Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: Apr 25, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

1.3K

Semi-supervised learning of statistical models for natural language understanding.

Deyu Zhou1, Yulan He2

  • 1School of Computer Science and Engineering, Key Laboratory of Computer Network and Information Integration, Ministry of Education, Southeast University, Nanjing 210096, China.

Thescientificworldjournal
|August 26, 2014
PubMed
Summary

This study introduces a new framework for training natural language understanding models using abstract semantic annotations, reducing the need for expensive fully annotated data. The framework significantly improves performance over existing methods.

Related Concept Videos

Language and Cognition01:27

Language and Cognition

874
Language serves as a bridge between ideas and communication, influencing how individuals perceive and interact with the world. Psychologists have long debated whether language shapes thought or vice versa. This discussion gained grip with Edward Sapir and Benjamin Lee Whorf in the 1940s, who proposed that language determines thought, a concept known as linguistic determinism. They suggested that the vocabulary and structure of a language influence how its speakers think and perceive reality.
874

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

PRIMARY-AI: outcomes-based standards to safeguard primary care in the AI era.

Nature medicine·2026
Same author

Open LLM-based actionable incidental finding extraction from [<sup>18</sup>F]fluorodeoxyglucose PET-CT radiology reports.

Frontiers in digital health·2025
Same author

Ultrasound and Freeze-Thaw Pretreatments Alter Cell Wall Structure and Water Status to Improve Vacuum Freeze-Drying Efficiency and Quality Attributes of Plums.

Journal of food science·2025
Same author

Improving Medical Visual Representation Learning With Pathological-Level Cross-Modal Alignment and Correlation Exploration.

IEEE journal of biomedical and health informatics·2025
Same author

SGRRG: Leveraging radiology scene graphs for improved and abnormality-aware radiology report generation.

Computerized medical imaging and graphics : the official journal of the Computerized Medical Imaging Society·2025
Same author

Uncertainty-aware automatic TNM staging classification for [<sup>18</sup>F] Fluorodeoxyglucose PET-CT reports for lung cancer utilising transformer-based language models and multi-task learning.

BMC medical informatics and decision making·2024

Area of Science:

  • Computational Linguistics
  • Artificial Intelligence
  • Machine Learning

Background:

  • Natural language understanding (NLU) aims to create computational models for mapping sentences to semantic representations.
  • Training statistical NLU models typically requires expensive, fully annotated data.
  • Existing methods like the hidden vector state (HVS) model utilize abstract semantic annotations but have limitations.

Purpose of the Study:

  • To propose a novel framework for training statistical NLU models without relying on fully annotated data.
  • To automatically induce derivation rules for mapping sentences to semantic meaning representations.
  • To evaluate the framework's effectiveness using conditional random fields (CRFs) and hidden Markov support vector machines (HM-SVMs).

Main Methods:

Related Experiment Videos

Last Updated: Apr 25, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

1.3K
  • Utilizing a framework that accepts sentences labeled with abstract semantic annotations.
  • These annotations capture underlying semantic structural relations without explicit word/semantic tag alignment.
  • Applying the learning framework to train CRFs and HM-SVMs.

    • Main Results:

    • Both CRFs and HM-SVMs trained with the proposed framework outperformed the baseline HVS model.
    • The framework demonstrated superior performance compared to a hybrid HVS-HM-SVM approach.
    • A relative error reduction rate of approximately 25% and 15% in F-measure was achieved against baseline models.

    Conclusions:

    • The proposed framework effectively trains statistical NLU models using abstract semantic annotations.
    • It offers a cost-effective alternative to fully annotated data, significantly improving performance.
    • The framework provides a robust method for inducing semantic representations in natural language processing.