Search

Insect-pests significantly impact global agricultural productivity and quality. Effective management involves identifying the full insect community, including beneficial insects and harmful pests, to develop and implement integrated pest management strategies. Automated identification of insects under real-world conditions presents several challenges, including differentiating similar-looking species, intra-species dissimilarity and inter-species similarity, several life cycle stages, camouflage, diverse imaging conditions, and variability in insect orientation. A deep-learning model, InsectNet, is proposed to address these challenges. InsectNet is endowed with five key features: (a) utilization of a large dataset of insect images collected through citizen science; (b) label-free self-supervised learning for large models; (c) improving prediction accuracy for species with a small sample size; (d) enhancing model trustworthiness; and (e) democratizing access through streamlined MLOps. This approach allows accurate identification (>96% accuracy) of over 2500 insect species, including pollinator (e.g., butterflies, bees), parasitoid (e.g., some wasps and flies), predator species (e.g., lady beetles, mantises, dragonflies) and harmful pest species (e.g., armyworms, cutworms, grasshoppers, stink bugs). InsectNet can identify invasive species, provide fine-grained insect species identification, and work effectively in challenging backgrounds. It also can abstain from making predictions when uncertain, facilitating seamless human intervention and making it a practical and trustworthy tool. InsectNet can guide citizen science data collection, especially for invasive species where early detection is crucial. Similar approaches may transform other agricultural challenges like disease detection and underscore the importance of data collection, particularly through citizen science efforts..

We present a high-level domain-specific language (DSL) interface to drive an adaptive incomplete $k$-d tree-based framework for finite element (FEM) solutions to PDEs. This DSL provides three key advances: (a) it abstracts out the complexity of implementing non-trivial FEM formulations, (b) it simplifies deploying these formulations on arbitrarily complicated and adaptively refined meshes, and (c) it exhibits good parallel performance. Taken together, the DSL interface allows end-users to rapidly and efficiently prototype new mathematical approaches, and deploy them on large clusters for solving practical problems. We illustrate this DSL by implementing a workflow for solving PDEs using the recently developed shifted boundary method (SBM). The SBM requires approximating relatively complicated integrals over boundary surfaces. Using a high-level DSL greatly simplifies this process and allows rapid exploration of variations. We demonstrate these tools on a variety of 2-D and 3-D configurations. With fewer than 20 lines of input, we can produce a parallel code that scales well to thousands of processes. This generated code is made accessible and readable to be easily modified and hand-tuned, making this tool useful even to experts with the target software.

IntroductionComputer vision and deep learning (DL) techniques have succeeded in a wide range of diverse fields. Recently, these techniques have been successfully deployed in plant science applications to address food security, productivity, and environmental sustainability problems for a growing global population. However, training these DL models often necessitates the large-scale manual annotation of data which frequently becomes a tedious and time-and-resource- intensive process. Recent advances in self-supervised learning (SSL) methods have proven instrumental in overcoming these obstacles, using purely unlabeled datasets to pre-train DL models.MethodsHere, we implement the popular self-supervised contrastive learning methods of NNCLR Nearest neighbor Contrastive Learning of visual Representations) and SimCLR (Simple framework for Contrastive Learning of visual Representations) for the classification of spatial orientation and segmentation of embryos of maize kernels. Maize kernels are imaged using a commercial high-throughput imaging system. This image data is often used in multiple downstream applications across both production and breeding applications, for instance, sorting for oil content based on segmenting and quantifying the scutellum’s size and for classifying haploid and diploid kernels.Results and discussionWe show that in both classification and segmentation problems, SSL techniques outperform their purely supervised transfer learning-based counterparts and are significantly more annotation efficient. Additionally, we show that a single SSL pre-trained model can be efficiently finetuned for both classification and segmentation, indicating good transferability across multiple downstream applications. Segmentation models with SSL-pretrained backbones produce DICE similarity coefficients of 0.81, higher than the 0.78 and 0.73 of those with ImageNet-pretrained and randomly initialized backbones, respectively. We observe that finetuning classification and segmentation models on as little as 1% annotation produces competitive results. These results show SSL provides a meaningful step forward in data efficiency with agricultural deep learning and computer vision.

Finite element modeling of charged species transport has enabled the analysis, design, and optimization of a diverse array of electrochemical and electrokinetic devices. These systems are represented by the Poisson-Nernst-Planck (PNP) equations coupled with the Navier-Stokes (NS) equation. Direct numerical simulation (DNS) to accurately capture the spatio-temporal variation of ion concentration and current flux remains challenging due to the (a) small critical dimension of the electric double layer (EDL), (b) stiff coupling, large advective effects, and steep gradients close to boundaries, and (c) complex geometries exhibited by electrochemical devices. In the current study, we address these challenges by presenting a direct numerical simulation framework that incorporates: (a) a variational multiscale (VMS) treatment, (b) a block-iterative strategy in conjunction with semi-implicit (for NS) and implicit (for PNP) time integrators, and (c) octree based adaptive mesh refinement. The VMS formulation provides numerical stabilization critical for capturing the electro-convective instabilities often observed in engineered devices. The block-iterative strategy decouples the difficulty of non-linear coupling between the NS and PNP equations and allows using tailored numerical schemes separately for NS and PNP equations. The carefully designed second-order, hybrid implicit methods circumvent the harsh timestep requirements of explicit time steppers, thus enabling simulations over longer time horizons. Finally, the octree-based meshing allows efficient and targeted spatial resolution of the EDL. These features are incorporated into a massively parallel computational framework, enabling the simulation of realistic engineering electrochemical devices. The numerical framework is illustrated using several challenging canonical examples.

Deep learning-based approaches have produced models with good insect classification accuracy; Most of these models are conducive for application in controlled environmental conditions. One of the primary emphasis of researchers is to implement identification and classification models in the real agriculture fields, which is challenging because input images that are wildly out of the distribution (e.g., images like vehicles, animals, humans, or a blurred image of an insect or insect class that is not yet trained on) can produce an incorrect insect classification. Out-of-distribution (OOD) detection algorithms provide an exciting avenue to overcome these challenge as it ensures that a model abstains from making incorrect classification prediction of non-insect and/or untrained insect class images. We generate and evaluate the performance of state-of-the-art OOD algorithms on insect detection classifiers. These algorithms represent a diversity of methods for addressing an OOD problem. Specifically, we focus on extrusive algorithms, i.e., algorithms that wrap around a well-trained classifier without the need for additional co-training. We compared three OOD detection algorithms: (i) Maximum Softmax Probability, which uses the softmax value as a confidence score, (ii) Mahalanobis distance-based algorithm, which uses a generative classification approach; and (iii) Energy-Based algorithm that maps the input data to a scalar value, called energy. We performed an extensive series of evaluations of these OOD algorithms across three performance axes: (a) \textit{Base model accuracy}: How does the accuracy of the classifier impact OOD performance? (b) How does the \textit{level of dissimilarity to the domain} impact OOD performance? and (c) \textit{Data imbalance}: How sensitive is OOD performance to the imbalance in per-class sample size?

The accurate and efficient simulation of Partial Differential Equations (PDEs) in and around arbitrarily defined geometries is critical for many application domains. Immersed boundary methods (IBMs) alleviate the usually laborious and time-consuming process of creating body-fitted meshes around complex geometry models (described by CAD or other representations, e.g., STL, point clouds), especially when high levels of mesh adaptivity are required. In this work, we advance the field of IBM in the context of the recently developed Shifted Boundary Method (SBM). In the SBM, the location where boundary conditions are enforced is shifted from the actual boundary of the immersed object to a nearby surrogate boundary, and boundary conditions are corrected utilizing Taylor expansions. This approach allows choosing surrogate boundaries that conform to a Cartesian mesh without losing accuracy or stability. Our contributions in this work are as follows: (a) we show that the SBM numerical error can be greatly reduced by an optimal choice of the surrogate boundary, (b) we mathematically prove the optimal convergence of the SBM for this optimal choice of the surrogate boundary, (c) we deploy the SBM on massively parallel octree meshes, including algorithmic advances to handle incomplete octrees, and (d) we showcase the applicability of these approaches with a wide variety of simulations involving complex shapes, sharp corners, and different topologies. Specific emphasis is given to Poisson's equation and the linear elasticity equations.

We present a full space-time numerical solution of the advection-diffusion equation using a continuous Galerkin finite element method. The Galerkin/least-square method is employed to ensure stability of the discrete variational problem. In the full space-time formulation, time is considered another dimension, and the time derivative is interpreted as an additional advection term of the field variable. We derive a priori error estimates and illustrate spatio-temporal convergence with several numerical examples. We also derive a posteriori error estimates, which coupled with adaptive space-time mesh refinement provide efficient and accurate solutions. The accuracy of the space-time solutions is illustrated against analytical solutions as well as against numerical solutions using a conventional time-marching algorithm.

Accurate simulation of the printing process is essential for improving print quality, reducing waste, and optimizing the printing parameters of extrusion-based additive manufacturing. Traditional additive manufacturing simulations are very compute-intensive and are not scalable to simulate even moderately-sized geometries. In this paper, we propose a general framework for creating a digital twin of the dynamic printing process by performing physics simulations with the intermediate print geometries. Our framework takes a general extrusion-based additive manufacturing G-code, generates an analysis-suitable voxelized geometry representation from the print schedule, and performs physics-based (transient thermal and phase change) simulations of the printing process. Our approach leverages parallel adaptive octree meshes for both voxelated geometry representation as well as for fast simulations to address real-time predictions. We demonstrate the effectiveness of our method by simulating the printing of complex geometries at high voxel resolutions with both sparse and dense infills. Our results show that this approach scales to high voxel resolutions and can predict the transient heat distribution as the print progresses. This work lays the computational and algorithmic foundations for building real-time digital twins and performing rapid virtual print sequence exploration to improve print quality and further reduce material waste., Comment: 13 pages

Many physical systems have underlying safety considerations that require that the strategy deployed ensures the satisfaction of a set of constraints. Further, often we have only partial information on the state of the system. We study the problem of safe real-time decision making under uncertainty. In this paper, we formulate a conservative stochastic contextual bandit formulation for real-time decision making when an adversary chooses a distribution on the set of possible contexts and the learner is subject to certain safety/performance constraints. The learner observes only the context distribution and the exact context is unknown, and the goal is to develop an algorithm that selects a sequence of optimal actions to maximize the cumulative reward without violating the safety constraints at any time step. By leveraging the UCB algorithm for this setting, we propose a conservative linear UCB algorithm for stochastic bandits with context distribution. We prove an upper bound on the regret of the algorithm and show that it can be decomposed into three terms: (i) an upper bound for the regret of the standard linear UCB algorithm, (ii) a constant term (independent of time horizon) that accounts for the loss of being conservative in order to satisfy the safety constraint, and (ii) a constant term (independent of time horizon) that accounts for the loss for the contexts being unknown and only the distribution being known. To validate the performance of our approach we perform extensive simulations on synthetic data and on real-world maize data collected through the Genomes to Fields (G2F) initiative.

Neural network-based approaches for solving partial differential equations (PDEs) have recently received special attention. However, the large majority of neural PDE solvers only apply to rectilinear domains, and do not systematically address the imposition of Dirichlet/Neumann boundary conditions over irregular domain boundaries. In this paper, we present a framework to neurally solve partial differential equations over domains with irregularly shaped (non-rectilinear) geometric boundaries. Our network takes in the shape of the domain as an input (represented using an unstructured point cloud, or any other parametric representation such as Non-Uniform Rational B-Splines) and is able to generalize to novel (unseen) irregular domains; the key technical ingredient to realizing this model is a novel approach for identifying the interior and exterior of the computational grid in a differentiable manner. We also perform a careful error analysis which reveals theoretical insights into several sources of error incurred in the model-building process. Finally, we showcase a wide variety of applications, along with favorable comparisons with ground truth solutions.

We present a technique to perform interface-resolved simulations of complex breakup dynamics in two-phase flows using the Cahn-Hilliard Navier-Stokes equations. The method dynamically decreases the interface thickness parameter in relevant regions and simultaneously increases local mesh resolution, preventing numerical artifacts. We perform a detailed numerical simulation of pulsed jet atomization that shows a complex cascade of break-up mechanisms involving sheet rupture and filament formation. To understand the effect of refinement on the breakup, we analyze the droplet size distribution. The proposed approach opens up resolved simulations for various multiphase flow phenomena., Comment: 6 pages, 5 figures

High-fidelity flow simulations are indispensable when analyzing systems exhibiting multiphase flow phenomena. The accuracy of multiphase flow simulations is strongly contingent upon the finest mesh resolution used to represent the fluid-fluid interfaces. However, the increased resolution comes at a higher computational cost. In this work, we propose algorithmic advances that aim to reduce the computational cost without compromising on the physics by selectively detecting key regions of interest (droplets/filaments) that require significantly higher resolution. The framework uses an adaptive octree-based meshing framework that is integrated with PETSc's linear algebra solvers. We demonstrate scaling of the framework up to 114,688 processes on TACC's Frontera. Finally, we deploy the framework to simulate one of the most resolved simulations of primary jet atomization. This simulation -- equivalent to 35 trillion grid points on a uniform grid -- is 64 times larger than the current state-of-the-art simulations and provides unprecedented insights into an important flow physics problem with a diverse array of engineering applications., Comment: 12 pages, 9 figures; Accepted for publication in Proceedings of 2023 IEEE International Parallel and Distributed Processing Symposium (IPDPS)

Linking the highly complex morphology of organic photovoltaic (OPV) thin films to their charge transport properties is critical for achieving high performance material system that serves as a cost-efficient approach for energy harvesting. In this paper, a novel unsupervised feature engineering framework is developed and used to establish reduced-order structure-property linkages for OPV films. This framework takes advantage of digital image processing algorithms to identify the salient material features of OPVs undergoing the charge transport phenomenon. These material states are then used to obtain a low-dimensional representation of OPV microstructures via 2-point spatial correlations and principal component analysis. It is found that in addition to the material PC scores, two distance-based metrics are required to complete the microstructure quantification of complex OPVs. A localized version of the Gaussian process (laGP) is then used to link the material PC scores as well as the two distance-based metrics to the short-circuit current of OPVs. It is demonstrated that the unsupervised feature engineering framework presented in this paper in conjunction with the laGP can lead to high-fidelity and accurate data-driven structure-property linkages for OPV films.

The paradigm of differentiable programming has significantly enhanced the scope of machine learning via the judicious use of gradient-based optimization. However, standard differentiable programming methods (such as autodiff) typically require that the machine learning models be differentiable, limiting their applicability. Our goal in this paper is to use a new, principled approach to extend gradient-based optimization to functions well modeled by splines, which encompass a large family of piecewise polynomial models. We derive the form of the (weak) Jacobian of such functions and show that it exhibits a block-sparse structure that can be computed implicitly and efficiently. Overall, we show that leveraging this redesigned Jacobian in the form of a differentiable "layer" in predictive models leads to improved performance in diverse applications such as image segmentation, 3D point cloud reconstruction, and finite element analysis., 9 pages, accepted in Neurips 2021

We consider a mesh-based approach for training a neural network to produce field predictions of solutions to parametric partial differential equations (PDEs). This approach contrasts current approaches for "neural PDE solvers" that employ collocation-based methods to make point-wise predictions of solutions to PDEs. This approach has the advantage of naturally enforcing different boundary conditions as well as ease of invoking well-developed PDE theory -- including analysis of numerical stability and convergence -- to obtain capacity bounds for our proposed neural networks in discretized domains. We explore our mesh-based strategy, called NeuFENet, using a weighted Galerkin loss function based on the Finite Element Method (FEM) on a parametric elliptic PDE. The weighted Galerkin loss (FEM loss) is similar to an energy functional that produces improved solutions, satisfies a priori mesh convergence, and can model Dirichlet and Neumann boundary conditions. We prove theoretically, and illustrate with experiments, convergence results analogous to mesh convergence analysis deployed in finite element solutions to PDEs. These results suggest that a mesh-based neural network approach serves as a promising approach for solving parametric PDEs with theoretical bounds.

Complex flow simulations are conventionally performed on HPC clusters. However, the limited availability of HPC resources and steep learning curve of executing on traditional supercomputer infrastructure has drawn attention towards deploying flow simulation software on the cloud. We showcase how a complex computational framework -- that can evaluate COVID-19 transmission risk in various indoor classroom scenarios -- can be abstracted and deployed on cloud services. The availability of such cloud-based personalized planning tools can enable educational institutions, medical institutions, public sector workers (courthouses, police stations, airports, etc.), and other entities to comprehensively evaluate various in-person interaction scenarios for transmission risk. We deploy the simulation framework on the Azure cloud framework, utilizing the Dendro-kT mesh generation tool and PETSc solvers. The cloud abstraction is provided by RocketML cloud infrastructure. We compare the performance of the cloud machines with state-of-the-art HPC machine TACC Frontera. Our results suggest that cloud-based HPC resources are a viable strategy for a diverse array of end-users to rapidly and efficiently deploy simulation software., Comment: Accepted for publication at SuperCompCloud: 5th Workshop on Interoperability of Supercomputing and Cloud Technologies

The Cahn-Hilliard Navier-Stokes (CHNS) system provides a computationally tractable model that can be used to effectively capture interfacial dynamics in two-phase fluid flows. In this work, we present a semi-implicit, projection-based finite element framework for solving the CHNS system. We use a projection-based semi-implicit time discretization for the Navier-Stokes equation and a fully-implicit time discretization for the Cahn-Hilliard equation. We use a conforming continuous Galerkin (cG) finite element method in space equipped with a residual-based variational multiscale (RBVMS) formulation. Pressure is decoupled using a projection step, which results in two linear positive semi-definite systems for velocity and pressure, instead of the saddle point system of a pressure-stabilized method. All the linear systems are solved using an efficient and scalable algebraic multigrid (AMG) method. We deploy this approach on a massively parallel numerical implementation using parallel octree-based adaptive meshes. The overall approach allows the use of relatively large time steps with much faster time-to-solve than similar fully-implicit methods. We present comprehensive numerical experiments showing detailed comparisons with results from the literature for canonical cases, including the single bubble rise and Rayleigh-Taylor instability., Comment: 51 pages, 18 Figures, version accepted for publication Journal of Computational Physics. arXiv admin note: text overlap with arXiv:2009.06628

Reliable seed yield estimation is an indispensable step in plant breeding programs geared towards cultivar development in major row crops. The objective of this study is to develop a machine learning (ML) approach adept at soybean [\textit{Glycine max} L. (Merr.)] pod counting to enable genotype seed yield rank prediction from in-field video data collected by a ground robot. To meet this goal, we developed a multi-view image-based yield estimation framework utilizing deep learning architectures. Plant images captured from different angles were fused to estimate the yield and subsequently to rank soybean genotypes for application in breeding decisions. We used data from controlled imaging environment in field, as well as from plant breeding test plots in field to demonstrate the efficacy of our framework via comparing performance with manual pod counting and yield estimation. Our results demonstrate the promise of ML models in making breeding decisions with significant reduction of time and human effort, and opening new breeding methods avenues to develop cultivars., 18 pages, 8 figures, and 3 Tables

Recent progress in scientific machine learning (SciML) has opened up the possibility of training novel neural network architectures that solve complex partial differential equations (PDEs). Several (nearly data free) approaches have been recently reported that successfully solve PDEs, with examples including deep feed forward networks, generative networks, and deep encoder-decoder networks. However, practical adoption of these approaches is limited by the difficulty in training these models, especially to make predictions at large output resolutions ($\geq 1024 \times 1024$). Here we report on a software framework for data parallel distributed deep learning that resolves the twin challenges of training these large SciML models - training in reasonable time as well as distributing the storage requirements. Our framework provides several out of the box functionality including (a) loss integrity independent of number of processes, (b) synchronized batch normalization, and (c) distributed higher-order optimization methods. We show excellent scalability of this framework on both cloud as well as HPC clusters, and report on the interplay between bandwidth, network topology and bare metal vs cloud. We deploy this approach to train generative models of sizes hitherto not possible, showing that neural PDE solvers can be viably trained for practical applications. We also demonstrate that distributed higher-order optimization methods are $2-3\times$ faster than stochastic gradient-based methods and provide minimal convergence drift with higher batch-size., 10 pages, 18 figures

Deep learning models have been successfully deployed for a diverse array of image-based plant phenotyping applications including disease detection and classification. However, successful deployment of supervised deep learning models requires large amount of labeled data, which is a significant challenge in plant science (and most biological) domains due to the inherent complexity. Specifically, data annotation is costly, laborious, time consuming and needs domain expertise for phenotyping tasks, especially for diseases. To overcome this challenge, active learning algorithms have been proposed that reduce the amount of labeling needed by deep learning models to achieve good predictive performance. Active learning methods adaptively select samples to annotate using an acquisition function to achieve maximum (classification) performance under a fixed labeling budget. We report the performance of four different active learning methods, (1) Deep Bayesian Active Learning (DBAL), (2) Entropy, (3) Least Confidence, and (4) Coreset, with conventional random sampling-based annotation for two different image-based classification datasets. The first image dataset consists of soybean [Glycine max L. (Merr.)] leaves belonging to eight different soybean stresses and a healthy class, and the second consists of nine different weed species from the field. For a fixed labeling budget, we observed that the classification performance of deep learning models with active learning-based acquisition strategies is better than random sampling-based acquisition for both datasets. The integration of active learning strategies for data annotation can help mitigate labelling challenges in the plant sciences applications particularly where deep domain knowledge is required.

Deep learning techniques have been successfully deployed for automating plant stress identification and quantification. In recent years, there is a growing push towards training models that are interpretable -i.e. that justify their classification decisions by visually highlighting image features that were crucial for classification decisions. The expectation is that trained network models utilize image features that mimic visual cues used by plant pathologists. In this work, we compare some of the most popular interpretability methods: Saliency Maps, SmoothGrad, Guided Backpropogation, Deep Taylor Decomposition, Integrated Gradients, Layer-wise Relevance Propagation and Gradient times Input, for interpreting the deep learning model. We train a DenseNet-121 network for the classification of eight different soybean stresses (biotic and abiotic). Using a dataset consisting of 16,573 RGB images of healthy and stressed soybean leaflets captured under controlled conditions, we obtained an overall classification accuracy of 95.05 \%. For a diverse subset of the test data, we compared the important features with those identified by a human expert. We observed that most interpretability methods identify the infected regions of the leaf as important features for some -- but not all -- of the correctly classified images. For some images, the output of the interpretability methods indicated that spurious feature correlations may have been used to correctly classify them. Although the output explanation maps of these interpretability methods may be different from each other for a given image, we advocate the use of these interpretability methods as `hypothesis generation' mechanisms that can drive scientific insight., Comment: 15 pages, 6 figures

Finite element modeling of charged species transport has enabled analysis, design, and optimization of a diverse array of electrochemical and electrokinetic devices. These systems are represented by the Poisson-Nernst-Planck equations coupled with the Navier-Stokes equation, with a key quantity of interest being the current at the system boundaries. Accurately computing the current flux is challenging due to the small critical dimension of the boundary layers (small Debye layer) that require fine mesh resolution at the boundaries. We resolve this challenge by using the Dirichlet-to-Neumanntransformation to weakly impose the Dirichlet conditions for the Poisson-Nernst-Planck equations. The results obtained with weakly imposed Dirichlet boundary conditions showed excellent agreement with those obtained when conventional boundary conditions with highly resolved mesh we reemployed. Furthermore, the calculated current flux showed faster mesh convergence using weakly imposed conditions compared to the conventionally imposed Dirichlet boundary conditions. We illustrate the approach on canonical 3D problems that otherwise would have been computationally intractable to solve accurately. This approach substantially reduces the computational cost of model-ing electrochemical systems., Comment: 26 pages, 14 figures

Reliable training of generative adversarial networks (GANs) typically require massive datasets in order to model complicated distributions. However, in several applications, training samples obey invariances that are \textit{a priori} known; for example, in complex physics simulations, the training data obey universal laws encoded as well-defined mathematical equations. In this paper, we propose a new generative modeling approach, InvNet, that can efficiently model data spaces with known invariances. We devise an adversarial training algorithm to encode them into data distribution. We validate our framework in three experimental settings: generating images with fixed motifs; solving nonlinear partial differential equations (PDEs); and reconstructing two-phase microstructures with desired statistical properties. We complement our experiments with several theoretical results.

Visual inspections for identifying focusing points in inertial microfluidic flows are prone to misinterpreting stable locations and focusing shifts in the case of non-trivial focusing patterns. We develop and deploy an approach for automating the calculation of focusing patterns for a general channel geometry, and thereby reduce the dependence on empirical/visual procedures to confirm the presence of stable locations. We utilize concepts from interpolation theory (to represent continuous force-fields using discrete points), and stability theory to identify "basins of attraction" and quantitatively identify stable equilibrium points. Our computational experiments reveal that predicting equilibrium points accurately requires upto $\times$10-20 times more refined force-maps that conventionally used, which highlights the spatial resolution required for an accurate representation of cross-sectional forces. These focusing patterns are validated using experimental results for a rectangular channel, and triangular channel with an apex angle of $90^\circ$. We then apply the approach to predict and explain focusing patterns and shifts for a $90^\circ$-isosceles triangular channel across a range of Reynolds numbers for $\frac{a}{H} = 0.4$ (particle-to-channel size ratio). We observe that the predicted focusing patterns match experiments well. The force-maps also reveal certain "clouds" of localized stable points, which aid in explaining the onset of bifurcation observed in experiments. The current algorithm is agnostic to channel cross-sections and straight/curved channels, which could pave way to generating a library of focusing patterns as a function of channel geometry, and $Re$, to assist in design of novel devices for tailored particle-streams., Comment: 30 pages, 50 figures

Microfluidic devices are utilized to control and direct flow behavior in a wide variety of applications, particularly in medical diagnostics. A particularly popular form of microfluidics -- called inertial microfluidic flow sculpting -- involves placing a sequence of pillars to controllably deform an initial flow field into a desired one. Inertial flow sculpting can be formally defined as an inverse problem, where one identifies a sequence of pillars (chosen, with replacement, from a finite set of pillars, each of which produce a specific transformation) whose composite transformation results in a user-defined desired transformation. Endemic to most such problems in engineering, inverse problems are usually quite computationally intractable, with most traditional approaches based on search and optimization strategies. In this paper, we pose this inverse problem as a Reinforcement Learning (RL) problem. We train a DoubleDQN agent to learn from this environment. The results suggest that learning is possible using a DoubleDQN model with the success frequency reaching 90% in 200,000 episodes and the rewards converging. While most of the results are obtained by fixing a particular target flow shape to simplify the learning problem, we later demonstrate how to transfer the learning of an agent based on one target shape to another, i.e. from one design to another and thus be useful for a generic design of a flow shape., Neurips 2018 Deep RL workshop

A key problem in computational material science deals with understanding the effect of material distribution (i.e., microstructure) on material performance. The challenge is to synthesize microstructures, given a finite number of microstructure images, and/or some physical invariances that the microstructure exhibits. Conventional approaches are based on stochastic optimization and are computationally intensive. We introduce three generative models for the fast synthesis of binary microstructure images. The first model is a WGAN model that uses a finite number of training images to synthesize new microstructures that weakly satisfy the physical invariances respected by the original data. The second model explicitly enforces known physical invariances by replacing the traditional discriminator in a GAN with an invariance checker. Our third model combines the first two models to reconstruct microstructures that respect both explicit physics invariances as well as implicit constraints learned from the image data. We illustrate these models by reconstructing two-phase microstructures that exhibit coarsening behavior. The trained models also exhibit interesting latent variable interpolation behavior, and the results indicate considerable promise for enforcing user-defined physics constraints during microstructure synthesis.

PARyOpt is a python based implementation of the Bayesian optimization routine designed for remote and asynchronous function evaluations. Bayesian optimization is especially attractive for computational optimization due to its low cost function footprint as well as the ability to account for uncertainties in data. A key challenge to efficiently deploy any optimization strategy on distributed computing systems is the synchronization step, where data from multiple function calls is assimilated to identify the next campaign of function calls. Bayesian optimization provides an elegant approach to overcome this issue via asynchronous updates. We formulate, develop and implement a parallel, asynchronous variant of Bayesian optimization. The framework is robust and resilient to external failures. We show how such asynchronous evaluations help reduce the total optimization wall clock time for a suite of test problems. Additionally, we show how the software design of the framework allows easy extension to response surface reconstruction (Kriging), providing a high performance software for autonomous exploration. The software is available on PyPI, with examples and documentation., 14 pages, 5 figures, 3 tables

Corn yield prediction is beneficial as it provides valuable information about production and prices prior the harvest. Publicly available high-quality corn yield prediction can help address emergent information asymmetry problems and in doing so improve price efficiency in futures markets. This paper is the first to employ Long Short-Term Memory (LSTM), a special form of Recurrent Neural Network (RNN) method to predict corn yields. A cross sectional time series of county-level corn yield and hourly weather data made the sample space large enough to use deep learning technics. LSTM is efficient in time series prediction with complex inner relations, which makes it suitable for this task. The empirical results from county level data in Iowa show promising predictive power relative to existing survey based methods., 26 pages, 10 figures

Our overarching goal is to develop an accurate and explainable model for plant disease identification using hyperspectral data. Charcoal rot is a soil borne fungal disease that affects the yield of soybean crops worldwide. Hyperspectral images were captured at 240 different wavelengths in the range of 383 - 1032 nm. We developed a 3D Convolutional Neural Network model for soybean charcoal rot disease identification. Our model has classification accuracy of 95.73\% and an infected class F1 score of 0.87. We infer the trained model using saliency map and visualize the most sensitive pixel locations that enable classification. The sensitivity of individual wavelengths for classification was also determined using the saliency map visualization. We identify the most sensitive wavelength as 733 nm using the saliency map visualization. Since the most sensitive wavelength is in the Near Infrared Region(700 - 1000 nm) of the electromagnetic spectrum, which is also the commonly used spectrum region for determining the vegetation health of the plant, we were more confident in the predictions using our model.

Recent advances in efficiency of organic photovoltaics are driven by judicious selection of processing conditions that result in a “desired” morphology. An important theme of morphology research is quantifying the effect of processing conditions on morphology and relating it to device efficiency. State-of-the-art morphology quantification methods provide film-averaged or 2D-projected features that only indirectly correlate with performance, making causal reasoning nontrivial. Accessing the 3D distribution of material, however, provides a means of directly mapping processing to performance. In this paper, two recently developed techniques are integrated—reconstruction of 3D morphology and subsequent conversion into intuitive morphology descriptors —to comprehensively image and quantify morphology. These techniques are applied on films generated by doctor blading and spin coating, additionally investigating the effect of thermal annealing. It is found that morphology of all samples exhibits very high connectivity to electrodes. Not surprisingly, thermal annealing consistently increases the average domain size in the samples, aiding exciton generation. Furthermore, annealing also improves the balance of interfaces, enhancing exciton dissociation. A comparison of morphology descriptors impacting each stage of photophysics (exciton generation, dissociation, and charge transport) reveals that spin-annealed sample exhibits superior morphology-based performance indicators. This suggests substantial room for improvement of blade-based methods (process optimization) for morphology tuning to enhance performance of large area devices.

Availability of an explainable deep learning model that can be applied to practical real world scenarios and in turn, can consistently, rapidly and accurately identify specific and minute traits in applicable fields of biological sciences, is scarce. Here we consider one such real world example viz., accurate identification, classification and quantification of biotic and abiotic stresses in crop research and production. Up until now, this has been predominantly done manually by visual inspection and require specialized training. However, such techniques are hindered by subjectivity resulting from inter- and intra-rater cognitive variability. Here, we demonstrate the ability of a machine learning framework to identify and classify a diverse set of foliar stresses in the soybean plant with remarkable accuracy. We also present an explanation mechanism using gradient-weighted class activation mapping that isolates the visual symptoms used by the model to make predictions. This unsupervised identification of unique visual symptoms for each stress provides a quantitative measure of stress severity, allowing for identification, classification and quantification in one framework. The learnt model appears to be agnostic to species and make good predictions for other (non-soybean) species, demonstrating an ability of transfer learning.

Charcoal rot is a fungal disease that thrives in warm dry conditions and affects the yield of soybeans and other important agronomic crops worldwide. There is a need for robust, automatic and consistent early detection and quantification of disease symptoms which are important in breeding programs for the development of improved cultivars and in crop production for the implementation of disease control measures for yield protection. Current methods of plant disease phenotyping are predominantly visual and hence are slow and prone to human error and variation. There has been increasing interest in hyperspectral imaging applications for early detection of disease symptoms. However, the high dimensionality of hyperspectral data makes it very important to have an efficient analysis pipeline in place for the identification of disease so that effective crop management decisions can be made. The focus of this work is to determine the minimal number of most effective hyperspectral bands that can distinguish between healthy and diseased specimens early in the growing season. Healthy and diseased hyperspectral data cubes were captured at 3, 6, 9, 12, and 15 days after inoculation. We utilized inoculated and control specimens from 4 different genotypes. Each hyperspectral image was captured at 240 different wavelengths in the range of 383 to 1032 nm. We used a combination of genetic algorithm as an optimizer and support vector machines as a classifier for identification of maximally effective band combinations. A binary classification between healthy and infected samples using six selected band combinations obtained a classification accuracy of 97% and a F1 score of 0.97 for the infected class. The results demonstrated that these carefully chosen bands are more informative than RGB images, and could be used in a multispectral camera for remote identification of charcoal rot infection in soybean.

A computational framework integrating optimization algorithms, parallel computing and plant physiology was developed to explore crop ideotype design. The backbone of the framework is a plant physiology model that accurately tracks water use (i.e. a plant hydraulic model) coupled with mass transport (CO2 exchange and transport), energy conversion (leaf temperature due to radiation, convection and mass transfer) and photosynthetic biochemistry of an adult maize plant. For a given trait configuration, soil parameters and hourly weather data, the model computes water use and photosynthetic output over the life of an adult maize plant. We coupled this validated model with a parallel, meta-heuristic optimization algorithm, specifically a genetic algorithm (GA), to identify trait sets (ideotypes) that resulted in desired water use behavior of the adult maize plant. We detail features of the model as well as the implementation details of the coupling with the optimization framework and deployment on high performance computing platforms. We illustrate a representative result of this framework by identifying maize ideotypes with optimized photosynthetic yields using weather and soil conditions corresponding to Davis, CA. Finally, we show how the framework can be used to identify broad ideotype trends that can inform breeding efforts. The developed presented tool has the potential to inform the development of future climate-resilient crops.

Deep learning became the method of choice in recent year for solving a wide variety of predictive analytics tasks. For sequence prediction, recurrent neural networks (RNN) are often the go-to architecture for exploiting sequential information where the output is dependent on previous computation. However, the dependencies of the computation lie in the latent domain which may not be suitable for certain applications involving the prediction of a step-wise transformation sequence that is dependent on the previous computation only in the visible domain. We propose that a hybrid architecture of convolution neural networks (CNN) and stacked autoencoders (SAE) is sufficient to learn a sequence of actions that nonlinearly transforms an input shape or distribution into a target shape or distribution with the same support. While such a framework can be useful in a variety of problems such as robotic path planning, sequential decision-making in games, and identifying material processing pathways to achieve desired microstructures, the application of the framework is exemplified by the control of fluid deformations in a microfluidic channel by deliberately placing a sequence of pillars. Learning of a multistep topological transform has significant implications for rapid advances in material science and biomedical applications.

Crop yield forecasting is the methodology of predicting crop yields prior to harvest. The availability of accurate yield prediction frameworks have enormous implications from multiple standpoints, including impact on the crop commodity futures markets, formulation of agricultural policy, as well as crop insurance rating. The focus of this work is to construct a corn yield predictor at the county scale. Corn yield (forecasting) depends on a complex, interconnected set of variables that include economic, agricultural, management and meteorological factors. Conventional forecasting is either knowledge-based computer programs (that simulate plant-weather-soil-management interactions) coupled with targeted surveys or statistical model based. The former is limited by the need for painstaking calibration, while the latter is limited to univariate analysis or similar simplifying assumptions that fail to capture the complex interdependencies affecting yield. In this paper, we propose a data-driven approach that is "gray box" i.e. that seamlessly utilizes expert knowledge in constructing a statistical network model for corn yield forecasting. Our multivariate gray box model is developed on Bayesian network analysis to build a Directed Acyclic Graph (DAG) between predictors and yield. Starting from a complete graph connecting various carefully chosen variables and yield, expert knowledge is used to prune or strengthen edges connecting variables. Subsequently the structure (connectivity and edge weights) of the DAG that maximizes the likelihood of observing the training data is identified via optimization. We curated an extensive set of historical data (1948-2012) for each of the 99 counties in Iowa as data to train the model., Comment: 8 pages, In Proceedings of the 22nd ACM SIGKDD Workshop on Data Science for Food, Energy and Water , 2016 (San Francisco, CA, USA)

Current wind turbine simulations successfully use turbulence generating tools for modeling behavior. However, they lack the ability to reproduce variabilities in wind dynamics and inherent stochastic structures (like temporal and spatial coherences, sporadic bursts, high shear regions). This necessitates a more realistic parameterization of the wind that encodes location-, topography-, diurnal-, seasonal and stochastic affects. In this work, we develop a hierarchical temporal and spatial decomposition (TSD) of large-scale meteorology data to construct a low-dimensional yet realistic stochastic wind flow model. Starting from a large data set (Crop Wind-energy EXperiment), a low dimensional stochastic parameterization is constructed. The resulting time-resolved stochastic model encodes the statistical behavior exhibited by the actual wind flow. Moreover, the temporal modes encode the variation of wind speed in the mean sense and resolve diurnal variations and temporal correlation while the spatial modes provide deeper insight into spatial coherence of the wind field - which is a key aspect in current wind turbine sizing, design and classification. Comparison of several important turbulent properties between the simulated wind flow and the original dataset show the utility of the framework. We envision this framework as a useful complement to existing wind simulation codes., Comment: 34 pages, 11 figures

The atomic force microscope (AFM) is a versatile, high-resolution tool used to characterize the topography and material properties of a large variety of specimens at nano-scale. The interaction of the micro-cantilever tip with the specimen causes cantilever deflections that are measured by an optical sensing mechanism and subsequently utilized to construct the sample topography. Recent years have seen increased interest in using the AFM to characterize soft specimens like gels and live cells. This remains challenging due to the complex and competing nature of tip-sample interaction forces - large tip-sample interaction force is necessary to achieve good signal-to-noise ratios; However, large force tends to deform and destroy soft samples. In situ estimation of the local tip-sample interaction force is needed to control the AFM cantilever motion and prevent destruction of soft samples while maintaining a good signal-to-noise ratio. This necessitates the ability to rapidly estimate the tip-sample forces from the cantilever deflection during operation. This paper proposes a first approach to a near real-time framework for tip-sample force inversion. We pose the inverse problem of extracting the tip-sample force as an unconstrained optimization problem. A fast, parallel forward solver is developed by utilizing graphical processing units (GPU). This forward solver shows an effective 30000 fold speed-up over a comparable CPU implementation, resulting in milli-second calculation times. The forward solver is coupled with a GPU based particle-swarm optimization implementation. We illustrate the framework on three classes of tip-sample interaction inversions. Each of these inversions is performed in sub-second timings, showing potential for integration with on-line AFM imaging and material characterization

Organic solar cells have the potential for widespread usage due to their promise of low cost, roll-to-roll manufacturability, and mechanical flexibility. However, deployment is impeded by their relatively low power conversion efficiencies. The last decade has seen significant progress in enhancing the power conversion of these devices through various strategies. One such approach is based on morphology control. This is because morphology affects all phenomena involved in solar conversion: light absorption and electron-hole pair (exciton) generation; exciton diffusion and dissociation into free charges; and transport of charges to the electrodes. Progress in experimental characterization and computational modeling now allow reconstruction and imaging of the thin film morphology. This opens up the possibility of rationally linking fabrication with morphology, as well as morphology with performance. In this context, a comprehensive set of computational tools to rapidly quantify and classify the heterogeneous internal structure of thin films will be invaluable in linking process, structure and property. We present a novel graph-based framework to efficiently construct a broad suite of physically meaningful morphology descriptors. These morphology descriptors are further classified according to the physical subprocesses within an organic solar cells. The approach is motivated by the equivalence between a discretized morphology and a labeled, weighted, undirected graph. We utilize this approach to pose key questions related to structure characterization. We subsequently construct estimates and upper bounds of various efficiencies. The approach is showcased by characterizing the effect of thermal annealing on time-evolution of a thin film morphology. We conclude by formulating natural extensions to characterize crystallinity and anisotropy of the morphology using the framework.

We present an adaptive methodology for the solution of (linear and) non-linear time dependent problems that is especially tailored for massively parallel computations. The basic concept is to solve for large blocks of space-time unknowns instead of marching sequentially in time. The methodology is a combination of a computationally efficient implementation of a parallel-in-space time finite element solver coupled with a posteriori space-time error estimates and a parallel mesh generator. While we focus on spatial adaptivity in this work, the methodology enables simultaneous adaptivity in both space and time domains. We explore this basic concept in the context of a variety of time-steppers including ?-schemes and Backward Difference Formulas. We specifically illustrate this framework with applications involving time dependent linear, quasi-linear and semi-linear diffusion equations. We focus on investigating how the coupled space-time refinement indicators for this class of problems aspect spatial adaptivity. Finally, we show good scaling behavior up to 150,000 processors on the NCSA Blue Waters machine. This conceptually simple methodology enables scaling on next generation multi-core machines by simultaneously solving for large number of time-steps, and reduces computational overhead by locally refining spatial blocks that can track localized features. This methodology also opens up the possibility of efficiently incorporating adjoint equations for error estimators and inverse