Understand
Due to its constrained support, the Dirichlet distribution is uniquely suited to many applications.
- The constraints that make it powerful, however, can also hinder practical implementations, particularly those utilizing Markov Chain Monte Carlo (MCMC) techniques such as Hamiltonian Monte Carlo.
- I demonstrate a series of transformations that reshape the canonical Dirichlet distribution into a form much more amenable to MCMC algorithms.
Built on
R. C. H. Cheng, in Handbook of Simulation: Principles, Methodology, Advances, Applications, and Practice , edited by Banks (Wiley & Sons, New York, 1998)
1998
Earlier work this paper cites.
G. Marsaglia and W. W. Tsang, ACM Transactions on Mathematical Software, 26
2000
Earlier work this paper cites.
S. Hassani, Mathematical Physics: A Modern Introduction to Its Foundations (Springer, New York, 2002)
2002
Earlier work this paper cites.
Similar
D. MacKay, Information Theory, Inference, and Machine Learning (Cambridge University Press, New York, 2003)
2003
Cited alongside, same era.
J. Skilling, in Maximum Entropy and Bayesian methods in science and engineering , American Institute of Physics Conference Series, Vol. 735, edited by G. Erikson, J. T. Bercher, & C. R Smith (AIP Press, 2004) pp. 395–405
2004
Cited alongside, same era.
C. Bishop, Pattern Classification and Machine Learning (Springer, New York, 2007)
2007
Cited alongside, same era.
Then
R. Neal, in Handbook of Markov Chain Monte Carlo , edited by S. Brooks, A. Gelman, G. L. Jones, and X.-L. Meng (CRC Press, New York, 2011)
2011
Closest in time.
2011
Closest in time.
Beyond the bibliography
alphaXiv searches the wider corpus for related work and actual follow-ups.
alphaXiv is searching for related work…