Samsung electronics co., ltd. (20240161007). METHOD AND DEVICE WITH AUTOMATIC LABELING simplified abstract

From WikiPatents
Revision as of 02:53, 23 May 2024 by Wikipatents (talk | contribs) (Creating a new page)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

METHOD AND DEVICE WITH AUTOMATIC LABELING

Organization Name

samsung electronics co., ltd.

Inventor(s)

Heewon Kim of Suwon-si (KR)

Hyun Sung Chang of Suwon-si (KR)

Jaeyun Lee of Suwon-si (KR)

Kiho Cho of Suwon-si (KR)

METHOD AND DEVICE WITH AUTOMATIC LABELING - A simplified explanation of the abstract

This abstract first appeared for US patent application 20240161007 titled 'METHOD AND DEVICE WITH AUTOMATIC LABELING

Simplified Explanation

The abstract describes a method for training models to predict confidences of labels for data samples, correcting incorrect labels, and automatically correcting other incorrect labels with estimated correct labels.

  • Training a first model to predict label confidences for data samples in a training dataset.
  • Using a corrected data sample obtained by correcting an incorrect label based on a corresponding confidence detected by the first model and an estimated corrected label generated by a second model.
  • Training the second model to estimate correct labels for the data samples.
  • Estimating a correct other label corresponding to another incorrect label detected based on a corresponding confidence generated by the first model.
  • Automatically correcting the other incorrect label with the estimated correct other label.

Potential Applications

This technology could be applied in various fields such as:

  • Natural language processing
  • Image recognition
  • Fraud detection

Problems Solved

This technology helps in:

  • Improving the accuracy of label predictions
  • Correcting mislabeled data samples
  • Enhancing the overall performance of machine learning models

Benefits

The benefits of this technology include:

  • Increased efficiency in training models
  • Enhanced data quality
  • Improved decision-making based on more accurate predictions

Potential Commercial Applications

The technology could be utilized in industries such as:

  • Healthcare for patient diagnosis
  • Finance for fraud detection
  • E-commerce for personalized recommendations

Possible Prior Art

One possible prior art could be the use of ensemble learning techniques to improve model predictions by combining multiple models' outputs.

What are the potential limitations of this technology?

This technology may have limitations in:

  • Handling highly complex datasets
  • Scalability for large-scale applications

How does this technology compare to existing methods for label correction in machine learning?

This technology stands out by:

  • Automatically correcting incorrect labels based on confidence levels
  • Training models to estimate correct labels for data samples


Original Abstract Submitted

a processor-implemented method includes training a first model to predict confidences of labels for data samples in a training dataset, including using a corrected data sample obtained by correcting an incorrect label based on a corresponding confidence detected by the first model and an estimated corrected label generated by a second model; training the second model to estimate correct labels for the data samples, including estimating a correct other label corresponding to another incorrect label detected based on a corresponding confidence generated by the first model with respect to the other incorrect label; and automatically correcting the other incorrect label with the estimated correct other label.