Search for: All records

Parallel tempering (PT), also known as replica exchange, is the go-to workhorse for simulations of multi-modal distributions. The key to the success of PT is to adopt efficient swap schemes. The popular deterministic even-odd (DEO) scheme exploits the non-reversibility property and has successfully reduced the communication cost from O(P 2) to O(P) given sufficient many P chains. However, such an innovation largely disappears in big data problems due to the limited chains and extremely few bias-corrected swaps. To handle this issue, we generalize the DEO scheme to promote the non-reversibility and obtain an appealing communication cost O(P log P) based on the optimal window size. In addition, we also analyze the bias when we adopt stochastic gradient descent (SGD) with large and constant learning rates as exploration kernels. Such a user-friendly nature enables us to conduct large-scale uncertainty approximation tasks without much tuning costs. 

We propose an interacting contour stochastic gradient Langevin dynamics (IC-SGLD) sampler, an embarrassingly parallel multiple-chain contour stochastic gradient Langevin dynamics (CSGLD) sampler with efficient interactions. We show that ICSGLD can be theoretically more efficient than a single-chain CSGLD with an equivalent computational budget. We also present a novel random-field function, which facilitates the estimation of self-adapting parameters in big data and obtains free mode explorations. Empirically, we compare the proposed algorithm with popular benchmark methods for posterior sampling. The numerical results show a great potential of ICSGLD for large-scale uncertainty estimation tasks. 

In this work, a Pt catalyst supported on an equimolar Al 2 O 3 –CeO 2 binary oxide (Pt–Al–Ce) was prepared and applied in photo-thermo-chemical dry reforming of methane (DRM) driven by concentrated solar irradiation. It was found that the Pt–Al–Ce catalyst showed good stability in DRM reactions and significant enhancements in H 2 and CO production rates compared with Pt/CeO 2 (Pt–Ce) and Pt/Al 2 O 3 (Pt–Al) catalysts. At a reaction temperature of 700 °C under 30-sun equivalent solar irradiation, the Pt–Al–Ce catalyst exhibits a stable DRM catalytic performance at a H 2 production rate of 657 mmol g −1 h −1 and a CO production rate of 666 mmol g −1 h −1 , with the H 2 /CO ratio almost equal to unity. These production rates and the H 2 /CO ratio were significantly higher than those obtained in the dark at the same temperature. The light irradiation was found to induce photocatalytic activities on Pt–Al–Ce and reduce the reaction activation energy. In situ diffuse reflectance infrared Fourier transform spectroscopy ( in situ DRIFTS) was applied to identify the active intermediates in the photo-thermo-chemical DRM process, which were bidentate/monodentate carbonate, absorbed CO on Pt, and formate. The benefits of the binary Al 2 O 3 –CeO 2 substrate could be ascribed to Al 2 O 3 promoting methane dissociation while CeO 2 stabilized and eliminated possible coke formation, leading to high catalytic DRM activity and stability.