Particle Gibbs with Ancestor Sampling for Probabilistic Programs
This work addresses a bottleneck in probabilistic programming inference, offering an incremental improvement for researchers and practitioners in this domain.
The paper tackled the problem of particle degeneracy and low effective sample size in particle Markov chain Monte Carlo methods for probabilistic program inference by adapting ancestor resampling to this setting, resulting in nontrivial performance gains.
Particle Markov chain Monte Carlo techniques rank among current state-of-the-art methods for probabilistic program inference. A drawback of these techniques is that they rely on importance resampling, which results in degenerate particle trajectories and a low effective sample size for variables sampled early in a program. We here develop a formalism to adapt ancestor resampling, a technique that mitigates particle degeneracy, to the probabilistic programming setting. We present empirical results that demonstrate nontrivial performance gains.