A5010429017- Machine Learning in Materials Science
- Organic Electronics and Photovoltaics
- Block Copolymer Self-Assembly
- Advanced Memory and Neural Computing
- Rheology and Fluid Dynamics Studies
- Hydrocarbon exploration and reservoir analysis
- Phase Equilibria and Thermodynamics
- Conducting polymers and applications
- Polymer crystallization and properties
- Computer Graphics and Visualization Techniques
- Topology Optimization in Engineering
- Electron and X-Ray Spectroscopy Techniques
- Perovskite Materials and Applications
- Molecular Junctions and Nanostructures
- Silicon and Solar Cell Technologies
- Machine Learning and Data Classification
- Advanced Multi-Objective Optimization Algorithms
- Industrial Vision Systems and Defect Detection
- Thin-Film Transistor Technologies
- Generative Adversarial Networks and Image Synthesis
- Model Reduction and Neural Networks
- Advanced Bandit Algorithms Research
- Numerical Methods and Algorithms
- Advanced Surface Polishing Techniques
- Cell Image Analysis Techniques
Iowa State University
2015-2022
Accelerating the development of π-conjugated molecules for applications such as energy generation and storage, catalysis, sensing, pharmaceuticals, (semi)conducting technologies requires rapid accurate evaluation electronic, redox, or optical properties. While high-throughput computational screening has proven to be a tremendous aid in this regard, machine learning (ML) other data-driven methods can further enable orders magnitude reduction time while at same providing dramatic increases...
Abstract Recent advances in efficiency of organic photovoltaics are driven by judicious selection processing conditions that result a “desired” morphology. An important theme morphology research is quantifying the effect on and relating it to device efficiency. State‐of‐the‐art quantification methods provide film‐averaged or 2D‐projected features only indirectly correlate with performance, making causal reasoning nontrivial. Accessing 3D distribution material, however, provides means...
Reliable training of generative adversarial networks (GANs) typically require massive datasets in order to model complicated distributions. However, several applications, samples obey invariances that are \textit{a priori} known; for example, complex physics simulations, the data universal laws encoded as well-defined mathematical equations. In this paper, we propose a new modeling approach, InvNet, can efficiently spaces with known invariances. We devise an algorithm encode them into...
A key problem in computational material science deals with understanding the effect of distribution (i.e., microstructure) on performance. The challenge is to synthesize microstructures, given a finite number microstructure images, and/or some physical invariances that exhibits. Conventional approaches are based stochastic optimization and computationally intensive. We introduce three generative models for fast synthesis binary images. first model WGAN uses training images new...
Abstract Charge transport in molecular solids, such as semiconducting polymers, is strongly affected by packing and structural order over several length scales. Conventional approaches to modeling these phenomena range from analytical models numerical using quantum mechanical calculations. While cannot account for detailed effects, are expensive exhaustive (and statistically significant) analysis. Here, we report a computationally scalable methodology graph theory explore the influence of...
Generative Adversarial Networks (GANs), while widely successful in modeling complex data distributions, have not yet been sufficiently leveraged scientific computing and design. Reasons for this include the lack of flexibility GANs to represent discrete-valued image data, as well control over physical properties generated samples. We propose a new conditional generative approach (InvNet) that efficiently enables images, allowing their parameterized geometric statistical properties. evaluate...
As new generations of thin-film semiconductors are moving toward solution-based processing, the development printing formulations will require information pertaining to free energies mixing complex mixtures. From standpoint in silico material design, this move necessitates methods that can accurately and quickly evaluate these order maximize processing speed reproducibility. Here, we make use molecular dynamics (MD) simulations, combination with two-phase thermodynamic (2PT) model, explore...
Fibers made from polymer blends have conventionally enjoyed wide use, particularly in textiles. This applicability is primarily aided by the ease of manufacturing such fibers. More recently, ability to tailor internal morphology blend fibers carefully designing processing conditions has enabled be used technologically relevant applications. Some examples include anisotropic insulating properties for heat and wicking moisture, coaxial morphologies optical applications as well with high...
PARyOpt is a python based implementation of the Bayesian optimization routine designed for remote and asynchronous function evaluations. especially attractive computational due to its low cost footprint as well ability account uncertainties in data. A key challenge efficiently deploy any strategy on distributed computing systems synchronization step, where data from multiple calls assimilated identify next campaign calls. provides an elegant approach overcome this issue via updates. We...
PARyOpt 1 is a Python based implementation of the Bayesian optimization routine designed for remote and asynchronous function evaluations. especially attractive computational due to its low cost footprint as well ability account uncertainties in data. A key challenge efficiently deploy any strategy on distributed computing systems synchronization step, where data from multiple calls assimilated identify next campaign calls. provides an elegant approach overcome this issue via updates. We...
Molecular systems are analyzed via the construction of a molecular graph and quantifying resiliency for charge transport through metrics centrality, in context pathways between source drain electrodes.