Variational Optimization of Annealing Schedules
This work addresses a specific bottleneck in stochastic modeling for researchers using AIS, offering an incremental improvement over heuristic methods.
The paper tackles the problem of selecting annealing schedules for annealed importance sampling (AIS) to improve partition function estimates, proposing an algorithm that optimizes schedules by minimizing a derived error functional and demonstrating it outperforms conventional schemes with large quantization numbers.
Annealed importance sampling (AIS) is a common algorithm to estimate partition functions of useful stochastic models. One important problem for obtaining accurate AIS estimates is the selection of an annealing schedule. Conventionally, an annealing schedule is often determined heuristically or is simply set as a linearly increasing sequence. In this paper, we propose an algorithm for the optimal schedule by deriving a functional that dominates the AIS estimation error and by numerically minimizing this functional. We experimentally demonstrate that the proposed algorithm mostly outperforms conventional scheduling schemes with large quantization numbers.