Improving Markov Chain Monte Carlo Model Search for Data Mining

By Paolo Giudici, Robert Castelo. 2004

212 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Statistics and Probability, Probabilistic & Bayesian Methods

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Bernhard Schölkopf
2. Yoshua Bengio
3. Simon Lacoste-Julien
4. Stefan Bauer
5. Sergey Levine
6. Kevin P. Murphy
7. Francis R. Bach
8. Alex Smola

In Statistics and Probability

1. Stefan Bauer
2. Giusi Moffa
3. Jack Kuipers
4. Yoshua Bengio
5. Kevin P. Murphy
6. Simon Lacoste-Julien
7. Bernhard Schölkopf
8. Daniel Eaton

In Probabilistic & Bayesian Methods

1. Yoshua Bengio
2. Stefan Bauer
3. Simon Lacoste-Julien
4. Bernhard Schölkopf
5. Tristan Deleu
6. Qiang Shen
7. Stuart Aitken
8. Rónán Daly

Citations in the corpus

• Bayesian Structure Learning with Generative Flow Networks (2022, methodology): “Fortu-nately, we can filter out invalid actions using some mask m associated to the graph, built from the adjacency matrix of G and the transitive closure of its transpose, and that can be updated efficiently after the addition an edge (Giudici and Castelo, 2003).”
• Efficient Sampling and Structure Learning of Bayesian Networks (2018, unknown): “Other MCMC alternatives include large scale edge reversal (Grzegorczyk and Husmeier, 2008) and Gibbs sampling (Goudie and Mukherjee, 2016) moves.”
• On Inclusion-Driven Learning of Bayesian Networks (2003, methodology): “We assess the convergence rates of the e MC 3 and MC 3 algorithms by using several convergence diagnostics introduced by Giudici and Castelo (2003).”
• Learning Bayesian networks: approaches and issues (2011, methodology): “Giudici et al. (1999) provide a more efficient procedure to sample the chain, while Giudici and Castelo (2003) extend MC3 by using different moves in the state space and provide an analysis of the distribution of various domains.”
• Bayesian structure learning using dynamic programming and MCMC (2007, methodology): “In this paper, we propose to overcome both of these drawbacks by using the DP algorithm as a proposal distribution for Metropolis Hastings in the space of DAGs.”
• DiBS: Differentiable Bayesian Structure Learning (2021, methodology): “For MCMC, we only consider structure MCMC (MC(3)) [37] as a comparison.”
• Partition MCMC for Inference on Acyclic Digraphs (2015, methodology): “Moreover, rejection sampling from a larger fixed sized neighbourhood can be employed to avoid explicitly calculating the neighbourhood, which can further speed up the implementation (Giudici and Castelo, 2003).”