The data-driven future of high-energy-density physics.

High-energy-density physics is the field of physics concerned with studying matter at extremely high temperatures and densities. Such conditions produce highly nonlinear plasmas, in which several phenomena that can normally be treated independently of one another become strongly coupled. The study of these plasmas is important for our understanding of astrophysics, nuclear fusion and fundamental physics-however, the nonlinearities and strong couplings present in these extreme physical systems makes them very difficult to understand theoretically or to optimize experimentally. Here we argue that machine learning models and data-driven methods are in the process of reshaping our exploration of these extreme systems that have hitherto proved far too nonlinear for human researchers. From a fundamental perspective, our understanding can be improved by the way in which machine learning models can rapidly discover complex interactions in large datasets. From a practical point of view, the newest generation of extreme physics facilities can perform experiments multiple times a second (as opposed to approximately daily), thus moving away from human-based control towards automatic control based on real-time interpretation of diagnostic data and updates of the physics model. To make the most of these emerging opportunities, we suggest proposals for the community in terms of research design, training, best practice and support for synthetic diagnostics and data analysis.

Deep Neural Network Initialization With Decision Trees.

In this paper, a novel, automated process for constructing and initializing deep feedforward neural networks based on decision trees is presented. The proposed algorithm maps a collection of decision trees trained on the data into a collection of initialized neural networks with the structures of the networks determined by the structures of the trees. The tree-informed initialization acts as a warm-start to the neural network training process, resulting in efficiently trained, accurate networks. These models, referred to as "deep jointly informed neural networks" (DJINN), demonstrate high predictive performance for a variety of regression and classification data sets and display comparable performance to Bayesian hyperparameter optimization at a lower computational cost. By combining the user-friendly features of decision tree models with the flexibility and scalability of deep neural networks, DJINN is an attractive algorithm for training predictive models on a wide range of complex data sets.

Response to Comment on "Insulator-metal transition in dense fluid deuterium".

In their comment, Desjarlais et al claim that a small temperature drop occurs after isentropic compression of fluid deuterium through the first-order insulator-metal transition. We show that their calculations do not correspond to the experimental thermodynamic path, and that thermodynamic integrations with parameters from first-principles calculations produce results in agreement with our original estimate of the temperature drop.

Insulator-metal transition in dense fluid deuterium.

Dense fluid metallic hydrogen occupies the interiors of Jupiter, Saturn, and many extrasolar planets, where pressures reach millions of atmospheres. Planetary structure models must describe accurately the transition from the outer molecular envelopes to the interior metallic regions. We report optical measurements of dynamically compressed fluid deuterium to 600 gigapascals (GPa) that reveal an increasing refractive index, the onset of absorption of visible light near 150 GPa, and a transition to metal-like reflectivity (exceeding 30%) near 200 GPa, all at temperatures below 2000 kelvin. Our measurements and analysis address existing discrepancies between static and dynamic experiments for the insulator-metal transition in dense fluid hydrogen isotopes. They also provide new benchmarks for the theoretical calculations used to construct planetary models.