FICO has been granted 9 patents, 3 foreign and 6 U.S., related to digital decisioning in the areas of fraud, artificial intelligence (AI), and machine learning technology. FICO has 218 active patents to date.
FICO: Highlights:
- FICO continues to drive unique industry specific innovation in machine learning and technology innovations to successfully operationalize AI at scale.
- FICO’s current patent portfolio consists of 218 active patents, with an additional 71 patent applications filed and pending.
Global analytics software firm FICO has been granted 9 patents, 3 foreign and 6 U.S., related to digital decisioning in the areas of fraud, artificial intelligence (AI), and machine learning technology. FICO’s extensive industry expertise in building AI and machine learning models provides value to enterprises that adopt FICO Platform and other FICO solutions.
“Patents are a product of industry inspired innovation, and we continue to develop tremendous IP in AI, machine learning, and decision management,” said Scott Zoldi, chief analytics officer at FICO. “FICO solutions deliver real-time decisions for businesses worldwide. The power of better decisions and hyper-personalized customer experiences, as exemplified by FICO Platform, is being driven by new AI technologies responding to ever-changing industry demands.”
FICO’s new patents include:
- “Soft Segmentation Based Rules Optimization for Zero Detection Loss False Positive Reduction” – which covers rules-based financial crime detection for detecting money laundering and fraud while reducing false positive detections. The techniques include the use of topic models to determine semantic structures underlying transactions and the derivation of behavior archetypes based on those structures, with the result being the reduction of false positives while preserving the same true positive detection set.
- “Mobile Attribute Time-Series Profiling Analytics” – which claims a means to assess the likelihood that a transaction is abnormal based on a behavioral pattern of a mobile device. The behaviors associated with the device improve fraud detection for transactions that do not occur on the mobile device. This patent has been validated in the UK, France, and Germany and used in FICO Falcon Fraud Manager, FICO Fraud Predictor with Merchant Profiles, and FICO Platform.
- “Latent Feature Dimensionality Bounds for Robust Machine Learning on High Dimensional Datasets” – which describes a method to identify the number of optimal hidden nodes in a neural network model, leading to the development of more robust and stable machine learning models at the core of responsible AI. This patent is used in FICO Platform.
- “Assessing the Presence of Selective Omission via Collaborative Counterfactual Interventions” – which addresses a critical challenge in operationalizing models in open data environments, where lenders and users of models need to understand the likelihood that the customer is omitting key data that impact outcomes. This patent is used in FICO’s Open Banking Analytic models.
SOURCE: Businesswire