Gradient-based MCMC samplers for dynamic causal modelling

@inproceedings{Sengupta2016GradientbasedMS,
  title={Gradient-based MCMC samplers for dynamic causal modelling},
  author={Biswa Sengupta and Karl J. Friston and William D. Penny},
  booktitle={NeuroImage},
  year={2016}
}
In this technical note, we derive two MCMC (Markov chain Monte Carlo) samplers for dynamic causal models (DCMs). Specifically, we use (a) Hamiltonian MCMC (HMC-E) where sampling is simulated using Hamilton's equation of motion and (b) Langevin Monte Carlo algorithm (LMC-R and LMC-E) that simulates the Langevin diffusion of samples using gradients either on a Euclidean (E) or on a Riemannian (R) manifold. While LMC-R requires minimal tuning, the implementation of HMC-E is heavily dependent on… CONTINUE READING
Highly Cited
This paper has 35 citations. REVIEW CITATIONS

Citations

Publications citing this paper.
Showing 1-10 of 17 extracted citations

References

Publications referenced by this paper.
Showing 1-10 of 48 references

User Documentation for CVODES, and ODE Solver with Sensitivity Analysis Capabilities

  • A. Hindmarsh, R. Serban
  • Technical Report
  • 2002
Highly Influential
7 Excerpts

2015w. The geometric foundations of Hamiltonian Monte Carlo

  • Betancourt, Michael, +5 authors Mark
  • Stat. Sci. (under review)
  • 2015
1 Excerpt

Optimizing the integrator step size for Hamiltonian Monte Carlo (arXiv:1411.6669)

  • Betancourt, Michael, Byrne, Simon, Girolami, Mark
  • 2015
1 Excerpt

Second order MCMC for dynamic causal modelling (in preparation)

  • Sengupta, Biswa, J. Friston Karl, D. Penny Will
  • 2015
1 Excerpt

Similar Papers

Loading similar papers…