Timezone: »

Thermostat-assisted continuously-tempered Hamiltonian Monte Carlo for Bayesian learning
Rui Luo · Jianhong Wang · Yaodong Yang · Jun WANG · Zhanxing Zhu

Thu Dec 06 07:45 AM -- 09:45 AM (PST) @ Room 210 #44

In this paper, we propose a novel sampling method, the thermostat-assisted continuously-tempered Hamiltonian Monte Carlo, for the purpose of multimodal Bayesian learning. It simulates a noisy dynamical system by incorporating both a continuously-varying tempering variable and the Nos\'e-Hoover thermostats. A significant benefit is that it is not only able to efficiently generate i.i.d. samples when the underlying posterior distributions are multimodal, but also capable of adaptively neutralising the noise arising from the use of mini-batches. While the properties of the approach have been studied using synthetic datasets, our experiments on three real datasets have also shown its performance gains over several strong baselines for Bayesian learning with various types of neural networks plunged in.

Author Information

Rui Luo (University College London)
Jianhong Wang (UCL)
Yaodong Yang (AIG)
Zhanxing Zhu (Peking University)

More from the Same Authors