Approximate inference in BN

Edit Answers
1234567891011121314151617181920
Across
  1. 2. Process of stepping from one state to another in MCMC.
  2. 4. Statistical measure used to evaluate sample reliability.
  3. 5. Variable whose probability we want to compute.
  4. 7. MCMC method that samples each variable conditioned on others.
  5. 9. Set of all possible assignments of variables.
  6. 10. Observed variable values that guide inference.
  7. 11. Advantage of approximate inference in large networks.
  8. 12. When sampling results stabilize to true distribution.
  9. 17. Difference between estimated and true probabilities.
  10. 20. Sampling method where inconsistent samples are discarded.
Down
  1. 1. Family of algorithms based on repeated random sampling.
  2. 3. Type of inference used when exact methods are costly.
  3. 6. Sampling technique that weights samples based on evidence.
  4. 8. Sequence of states where next depends only on current.
  5. 13. Probability distribution updated after evidence.
  6. 14. Abbreviation: Markov Chain Monte Carlo.
  7. 15. Adjusts the influence of a sample in likelihood weighting.
  8. 16. Initial discarded samples in MCMC to avoid bias.
  9. 18. The process of estimating probabilities in a BN.
  10. 19. Method used to approximate probability distributions.