Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

System and method for active machine learning

An active learning, processor technology, applied in machine learning, neural learning methods, instruments, etc., can solve difficult, labor-intensive and other problems

Pending Publication Date: 2020-10-30
SAMSUNG ELECTRONICS CO LTD
View PDF0 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Precise annotation of sequences is not only labor-intensive, but may also require very specific domain knowledge, which is not easily accomplished by crowd-sourcing workers

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • System and method for active machine learning
  • System and method for active machine learning
  • System and method for active machine learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0034] The following discussion is described with reference to the accompanying drawings Figures 1 to 11E and various embodiments of the present disclosure. However, it should be understood that the present disclosure is not limited to these embodiments, and all changes and / or their equivalents or substitutions also belong to the scope of the present disclosure.

[0035] Existing active learning strategies rely on uncertainty measures derived from classifiers for query sample selection. Due to the complexity of the label space, these active learning algorithms are not easily scalable to solve sequence learning problems. Consider a tokenized sequence with p tokens, where each token can belong to k possible classes. In view of this, there are kp possible combinations of marker sequences. This complexity grows exponentially with the length of the output.

[0036] Existing active learning methods face two major challenges when dealing with sequence learning tasks: (i) the "co...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

An electronic device for active learning includes at least one memory and at least one processor coupled to at least one memory. At least one processor is configured to select one or more entries froma data set including unlabeled data based on a similarity between the one or more entries and labeled data. At least one processor is further configured to cause the one or more entries to be labeled.

Description

technical field [0001] The present disclosure generally relates to machine learning systems. More specifically, the present disclosure relates to systems and methods for active machine learning. Background technique [0002] Active learning (AL) is a method to solve the supervised learning problem without enough labels. Although active learning solutions for classification problems have been proposed, active learning algorithms for sequences are still not widely discussed. As interest in artificial intelligence has grown, many emerging problems have been defined within the context of sequence learning, including image captioning, machine translation, and natural language understanding. In contrast to classification tasks, which require only one label for an example, sequence learning tasks typically require a series of token-level labels for the entire sequence. Precise annotation of sequences is not only labor-intensive, but may also require very specific domain knowledg...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06N20/00G06V10/774
CPCG06N3/088G06V10/82G06V10/7715G06V20/70G06V10/774G06N3/044G06N3/045G06F18/213G06N20/00G06F18/22G06F18/214
Inventor 邓岳沈逸麟金红霞
Owner SAMSUNG ELECTRONICS CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Eureka Blog
Learn More
PatSnap group products