SoFi Technologies has been granted a patent for a system that uses machine learning to improve fairness metrics by training on historical data, receiving real-time data, and generating risk scores. The system includes processors and computer-readable media to perform these functions efficiently. GlobalData’s report on SoFi Technologies gives a 360-degree view of the company including its patenting strategy. Buy the report here.

According to GlobalData’s company profile on SoFi Technologies, Sensor guided flow mixing was a key innovation area identified from patents. SoFi Technologies's grant share as of February 2024 was 48%. Grant share is based on the ratio of number of grants to total number of patents.

Machine-learning model for fairness metrics and risk score generation

Source: United States Patent and Trademark Office (USPTO). Credit: SoFi Technologies Inc

A recently granted patent (Publication Number: US11928730B1) outlines a system and method for training a machine-learning model to improve fairness metrics. The system involves one or more processors executing computing instructions stored on non-transitory computer-readable media. The model is trained using historical data with maximization and minimization problems to enhance fairness metrics. Real-time data is received, and a risk score is generated based on the trained model and real-time data. The training process includes an estimation bundling technique to create a uniform predicted output by estimating convergence points, multipliers, regularization items, and solving augmented minimization problems. The system evaluates whether the predicted output meets fairness criteria and model prediction power criteria, updating control parameters and regenerating outputs as needed.

Furthermore, the patent describes the training process in detail, highlighting the parallel execution of maximization and minimization problems, comparison of protected groups against benchmark groups based on fairness metrics, and outputting risk scores for credit application approval decisions. The method involves updating control parameters and regenerating outputs if fairness criteria are not met, ensuring the model satisfies fairness and prediction power criteria. The system aims to address issues of bias and discrimination in machine learning models by incorporating fairness metrics into the training process, ultimately improving decision-making processes in various applications such as credit risk assessment. The patent provides a comprehensive framework for developing fair and accurate machine learning models, contributing to the advancement of ethical AI practices in the industry.

To know more about GlobalData’s detailed insights on SoFi Technologies, buy the report here.

Data Insights

From

The gold standard of business intelligence.

Blending expert knowledge with cutting-edge technology, GlobalData’s unrivalled proprietary data will enable you to decode what’s happening in your market. You can make better informed decisions and gain a future-proof advantage over your competitors.

GlobalData

GlobalData, the leading provider of industry intelligence, provided the underlying data, research, and analysis used to produce this article.

GlobalData Patent Analytics tracks bibliographic data, legal events data, point in time patent ownerships, and backward and forward citations from global patenting offices. Textual analysis and official patent classifications are used to group patents into key thematic areas and link them to specific companies across the world’s largest industries.