20230126695. ML MODEL DRIFT DETECTION USING MODIFIED GAN simplified abstract (Raytheon Company)

From WikiPatents
Jump to navigation Jump to search

ML MODEL DRIFT DETECTION USING MODIFIED GAN

Organization Name

Raytheon Company

Inventor(s)

Nicole M. Hatten of Allen TX (US)

ML MODEL DRIFT DETECTION USING MODIFIED GAN - A simplified explanation of the abstract

This abstract first appeared for US patent application 20230126695 titled 'ML MODEL DRIFT DETECTION USING MODIFIED GAN

Simplified Explanation

The patent application discusses devices, systems, and methods for detecting machine learning model drift. This involves monitoring and analyzing the output of a deployed machine learning model to determine if it is suffering from model drift. Here are the key points:

  • The method involves receiving data that defines the structure of a deployed machine learning model, including the number of layers and neurons in each layer, as well as the weights of each neuron.
  • The deployed machine learning model is operated within a modified generative adversarial network (GAN) architecture.
  • While the model is operating, the output of a hidden layer is recorded and a metric is determined based on this output.
  • The deployed machine learning model is then re-deployed, and the re-deployed model is monitored to see if it is suffering from model drift based on the metric.

Potential applications of this technology:

  • This technology can be applied in various industries where machine learning models are used, such as finance, healthcare, and manufacturing.
  • It can be used to detect and prevent model drift in predictive maintenance systems, ensuring accurate and reliable predictions for equipment maintenance.
  • In the financial sector, it can help identify changes in market behavior and adjust trading algorithms accordingly to avoid financial losses.

Problems solved by this technology:

  • Model drift is a common problem in machine learning, where the performance of a deployed model deteriorates over time due to changes in the underlying data distribution.
  • This technology provides a method to detect and monitor model drift, allowing for timely adjustments and improvements to the deployed machine learning model.
  • By detecting model drift, organizations can ensure that their machine learning models continue to provide accurate and reliable predictions, avoiding potential negative consequences.

Benefits of this technology:

  • The method provides a systematic approach to detect and monitor model drift, allowing organizations to proactively address any performance degradation in their machine learning models.
  • By identifying model drift early on, organizations can take corrective actions to improve the performance and reliability of their deployed models, leading to better decision-making and outcomes.
  • This technology can help organizations maintain the effectiveness of their machine learning models over time, ensuring that they continue to provide value and meet the desired objectives.


Original Abstract Submitted

discussed herein are devices, systems, and methods for machine learning (ml) model drift detection. a method can include receiving machine learning (ml) data defining a number of layers of neurons, a number of neurons per each layer, and weights for each neuron of a deployed ml model, operating the deployed ml model in a modified generative adversarial network (gan) architecture, while operating the deployed ml model, recording output of a hidden layer of the deployed ml model, determining a metric of the output, and re-deploying the deployed ml model and monitoring whether the re-deployed ml model is suffering from ml model drift based on the metric.

Retrieved from "http://wikipatents.org/index.php?title=20230126695._ML_MODEL_DRIFT_DETECTION_USING_MODIFIED_GAN_simplified_abstract_(Raytheon_Company)&oldid=21812"