Figures
Mode-hopping Markov Chain Monte Carlo.
This mountain view of the Stubai Alps illustrates a search and optimization strategy for rugged score landscapes. Each green circle represents a graphical model describing the interactions of four components. The height of the circle corresponds to its likelihood score. Mode-hopping constructs a representative chain of models where the decision upon acceptance or rejection of a new proposal is based on the score of the respective nearest local maximum (red circles at the mountain top). This reduces the cost of walking through deep valleys, and increases the search efficiency. See Niederberger et al., doi:10.1371/journal.pcbi.1002568.
Image Credit: Theresa Niederberger, Gene Center, Ludwig-Maximilians-Universität München.
Citation: (2012) PLoS Computational Biology Issue Image | Vol. 8(6) June 2012. PLoS Comput Biol 8(6): ev08.i06. https://doi.org/10.1371/image.pcbi.v08.i06
Published: June 28, 2012
Copyright: © 2012 Niederberger. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
This mountain view of the Stubai Alps illustrates a search and optimization strategy for rugged score landscapes. Each green circle represents a graphical model describing the interactions of four components. The height of the circle corresponds to its likelihood score. Mode-hopping constructs a representative chain of models where the decision upon acceptance or rejection of a new proposal is based on the score of the respective nearest local maximum (red circles at the mountain top). This reduces the cost of walking through deep valleys, and increases the search efficiency. See Niederberger et al., doi:10.1371/journal.pcbi.1002568.
Image Credit: Theresa Niederberger, Gene Center, Ludwig-Maximilians-Universität München.