Autonomous efficient experiment design for materials discovery with Bayesian model averaging

@article{Talapatra2018AutonomousEE,
  title={Autonomous efficient experiment design for materials discovery with Bayesian model averaging},
  author={Anjana Talapatra and Shahin Boluki and Thien C. Duong and Xiaoning Qian and Edward R. Dougherty and Raymundo Arr'oyave},
  journal={Physical Review Materials},
  year={2018}
}
The accelerated exploration of the materials space in order to identify configurations with optimal properties is an ongoing challenge. Current paradigms are typically centered around the idea of performing this exploration through high-throughput experimentation/computation. Such approaches, however, do not account fo the always present constraints in resources available. Recently, this problem has been addressed by framing materials discovery as an optimal experiment design. This work… 

Bayesian optimization with adaptive surrogate models for automated experimental design

This paper proposes a fully autonomous experimental design framework that uses more adaptive and flexible Bayesian surrogate models in a BO procedure, namely Bayesian multivariate adaptive regression splines and Bayesian additive regression trees that can overcome the weaknesses of widely used Gaussian process-based methods.

Bayesian Optimization in Materials Science: A Survey

A survey of Bayesian optimization approaches in materials science to increase cross-fertilization and avoid duplication of work is presented and common challenges and opportunities for joint research efforts are highlighted.

Navigating the design space of inorganic materials synthesis using statistical methods and machine learning.

The account collates discussion of design-of-experiments sampling methods, machine learning modeling, and active learning to develop experimental workflows that accelerate the experimental navigation of synthetic landscapes.

Uncertainty-aware Mixed-variable Machine Learning for Materials Design

This work surveys frequentist and Bayesian approaches to uncertainty quantification of machine learning with mixed variables, investigating the machine learning models’ predictive and uncertainty estimation capabilities, and provides interpretations of the observed performance differences.

Experiment Design Frameworks for Accelerated Discovery of Targeted Materials Across Scales

The need for optimal experiment design is discussed, the challenges in it’s implementation are discussed and some successful examples of materials discovery via experiment design are discussed.

Uncertainty Quantification Accounting for Model Discrepancy Within a Random Effects Bayesian Framework

With the advent of integrated computational materials engineering, there is a drive to exchange statistical confidence in a design obtained from repeated experimentation with one developed through
...

References

SHOWING 1-10 OF 84 REFERENCES

Accelerated search for materials with targeted properties by adaptive design

This strategy uses inference and global optimization to balance the trade-off between exploitation and exploration of the search space, and finds very low thermal hysteresis (ΔT) NiTi-based shape memory alloys, with Ti50.0Ni46.7Cu0.8Fe2.3Pd0.2 possessing the smallest ΔT (1.84 K).

Multi-objective Optimization for Materials Discovery via Adaptive Design

This work shows that the Maximin and Centroid design strategies, based on value of information criteria, are more efficient in determining points on the Pareto front from the data than random selection, pure exploitation of the surrogate model prediction or pure exploration by maximum uncertainty from the learning model.

Adaptive Strategies for Materials Design using Uncertainties

Comparing several adaptive design strategies using a data set of 223 M2AX family of compounds for which the elastic properties have been computed using density functional theory finds that selectors that use information about the prediction uncertainty outperform those that don’t.

Bayesian Optimization in a Billion Dimensions via Random Embeddings

Empirical results confirm that REMBO can effectively solve problems with billions of dimensions, provided the intrinsic dimensionality is low, and show thatREMBO achieves state-of-the-art performance in optimizing the 47 discrete parameters of a popular mixed integer linear programming solver.

Efficient Global Optimization of Expensive Black-Box Functions

This paper introduces the reader to a response surface methodology that is especially good at modeling the nonlinear, multimodal functions that often occur in engineering and shows how these approximating functions can be used to construct an efficient global optimization algorithm with a credible stopping rule.

Accelerated Materials Design of Lithium Superionic Conductors Based on First‐Principles Calculations and Machine Learning Algorithms

Concerted efforts continue to be made in the search for superior lithium-ion conducting solids to replace the highly reactive liquid electrolytes typically used in rechargeable batteries.

High Dimensional Bayesian Optimisation and Bandits via Additive Models

It is demonstrated that the method outperforms naive BO on additive functions and on several examples where the function is not additive, and it is proved that, for additive functions the regret has only linear dependence on $D$ even though the function depends on all$D$ dimensions.

Quantifying the Objective Cost of Uncertainty in Complex Dynamical Systems

The proposed Bayesian framework for objective-based uncertainty quantification (UQ), which quantifies the uncertainty in a given system based on the expected increase of the operational cost, provides a practical way of quantifying the effect of various types of system uncertainties on the operation of interest.
...