An official website of the United States government
Abstract Gravity waves (GWs) make crucial contributions to the middle atmospheric circulation. Yet, their climate model representation remains inaccurate, leading to key circulation biases. This study introduces a set of three neural networks (NNs) that learn to predict GW fluxes (GWFs) from multiple years of high‐resolution ERA5 reanalysis. The three NNs: a ANN, a ANN‐CNN, and an Attention UNet embed different levels of horizontal nonlocality in their architecture and are capable of representing nonlocal GW effects that are missing from current operational GW parameterizations. The NNs are evaluated offline on both time‐averaged statistics and time‐evolving flux variability. All NNs, especially the Attention UNet, accurately recreate the global GWF distribution in both the troposphere and the stratosphere. Moreover, the Attention UNet most skillfully predicts the transient evolution of GWFs over prominent orographic and nonorographic hotspots, with the model being a close second. Since even ERA5 does not resolve a substantial portion of GWFs, this deficiency is compensated by subsequently applying transfer learning on the ERA5‐trained ML models for GWFs from a 1.4 km global climate model. It is found that the re‐trained models both (a) preserve their learning from ERA5, and (b) learn to appropriately scale the predicted fluxes to account for ERA5's limited resolution. Our results highlight the importance of embedding nonlocal information for a more accurate GWF prediction and establish strategies to complement abundant reanalysis data with limited high‐resolution data to develop machine learning‐driven parameterizations for missing mesoscale processes in climate models.
more »
« less
Can Deep Learning Extract Useful Information about Energy Dissipation and Effective Hydraulic Conductivity from Gridded Conductivity Fields?
We confirm that energy dissipation weighting provides the most accurate approach to determining the effective hydraulic conductivity (Keff) of a binary K grid. A deep learning algorithm (UNET) can infer Keff with extremely high accuracy (R2 > 0.99). The UNET architecture could be trained to infer the energy dissipation weighting pattern from an image of the K distribution, although it was less accurate for cases with highly localized structures that controlled flow. Furthermore, the UNET architecture learned to infer the energy dissipation weighting even if it was not trained directly on this information. However, the weights were represented within the UNET in a way that was not immediately interpretable by a human user. This reiterates the idea that even if ML/DL algorithms are trained to make some hydrologic predictions accurately, they must be designed and trained to provide each user-required output if their results are to be used to improve our understanding of hydrologic systems.
more »
« less
- Award ID(s):
- 1740858
- PAR ID:
- 10294761
- Date Published:
- Journal Name:
- Water
- Volume:
- 13
- Issue:
- 12
- ISSN:
- 2073-4441
- Page Range / eLocation ID:
- 1668
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Risk assessment of infrastructure exposed to ice-rich permafrost hazards is essential for climate change adaptation in the Arctic. As this process requires up-to-date, comprehensive, high-resolution maps of human-built infrastructure, gaps in such geospatial information and knowledge of the applications required to produce it must be addressed. Therefore, this study highlights the ongoing development of a deep learning approach to efficiently map the Arctic built environment by detecting nine different types of structures (detached houses, row houses, multi-story blocks, non-residential buildings, roads, runways, gravel pads, pipelines, and storage tanks) from recently-acquired Maxar commercial satellite imagery (<1 m resolution). We conducted a multi-objective comparison, focusing on generalization performance and computational cost, of nine different semantic segmentation architectures. K-fold cross validation was used to estimate the average F1-score of each architecture and the Friedman Aligned Ranks test with the Bergmann-Hommel posthoc procedure was applied to test for significant differences in generalization performance. ResNet-50-UNet++ performs significantly better than five out of the other eight candidate architectures; no significant difference was found in the pairwise comparisons of ResNet-50-UNet++ to ResNet-50-MANet, ResNet-101-MANet, and ResNet-101-UNet++. We then conducted a high-performance computing scaling experiment to compare the number of service units and runtime required for model inferencing on a hypothetical pan- Arctic scale dataset. We found that the ResNet-50-UNet++ model could save up to ~ 54% on service unit expenditure, or ~ 18% on runtime, when considering operational deployment of our mapping approach. Our results suggest that ResNet-50-UNet++ could be the most suitable architecture (out of the nine that were examined) for deep learning-enabled Arctic infrastructure mapping efforts. Overall, our findings regarding the differences between the examined CNN architectures and our methodological framework for multi-objective architecture comparison can provide a foundation that may propel future pan-Arctic GeoAI mapping efforts of infrastructure.more » « less
-
In this paper, we present a technique for estimating the geometry and reflectance of objects using only a camera, flashlight, and optionally a tripod. We propose a simple data capture technique in which the user goes around the object, illuminating it with a flashlight and capturing only a few images. Our main technical contribution is the introduction of a recursive neural architecture, which can predict geometry and reflectance at 2 k ×2 k resolution given an input image at 2 k ×2 k and estimated geometry and reflectance from the previous step at 2 k−1 ×2 k−1 . This recursive architecture, termed RecNet, is trained with 256×256 resolution but can easily operate on 1024×1024 images during inference. We show that our method produces more accurate surface normal and albedo, especially in regions of specular highlights and cast shadows, compared to previous approaches, given three or fewer input images.more » « less
-
This paper studies information theoretic secure aggregation in federated learning, which involves K distributed nodes and a central server. For security, the server can only recover aggregated updates of locally trained models, without any other information about the local users’ data being leaked. The secure aggregation process typically consists of two phases: the key sharing phase and the model aggregation phase. In previous research, a constraint on keys known as “uncoded groupwise keys” was introduced, and we adopt this constraint during the key sharing phase, where each set of S -users shares an independent key. During the model aggregation phase, each user transmits its encrypted model results to the server. To tolerate user dropouts in secure aggregation (i.e., some users may not respond), where up to K−U users may drop out and the identity of the surviving users is unpredictable in advance, at least two rounds of transmission are required in the model aggregation phase. In the first round, users send the masked models. Then, in the second round, based on the identity of the surviving users after the first round, these surviving users send additional messages that assist the server in decrypting the sum of the users’ trained models. Our goal is to minimize the number of transmissions in the two rounds. Additionally, we consider the potential impact of user collusion, where up to T users may collude with the server. This requires the transmissions to meet stricter security constraints, ensuring that the server cannot learn anything beyond the aggregated model updates, even if it colludes with any set of up to T users. For this more challenging problem, we propose schemes that ensure secure aggregation and achieve the capacity region when S∈{2}∪[K−U+1:K−T] . Experimental results conducted on Tencent Cloud also show that the proposed secure aggregation schemes improve the model aggregation time compared to the benchmark scheme.more » « less
-
Abstract Objective. UNet-based deep-learning (DL) architectures are promising dose engines for traditional linear accelerator (Linac) models. Current UNet-based engines, however, were designed differently with various strategies, making it challenging to fairly compare the results from different studies. The objective of this study is to thoroughly evaluate the performance of UNet-based models on magnetic-resonance (MR)-Linac-based intensity-modulated radiation therapy (IMRT) dose calculations.Approach. The UNet-based models, including the standard-UNet, cascaded-UNet, dense-dilated-UNet, residual-UNet, HD-UNet, and attention-aware-UNet, were implemented. The model input is patient CT and IMRT field dose in water, and the output is patient dose calculated by DL model. The reference dose was calculated by the Monaco Monte Carlo module. Twenty training and ten test cases of prostate patients were included. The accuracy of the DL-calculated doses was measured using gamma analysis, and the calculation efficiency was evaluated by inference time.Results. All the studied models effectively corrected low-accuracy doses in water to high-accuracy patient doses in a magnetic field. The gamma passing rates between reference and DL-calculated doses were over 86% (1%/1 mm), 98% (2%/2 mm), and 99% (3%/3 mm) for all the models. The inference times ranged from 0.03 (graphics processing unit) to 7.5 (central processing unit) seconds. Each model demonstrated different strengths in calculation accuracy and efficiency; Res-UNet achieved the highest accuracy, HD-UNet offered high accuracy with the fewest parameters but the longest inference, dense-dilated-UNet was consistently accurate regardless of model levels, standard-UNet had the shortest inference but relatively lower accuracy, and the others showed average performance. Therefore, the best-performing model would depend on the specific clinical needs and available computational resources.Significance. The feasibility of using common UNet-based models for MR-Linac-based dose calculations has been explored in this study. By using the same model input type, patient training data, and computing environment, a fair assessment of the models’ performance was present.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.3390/w13121668
