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: Jun 17, 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

Enhancing complaint locations prediction with image-space embedding representations and customized large language

Theng-Jia Law1, Choo-Yee Ting2, Hu Ng1

  • 1Centre of Big Data and Blockchain Technologies, CoE for Advanced Cloud, Faculty of Computing and Informatics, Multimedia University, Persiaran Multimedia, 63100, Cyberjaya, Selangor, Malaysia.

Scientific Reports
|June 15, 2026
PubMed
Summary

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

A failure in decryption process for bivariate polynomial reconstruction problem cryptosystem.

Heliyon·2024
Same author

Cryptanalysis of the SHMW signature scheme.

Heliyon·2024
Same author

Driving event recognition using machine learning and smartphones.

F1000Research·2023
Same author

A Machine Learning Approach to Predictive Modelling of Student Performance.

F1000Research·2022
Same author

Modelling sentiments based on objectivity and subjectivity with self-attention mechanisms.

F1000Research·2022
Same author

Risk stratification and assessment framework for international travel and border measures amidst the COVID-19 pandemic - A Malaysian perspective.

Travel medicine and infectious disease·2022

This study introduces an image-space embedding framework for predicting public complaint locations. The novel approach enhances urban management by improving spatial analysis and LLM integration for more accurate complaint prediction.

Area of Science:

  • Urban planning and management
  • Artificial intelligence and machine learning
  • Geospatial data analysis

Background:

  • Predicting public complaint locations is crucial for effective urban management but is hindered by complex urban data structures.
  • Current large language models (LLMs) struggle with spatial relationships when converting structured data to text.

Purpose of the Study:

  • To develop an innovative image-space embedding framework for enhanced complaint location prediction.
  • To integrate structured urban data with customized LLMs for improved spatial analysis.

Main Methods:

  • Transformed heterogeneous structured urban data into multi-channel, image-like representations.
  • Integrated these image-space representations with customized LLMs, replacing token embeddings with convolutional neural networks.
Keywords:
Convolutional layersCustomized large language modelsFine-tuningImage-space embeddingsPublic complaint predictionStructured urban data

Related Experiment Videos

Last Updated: Jun 17, 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

  • Compared the proposed framework against text-based LLM pipelines using 48,103 complaint records from Seberang Perai, Malaysia.

    • Main Results:

    • The image-space framework combined with customized CerebrasGPT-590M achieved the highest accuracy (74.2%), Cohen's kappa (71.8%), and F1-score (65.6%).
    • The proposed method required significantly less training time (19.531 minutes) compared to fully fine-tuned baseline LLMs.
    • Performance gains confirmed the effectiveness of image-space representations for complaint prediction.

    Conclusions:

    • The proposed image-space embedding framework offers an efficient solution for predicting complaint locations.
    • This approach advances the integration of structured data with LLMs in image space for urban management applications.