17931803. SUFFICIENCY ASSESSMENT OF MACHINE LEARNING MODELS THROUGH MAXIMUM DEVIATION simplified abstract (International Business Machines Corporation)

SUFFICIENCY ASSESSMENT OF MACHINE LEARNING MODELS THROUGH MAXIMUM DEVIATION

Organization Name

International Business Machines Corporation

Inventor(s)

Dennis Wei of Sunnyvale CA (US)

Rahul Nair of Dublin (IE)

Amit Dhurandhar of Yorktown Heights NY (US)

Kush Raj Varshney of Chappaqua NY (US)

Elizabeth Daly of Dublin (IE)

Moninder Singh of Farmington CT (US)

Michael Hind of Cortlandt Manor NY (US)

SUFFICIENCY ASSESSMENT OF MACHINE LEARNING MODELS THROUGH MAXIMUM DEVIATION - A simplified explanation of the abstract

This abstract first appeared for US patent application 17931803 titled 'SUFFICIENCY ASSESSMENT OF MACHINE LEARNING MODELS THROUGH MAXIMUM DEVIATION

Simplified Explanation

The patent application describes techniques for determining the sufficiency of machine learning models by measuring the maximum deviation of a supervised learning model from a reference model over a certification set.

• Measurement component measures maximum deviation of supervised learning model from reference model over certification set
• Analysis component determines sufficiency of supervised learning model based on maximum deviation

Potential Applications

This technology can be applied in various fields such as:

• Quality control in manufacturing processes
• Fraud detection in financial transactions
• Medical diagnosis and treatment planning

Problems Solved

This technology helps in:

• Ensuring the accuracy and reliability of machine learning models
• Identifying potential errors or biases in the models
• Improving decision-making based on machine learning predictions

Benefits

The benefits of this technology include:

• Increased trust in machine learning models
• Enhanced performance and efficiency of predictive models
• Reduction of errors and uncertainties in model predictions

Potential Commercial Applications

This technology can be commercially applied in:

• Healthcare industry for personalized medicine
• Financial sector for risk assessment and fraud detection
• Manufacturing industry for quality control and process optimization

Possible Prior Art

One possible prior art for this technology could be the use of cross-validation techniques to evaluate the performance of machine learning models. Another could be the use of sensitivity analysis to assess the robustness of predictive models.

How does this technology compare to existing methods for evaluating machine learning models?

This article does not provide a direct comparison with existing methods for evaluating machine learning models. It would be helpful to understand the specific advantages or limitations of this technique compared to traditional approaches.

What are the potential limitations or challenges in implementing this technology in real-world applications?

The article does not address the potential limitations or challenges in implementing this technology in real-world applications. It would be important to consider factors such as computational resources, data availability, and model complexity when applying this technique in practical settings.

Original Abstract Submitted

Techniques regarding determining sufficiency of one or more machine learning models are provided. For example, one or more embodiments described herein can comprise a system, which can comprise a memory that can store computer executable components. The system can also comprise a processor, operably coupled to the memory, and that can execute the computer executable components stored in memory. The computer executable components can comprise a measurement component that measures maximum deviation of a supervised learning model from a reference model over a certification set and an analysis component that determines sufficiency of the supervised learning model based at least in part on the maximum deviation.