Global analytics software leader FICO was awarded new patents by the U.S. Patent and Trademark Office, encompassing technologies in fraud, artificial intelligence, machine learning, decision management and cybersecurity.
The new patents showcase FICO’s continuous innovations and leadership in operationalizing AI to deliver tangible results.
To date, FICO’s patent portfolio includes over 200 U.S. and international patents. These patents represent innovative technologies aimed at helping drive profitability, customer satisfaction, customer protection, and growth across industries such as financial services, telecommunications, healthcare, retail, transportation, supply chain, and more.
The technology covered by FICO’s new patents include:
• A Latent-Space Misalignment Measure of Responsible AI for Machine Learning Models - This method covers technological improvements that identify misaligned data in a latent space for machine learning models. The innovation provides valuable new techniques to measure the degree of misalignment of a given data source with respect to an underlying machine learning model, both during development and production, and thereby furthers a core tenant of Responsible AI.
• Enhanced Data Privacy Through Structure-preserving Autoencoders with Latent Space Augmentation - This is a method for generating refined (de-identified and anonymized) synthetic data from one or more sources of data.
• Method and Apparatus for Analyzing Coverage, Bias, and Model Explanations in Large Dimensional Modeling Data - This method improves the fairness and transparency of AI models by identifying biases and weaknesses in large, complex datasets. By analyzing how AI models make decisions and the data on which they are built, it helps ensure the models are more accurate and explainable, which is crucial for building trust and accountability in AI systems.
• Latent Feature Dimensionality Bounds for Robust Machine Learning on High Dimensional Datasets - This technology optimizes architectures of neural network models by compactly processing complex data, enhancing their stability and reliability. It helps ensure better performance when handling large, high-dimensional datasets, making machine learning models more accurate and dependable.
• Data Distillery for Signal Detection - This system identifies important patterns or "signals" within vast datasets, helping businesses prioritize key insights to utilize in models and risk strategies. It streamlines data analysis of relevant business signals making decision management more efficient and effective.
• Facial Recognition for User Authentication - This technology enhances user authentication by leveraging facial recognition to securely verify identities. Its integration with transaction fraud detection can strengthen the identification of potentially compromised accounts.
• System and Method for Linearizing Messages from Data Sources for Optimized High-Performance Processing in a Stream Processing System - This technology is useful for handling large streams of real-time data, such as in financial or customer service systems.
• Interpretable Feature Discovery with Grammar-based Bayesian Optimization - This method uses Bayesian search to discover and optimize key features in data, making machine learning models more understandable and interpretable. It helps ensure AI decisions are transparent, making it easier to trust the outcomes generated by these systems.
• Managing Missing Values In Datasets For Machine Learning Models - This technology addresses missing data elements and handling in machine learning models by filling in gaps. It provides strategies for AI systems to remain reliable when data elements are incomplete.
• Attributing Reasons to Predictive Model Scores - This improvement helps explain the factors that influence predictive model scores, making it easier to understand how decisions are made and providing rationale to consumers. It’s particularly valuable in fraud detection systems like FICO® Falcon® Fraud Manager and any transaction analytic AI, ensuring transparency in transaction patterns driving outcomes and trust in scoring.
• Cyber Security Adaptive Analytics Threat Monitoring System and Method - This system detects malware and other security threats by analyzing DNS and other network communication messages to assess threats. It provides businesses with real-time insights into potential attacks on key information and financial systems, helping them protect against cyber-attacks and data being actively infiltrated.
"At FICO, innovation is at the heart of everything we do, driven by talented individuals who are dedicated to solving complex challenges. Developing patented technologies requires time, expertise and collaboration to turn bold ideas into realities that drive innovation and transform industries,” said Nikhil Behl, executive vice president at FICO. “FICO’s chief analytics officer, Dr. Scott Zoldi, has been listed as an inventor on 101 patents, in collaboration with other data and analytic scientists, and he is also named on an additional 40+ patent applications in process. This accomplishment reflects Dr. Zoldi's integral role in driving FICO's AI innovation forward. His ability to transform complex challenges into impactful solutions has set the standard for what we can achieve when we put solving customer challenges at the heart of our innovations.”
“These new patents reflect our commitment to developing innovative new AI technologies and solutions that enable our clients to operationalize and make smarter, data-driven decisions,” said Dr. Scott Zoldi, chief analytics officer at FICO. "We innovate to meet and anticipate our clients’ needs, ensuring that the technology we deliver addresses key industry business challenges and support our client’s long-term success."