Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Oct 5, 2025

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

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

Published on: December 6, 2024

Semi-Supervised Heterogeneous Domain Adaptation: Theory and Algorithms.

Zhen Fang, Jie Lu, Feng Liu

IEEE Transactions on Pattern Analysis and Machine Intelligence

|January 27, 2022

Summary

This study introduces a theoretical framework for semi-supervised heterogeneous domain adaptation (SsHeDA), explaining how source data improves target domain classification. Two new algorithms, KHDA and JMEA, are proposed to leverage this theory for better adaptation.

Related Concept Videos

You might also read

Related Articles

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

Sort by

Same author

Effect of Medicaid coverage of tobacco-dependence treatments on smoking cessation.

International journal of environmental research and public health·2010

Same author

Cytokine and autoantibody patterns in acute liver failure.

Journal of immunotoxicology·2009

Same author

A novel scoring system for prognostic prediction in d-galactosamine/lipopolysaccharide-induced fulminant hepatic failure BALB/c mice.

BMC gastroenterology·2009

Same author

Mammalian target of rapamycin signaling pathway contributes to glioma progression and patients' prognosis.

The Journal of surgical research·2009

Same author

Estrogen receptor neurobiology and its potential for translation into broad spectrum therapeutics for CNS disorders.

Current molecular pharmacology·2009

Same author

Transcriptional and post-translational regulation of adiponectin.

The Biochemical journal·2009

Area of Science:

Machine Learning
Computer Science
Artificial Intelligence

Background:

Semi-supervised heterogeneous domain adaptation (SsHeDA) addresses classification with limited labeled target data by using a separate source domain.
Existing SsHeDA methods lack a theoretical basis for their effectiveness.
Bridging the domain gap with limited labeled data is a key challenge.

Purpose of the Study:

To provide a theoretical foundation for understanding SsHeDA.
To explain how labeled source and unlabeled target data reduce target risk.
To develop novel algorithms guided by this theory.

Main Methods:

Leveraging the compatibility condition from semi-supervised probably approximately correct (PAC) theory.
Proving generalization error bounds for SsHeDA.

Related Experiment Videos

Last Updated: Oct 5, 2025

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

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

Published on: December 6, 2024

Developing two algorithms: kernel heterogeneous domain alignment (KHDA) and joint mean embedding alignment (JMEA).

Main Results:

KHDA offers faster training on smaller datasets.
JMEA achieves higher accuracy on larger datasets.
Both algorithms demonstrate strong performance in image and text classification tasks.

Conclusions:

The proposed theory provides a principled understanding of SsHeDA.
KHDA and JMEA represent effective algorithmic solutions for SsHeDA.
The developed methods outperform existing baselines in accuracy and efficiency.