1 / 11

Understanding Distinctions in Probabilistic Graphical Models for Knowledge Engineering

This wrap-up explores essential distinctions in probabilistic graphical models, covering template-based versus specific models, directed versus undirected graphs, as well as generative and discriminative approaches. We also delve into variable types, including target, observed, and latent variables, while discussing causal versus non-causal ordering. Important concepts such as parameters, local structure, structured conditional probability distributions (CPDs), and the processes of model testing and iterative refinement are highlighted. This comprehensive overview serves as a foundation for advancements in knowledge engineering.

tanuja
Télécharger la présentation

Understanding Distinctions in Probabilistic Graphical Models for Knowledge Engineering

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Representation Probabilistic Graphical Models Wrapup Knowledge Engineering

  2. Important Distinctions • Template based versus specific • Directed versus undirected • Generative versus discriminative • Hybrids are also common

  3. Important Distinctions Template-based Specific

  4. Important Distinctions Generative Discriminative

  5. Variable Types • Target • Observed • Including complex, constructed features • Latent GMT … W1 W2 W3 Wk

  6. Structure • Causal versus non-causal ordering GMT … … W1 W1 W2 W2 W3 W3 Wk Wk GMT

  7. Extending the Conversation

  8. Parameters: Values • What matters: • Zeros • Orders of magnitude • Relative values • Structured CPDs

  9. Parameters: Local Structure • Table CPDs are the exception

  10. Iterative Refinement • Model testing • Sensitivity analysis for parameters • Error analysis • Add features • Add dependencies

  11. END END END

More Related