
Fraud is vanquished to the highest degree possible
The member experience is seamless, secure and convenient
The credit union's brand maintains preferential top of mind, wallet and phone status at all times
An Impressive List of Early Adopters
A recent article cited that one of the pioneers in bringing early AI tools into banking and trading was Charles Schwab. Way back in 2011, the firm started using chart pattern recognition to both simplify complex trading activities and provide a more intuitive experience for active traders.
Many major financial brands followed suit, and are harnessing the power and analytics behind machine learning and AI today. PayPal, USAA and Capital One are simply a few of the companies that utilize this growing area of technology to detect fraud, reduce money laundering and, in some cases, identify the customer through voice recognition.
We All Share the Same Complaints
I hear it every time I visit a credit union. "Why didn't your neural net capture that?" "How is it possible that every time my CEO pumps a tank full of gas his card is declined?" and the classic "My member's vacation was impacted by fraud that occurred thousands of miles from where he was vacationing." Okay, I cannot disagree here that we need a better system that is more sensitive and capable of differentiating between legitimate commerce and fraudulent activity.
When your member hails a cab, purchases lunch at the airport, hops on a flight and ends up renting a car and staying in a hotel, there should be some sort of logic behind your fraud analysis that says "This is John on a routine business trip," and not "Deny John's payment card due to unusual activity."
This is where machine learning will excel as it builds a logical transactional profile for each member and evolves it over time, while still catching more fraud using a global model that is combined with a demographic user profile.
I was recently visiting with a credit union and discussing the merits of machine learning when the CEO exclaimed, "Oh! Machine Learning will actually do what we assumed for years that neural net tools were doing." Yes, exactly!
The Death of the Neural Net?
Many people have asked me if I see machine learning outpacing and eventually replacing classic neural net tools that have dominated our industry for such a long period of time. Admittedly, our fraud is drastically different today, while neural net tools have seen very little embellishments over the years.
This combination of rising fraud and lagging advancements in neural network technology pleads the case for something new that is faster and more capable of delivering a high-quality member experience - and that places the credit union at the top of the member's wallet.
Will CO-OP deploy machine learning that eventually replaces neural net technology? We may get to that juncture someday, but not before investing the time and resources needed to adequately compare the benefits. Rest assured - we won't rip out and replace anything before we have determined that the heavy lifting can be achieved with one tool. Meanwhile, come along for the ride with us. It promises to be an exciting one. Read More