Dell products l.p. (20240256914). SYSTEM AND METHOD FOR IDENTIFYING INPUT FEATURES CONTRIBUTING TO LATENT BIAS IN INFERENCE MODELS simplified abstract
Contents
SYSTEM AND METHOD FOR IDENTIFYING INPUT FEATURES CONTRIBUTING TO LATENT BIAS IN INFERENCE MODELS
Organization Name
Inventor(s)
OFIR Ezrielev of Be'er Sheva (IL)
FATEMEH Azmandian of Raynham MA (US)
SYSTEM AND METHOD FOR IDENTIFYING INPUT FEATURES CONTRIBUTING TO LATENT BIAS IN INFERENCE MODELS - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240256914 titled 'SYSTEM AND METHOD FOR IDENTIFYING INPUT FEATURES CONTRIBUTING TO LATENT BIAS IN INFERENCE MODELS
The abstract of the patent application describes methods, systems, and devices for providing computer-implemented services while managing inference models to reduce bias in the inferences provided. By identifying and removing input features with high contributions to latent bias from the training data set, new inference models can be trained to provide less biased inferences.
- Identification and removal of input features with high contributions to latent bias from the training data set
- Training of new inference models using the remaining input features
- Replacement of the old inference model with the new one to reduce bias in the inferences provided
- Management of inference models to reduce the likelihood of bias features
- Use of data processing systems to manage inference models
Potential Applications: - Improving the accuracy and fairness of computer-implemented services - Enhancing the reliability of inferences provided by data processing systems
Problems Solved: - Reducing bias in the inferences provided by computer-implemented services - Managing inference models to improve the quality of the services offered
Benefits: - Increased accuracy and fairness in the inferences provided - Enhanced trust in the computer-implemented services - Improved decision-making based on the inferences generated
Commercial Applications: Title: "Bias Reduction in Computer-Implemented Services" This technology can be applied in various industries such as finance, healthcare, and marketing to ensure fair and accurate inferences are provided to users. Market implications include improved customer satisfaction, regulatory compliance, and competitive advantage.
Prior Art: Readers can explore prior research on bias reduction in machine learning models, fairness in artificial intelligence, and data processing systems to gain a deeper understanding of the technology described in the patent application.
Frequently Updated Research: Stay informed about the latest advancements in bias reduction techniques, fairness algorithms, and data processing systems to enhance the effectiveness of managing inference models and reducing bias in computer-implemented services.
Questions about Bias Reduction in Computer-Implemented Services: 1. How does managing inference models help reduce bias in computer-implemented services? Managing inference models involves identifying and removing input features with high contributions to latent bias, training new models, and replacing the old ones to provide less biased inferences. 2. What are the potential applications of bias reduction technology in various industries? Bias reduction technology can be applied in finance, healthcare, marketing, and other sectors to improve the accuracy and fairness of inferences provided by computer-implemented services.
Original Abstract Submitted
methods, systems, and devices for providing computer-implemented services are disclosed. to provide the computer-implemented services, inference models used by data processing systems may be managed to reduce the likelihood of the inference models providing inferences indicative of bias features. the inference models may be managed by identifying input features in the training data set used to train the inference models with a high contribution to latent bias exhibited by the inference models. the input features may be removed from the training data set and a new inference model may be trained using the remainder of the input features. the new inference model may replace the inference model and, therefore, the inferences provided by the new inference model may be less likely to include latent bias thereby reducing bias in computer-implemented services provided using the inferences.
