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 Videos

A comparison of algorithms for inference and learning in probabilistic graphical models.

Brendan J Frey1, Nebojsa Jojic

  • 1Electrical and Computer Engineering Department, University of Toronto, 10 King's College Road, Toronto, ON M5S 3G4, Canada. frey@psi.toronto.edu

IEEE Transactions on Pattern Analysis and Machine Intelligence
|September 22, 2005
PubMed
Summary

This research explores graph-based probability models for artificial intelligence, enhancing data analysis and combinatorial problems. It reviews various inference and learning algorithms for complex AI systems.

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

An X-linked long non-coding RNA, PTCHD1-AS, and the core features of autism.

Nature·2026
Same author

Orthrus: toward evolutionary and functional RNA foundation models.

Nature methods·2026
Same author

Echoes in AI: Quantifying lack of plot diversity in LLM outputs.

Proceedings of the National Academy of Sciences of the United States of America·2025
Same author

Orthrus: Towards Evolutionary and Functional RNA Foundation Models.

bioRxiv : the preprint server for biology·2024
Same author

Donor activity is associated with US legislators' attention to political issues.

PloS one·2023
Same author

Compositional Processing Emerges in Neural Networks Solving Math Problems.

CogSci ... Annual Conference of the Cognitive Science Society. Cognitive Science Society (U.S.). Conference·2021

Area of Science:

  • Artificial Intelligence
  • Computer Vision
  • Machine Learning

Background:

  • The rise of big data enables advanced AI reasoning under uncertainty.
  • Current AI excels at pattern classification but struggles with large-scale data decomposition.
  • Automatic scene analysis in computer vision requires decomposing images into components.

Purpose of the Study:

  • To advocate for graph-based probability models in AI.
  • To review and compare exact and approximate inference and learning algorithms.
  • To demonstrate these methods in a computer vision application.

Main Methods:

  • Utilizing graph-based probability models.
  • Reviewing exact inference techniques.
  • Examining approximate algorithms: iterated conditional modes, expectation maximization (EM), Gibbs sampling, mean field, variational methods, and sum-product algorithm.

Related Experiment Videos

  • Applying techniques to a vision model of occluding objects.

    • Main Results:

    • Graph-based models offer effective representations for complex data.
    • Various inference algorithms provide efficient solutions for learning and prediction.
    • Comparison of algorithm performance using a free energy cost function.

    Conclusions:

    • Graph-based probability models are powerful tools for AI reasoning under uncertainty.
    • Efficient inference and learning algorithms are crucial for practical AI applications.
    • The reviewed techniques offer a unified framework for tackling complex problems in computer vision and beyond.