Engineering topics
Publications and source records attributed to Mueller, Juliane.
Today there has never been a more profound codependence and synergetic convergence, such as the one between the energy and IT sectors, and the industries will need to work together more closely to meet today’s growing data center power demands.
We present ideas from our ongoing work in adaptive computing - an optimization framework that allows us to strategically deploy various fidelity level experiments and simulations to guide decision making. The framework aims to enable uncertainty quantified scale-up of simulations and experiments, which causes increased complexity. A key feature of the framework is the integration of user-specified local model trustworthiness estimates. Adaptive sampling strategies allow us to optimally exploit the multiple fidelity level information and trustworthiness measures to arrive at the best decisions within a highly limited budget of objective function evaluations.
Climate modelling and analysis are facing new demands to enhance projections and climate information. Here, in this study, we argue that now is the time to push the frontiers of machine learning beyond state-of-the-art approaches, not only by developing machine-learning-based Earth system models with greater fidelity, but also by providing new capabilities through emulators for extreme event projections with large ensembles, enhanced detection and attribution methods for extreme events, and advanced climate model analysis and benchmarking. Utilizing this potential requires key machine learning challenges to be addressed, in particular generalization, uncertainty quantification, explainable artificial intelligence and causality. This interdisciplinary effort requires bringing together machine learning and climate scientists, while also leveraging the private sector, to accelerate progress towards actionable climate science.
Many engineering applications require the simultaneous optimization of multiple conflicting objective functions. Often, these objective functions are evaluated using highly accurate computer simulations that are computationally too expensive to be evaluated hundreds or thousands of times during optimization. Thus, the goal is to find good approximations of the Pareto front using as few of these expensive simulations as possible. Here, we describe an optimization approach based on surrogate models and diverse sampling strategies to accelerate the search for the Pareto solutions. We use a separate surrogate model for approximating each objective function and then we use the surrogate models to inform where additional expensive simulations should be run. The surrogate models are updated in an active learning framework whenever new information from the expensive simulations becomes available. The sampling strategies aim at balancing local improvements of the approximate Pareto front and global exploration to identify the extrema and fill in large gaps of the approximate Pareto front. We demonstrate on a large set of benchmark problems the effectiveness of the method for finding good approximations of the Pareto front.
We present our SciDAC FASTMath-HEP partnership results for tuning generative adversarial models (GANs) for high energy physics applications. The GANs are used in hybrid simulations to accelerate otherwise time-consuming computations. We optimize for both, prediction accuracy and variability with the goal to find GAN architectures that are reliable and robust.
Surrogate-based 0-th Order Global Optimization for black-box problems. This software is comprised of active learning algorithms based on surrogate models to solve black-box optimization problems and other scientific applications. See also: https://pypi.org/project/soogo/
This chapter reviews the basics of derivative-free optimization methods based on surrogate models and outlines how these methods can straightforwardly be applied to autonomously steering experimentation. It summarizes general solution approaches that use surrogate models and active learning. Surrogate modeling is often combined with active learning strategies, where in each iteration of the optimization algorithm, the surrogate model is used to identify which new inputs should be evaluated next and given the new input-output pair, the surrogate model is updated. Regardless of feasibility, the surrogate models for the constraints are updated in each iteration of the optimization algorithm, while the surrogate model for the objective function is only updated when a feasible point has been found. Similarly to the case of computationally cheap constraints, the surrogate models of the constraints should be incorporated into the definition of the auxiliary optimization problem that is solved to select new sample points.
Deep Learning (DL) models are increasingly used throughout the sciences. However, their performance and usefulness depend greatly on their architecture which is defined by hyperparameters such as the number of nodes, layers, the learning rate, etc. Tuning these hyperparameters is time-consuming because evaluating their performance requires a lengthy training step. Stochastic optimizers used in training lead to performance variability and potentially prediction reliability issues. In this talk, we will describe an automated optimization method based on surrogate models and active learning strategies for tuning DL model architectures. We take into account the prediction variability with the goal to identify architectures that make reliable and robust predictions. We demonstrate our developments on an application arising in particle physics.
We present a new software, HYPPO, that enables the automatic tuning of hyperparameters of various deep learning (DL) models. Unlike other hyperparameter optimization (HPO) methods, HYPPO uses adaptive surrogate models and directly accounts for uncertainty in model predictions to find accurate and reliable models that make robust predictions. Using asynchronous nested parallelism, we are able to significantly alleviate the computational burden of training complex architectures and quantifying the uncertainty. HYPPO is implemented in Python and can be used with both TensorFlow and PyTorch libraries. We demonstrate various software features on time-series prediction and image classification problems as well as a scientific application in computed tomography image reconstruction. Finally, we show that (1) we can reduce by an order of magnitude the number of evaluations necessary to find the most optimal region in the hyperparameter space and (2) we can reduce by two orders of magnitude the throughput for such HPO process to complete.
HYPPO is a hyperparameter optimization (HPO) software that uses surrogate modeling and uncertainty quantification to provide reliable neural network architectures. Similar state-of-the-art HPO technologies do not use such features and mostly perform random searches to explore the parameter space, which can easily become computationally expensive. Using HYPPO, the user is able to reduce by an order of magnitude the number of evaluations necessary to identify the most optimal region in the hyperparameter space. Finally, using asynchronous nested parallelism, we are able to decrease by two orders of magnitude the amount of time required to complete the HPO process.
Focal areas: Primary - (2) Predictive modeling through AI; Secondary - (3) Insight gleaned from complex data (both observed and simulated) using AI
Focus Area: Model predictability improvements (Primary); Data optimization (secondary); Data complexity insights (secondary). The Scientific Challenge: If we believe that a future under extreme conditions will look very differently from today, we can likely agree that ML/AI models trained on past and present datasets will not be adequate to make reliable predictions into the future. This is true for water cycling, as well as biogeochemistry and other Earth system components and behaviors. Additionally, ML/AI models are inherently non-physical. Despite the flourishing success of ML/AI in many applications, such as computer vision, natural language process, and gaming, even the most sophisticated AI models don’t understand the very basic physical laws. Therefore, a natural question is: Can we trust ML/AI based predictions of Earth system behaviors that are fundamentally driven by physical laws? So, are physics models with meticulous process representation a better choice? Not exactly. Physical models, when firmly rooted in first principles, work great at predicting behaviors of systems with a well-defined set of boundary conditions and variables. However, as a complex system, the number of parameters and the degree of complexity and dynamics in processes, coupling, and scale dependent emergent behaviors make the Earth system behaviors very challenging to predict with physical models composed of deterministic laws. In addition, due to the lack of fundamental understandings, physical models often implement empirical correlations derived from observations with biases from locality of data generation. Because correlation is not necessarily causation or comply with first principles, scaling of model predictions beyond locality is often invalid. Beyond the limitation of models, physics or AI, our knowledge of the Earth system is limited by the lack of observational technologies and resources. Insufficient data density, dimensionality and diversity only offer a sliced (or projected) view of the Earth system, e.g. Plato’s Cave analogy, limiting our capability to better understand and represent fundamental processes in models.
This whitepaper is responsive to focal area (2) Predictive modeling through the use of AI techniques and AI-derived model components: the use of AI and other tools to design a prediction system comprising of a hierarchy of models (e.g., AI driven model/component/parameterization selection). Here we describe scale-aware ML models for adaptive model refinement that allow us to bridge the spatial and/or temporal scales in simulation models and observation data for capturing and predicting extreme water cycles.
This whitepaper addresses the following focal area: (1) Data acquisition and assimilation enabled by machine learning, AI, and advanced methods including experimental/network design/optimization, unsupervised learning (including deep learning), and hardware-related efforts involving AI (e.g., edge computing)
This whitepaper is responsive to focal area (1) Data acquisition and assimilation enabled by machine learning (ML), Artificial Intelligence (AI), and advanced methods. Here we describe how Earth observations specific to water cycle disturbances can be collected in parallel with and integrated into future model development, and make use of the latest technologies other than AI/ ML such as 5G/satellite, edge computing, big data technologies, and cloud computing.
This whitepaper addresses to two focal areas – (3) Insight gleaned from complex data using Artificial Intelligence (AI), and other advanced techniques (primary), and (2) Predictive modeling through the use of AI techniques and AI-derived model components (secondary). This topic is directly relevant to four DOE Earth and Environmental Systems Science Division Grand Challenges: integrated water cycle, biogeochemistry, drivers and responses in the Earth system, and data-model integration.
Abstract not provided.
On behalf of the Vehicle Technologies Office of the U.S. Department of Energy, we are pleased to introduce the Fiscal Year (FY) 2019 Annual Progress Report for the Advanced Engine and Fuel Technologies Program. In support of the Vehicle Technology Office’s goal for future U.S. economic growth, the Program focuses on early-stage research and development to improve understanding of combustion processes, fuel properties, and emissions control technologies, generating knowledge and insight necessary for industry to cost-effectively develop the next generation of engines and fuels. One of the most promising and cost-effective approaches to improving the fuel economy of the U.S. vehicle fleet is to introduce the next generation of higher-efficiency, very-low-emission combustion engines that meet future federal emissions regulations into the passenger and commercial vehicle markets. Advanced fuel formulations that can incorporate non-petroleum-based blending agents could further enhance engine efficiency, reduce greenhouse gas emissions, and provide fuel diversification. Also, innovations in combustion, fuels, emissions control, air control, turbomachinery, and energy recovery could potentially increase fuel economy considerably compared to today’s vehicles. The expected national economic, environmental, and energy security benefits from these next-generation engines and fuels would be significant inasmuch as the majority of vehicles sold over the next several decades will still include an engine. The Program has set the following goals for passenger and commercial vehicle fuel economy improvements. By 2030, increase light-duty engine efficiency to demonstrate 35% improvement in passenger vehicle fuel economy (25% improvement from engine efficiency and 10% from fuel co-optimization) relative to a 2015 baseline vehicle, while meeting the U.S. Environmental Protection Agency Tier 3 Emission and Fuel Standards. By 2030, improve heavy-duty engine efficiency by 35% relative to a 2009 baseline vehicle and identify cost-effective high-performance fuels that can further increase efficiency up to an additional 4%, while meeting prevailing U.S. Environmental Protection Agency emissions standards. The Program utilized advanced combustion processes to increase engine efficiency, resulting in a modeled passenger vehicle fuel economy improvement of 19.4% (over a Model Year 2015 baseline) in FY 2019. This report highlights progress achieved by the Advanced Engine and Fuel Technologies Program during FY 2019. The nature, current focus, and recent progress of the Program are described together with summaries of National Laboratory, industry, and university projects that provide an overview of the exciting work being conducted to address critical technical barriers and challenges to commercializing the next generation of higher-efficiency engine, emissions control, and fuel technologies for passenger and commercial vehicles.