Such things happen every day in the UK: a customer checks out with a credit card at the grocery store, but the cashier says the card is rejected, but the things that have been bought are already installed, so I can imagine. However, the credit line of the card is good, the PIN code is correct, the customer has already purchased a lot of other things before, and now the credit card can't be really inexplicable.
For the financial services industry, these "false positives" have become an increasingly big problem. Not only customers and businesses are bothered, but also require staff to verify identity and unlock cards.
Every financial services company claims that they have a dedicated fraud detection system that distinguishes between normal use and fraud, but does not guarantee that domestic users' cards will not be rejected without cause. Regardless of how complex these systems claim to be, in fact, these false positives have already indicated problems within the system.
Cambridge
Therefore, in order to solve the problem of the system platform, a Cambridge-based company introduced adaptive behavior analysis technology, which can more accurately judge the quality of the transaction. The company is Featurespace, which sells platforms that use the ARIC engine and machine learning system to monitor the tiny details of complex events for anomalies.
Recently, another Cambridge-based company, Google's DeepMind, has attracted worldwide attention. The company also uses the concept of machine learning and Bayesian statistics. The system is famous for the robotic and human champions of the Go game, but we only saw a little bit of its power, because the machine will penetrate into the array of huge decision-making systems in the next decade.
Featurespace was a concept project of Cambridge University ten years ago. It was developed by David Excell (CTO) and Bill Fitzgerald (died in April 2014). The company's development has been slow and stable, and it has proved that the company is very hard. With the appointment of a new CEO in 2012, MarTIna King, the company's strength seems to have increased significantly. In 2014, investors formed a consortium with a total of $4.5 million in financing, including Mike Lynch, co-founder of Cambridge's famous company Autonomy.
In the conversation, MarTIna is obviously very busy (leading Yahoo UK and Capital Radio in a long career), she is warm and considerate, very flattering. Engineers are sometimes limited to technical details, but MarTIna's vision for Featurespace is clear and concise.
"When you're doing something new and different, it takes time to get people's approval. Then they know that there is a better way," MarTIna said, she firmly believes that despite adaptive behavior analysis. It used to be an esoteric laboratory technique, and everyone’s views are beginning to change. "This is the result of word of mouth."
Martina, through Mike Lynch, met Prof. Fitzgerald, co-founder of Featurespace. Although he died in 2014, she has been very lost since she talked about the professor.
On the other hand, Martina did not dare to compliment the fraud detection system used by banks. "They [financial services] have lost so much money," she sighed. "These attacks are so complicated that you can't predict their next step."
In the past few years, the problem of misjudgment and fraudulent fraud and customer dissatisfaction have become an important issue. But why have traditional systems struggled? Martina believes that the reason is simple, their fraud detection rules are based on known model patterns. As long as new things appear, it can lead to system crashes or management becomes expensive. Therefore, detecting fraudulent behavior eventually leads to expensive and manual processing.
ARIC simulates real-world more detailed fraud, discovering anomalies by understanding the context of the event. This is a theory that understands that the real world has gained traction in all forms of computer security, but does not show that successful implementation is trivial.
“It’s hard to do, you need to really understand the statistical analysis and provide this analysis in the enterprise,†Martina said. “We have created a market in many ways until recently that no one has talked about machine learning.â€
Its strength lies in the fact that all events, including fraud, are performed by or on behalf of humans. This gives Bayesian mathematical support, and ARIC not only grasps the meaning of understanding an event in its search, but also predicts what might happen next.
The challenge for Featurespace is simply to better detect fraud, cheaper and faster.
"When I wake up in the morning, how can I shorten my time?"
Martina estimates that the global response to false positives could reach $6.4 billion, a figure that accounts for a large percentage of real fraud. If the resolution of fraud becomes a heavy burden, then the problem needs to be resolved.
"We know that Featurespace can reduce alerts by 70%, so it doesn't take so many people to deal with misjudgments," Martina said.
future
The company has more and more customers, and Martina is particularly proud that it signed a five-year contract with mobile payment processor Zapp in 2015, and Zapp and Barclay are working together in the UK. Featurespace handles thousands of transactions per second, proving its depth and breadth of technology.
ARIC has also been used in a large US bank, as well as a range of other financial institutions, but Martina cannot disclose the company's name and it is necessary to keep its products confidential for competitive reasons.
The final decision on the company's continued growth is not just its customer list, but the entire state of the anomaly detection itself. With the road of Autonomy pioneering ten years ago, DeepMind is now world-famous, and Featurespace is also a place where people can be together.
At the turn of the millennium, Cambridge has transcended its reputation as an interesting computer science and a place of intelligent mathematics and established its position as one of the world's most important machine learning centers. Martina believes that computers are being used sparingly, not only to make the world work differently, but also for a better way. Otherwise, e-commerce will sink slowly because it can't provide reliable trading.
Induction Cooker Main Board,Motor Starter Button,Ignition Starter Switch For Foton,Main Spindle Induction Motor
Ningbo Zhenhai Rongda Electrical Appliance Co., Ltd. , https://www.centrifugalswitch.com