Logical Discrete Graphical Models Must Supplement Large Language Models for Information Synthesis.
Gregory CoppolaPublished in: CoRR (2024)
Keyphrases
- graphical models
- language model
- probabilistic model
- document retrieval
- approximate inference
- bayesian networks
- language modeling
- n gram
- information retrieval
- belief propagation
- random variables
- markov networks
- probabilistic inference
- test collection
- background knowledge
- statistical model
- machine learning
- statistical models
- mixture model
- maximum likelihood
- co occurrence
- information extraction
- prior knowledge
- map inference