An official website of the United States government
A state-space model (SSM) integrating physical parameters is proposed and developed in this work, to describe the increase of global average temperature and the subsequent changes in regional climate and hydrology. This SSM approach aims at providing updated and improved forecasts, based on observations and using Bayesian inference, and at facilitating flexible engineering decision-making schemes. Global climate model simulations are used for informing the distribution of the parameters of the SSM. The case study of the Colorado River Basin serves as a preliminary application of the method, to forecast changes in the upper basin natural flow. The method projects that the post-2000 low flow volume will continue, or become even lower on average, although such projections are subject to large uncertainty. Given the increasing need of climate projections in the design, operation, and management of infrastructure, the SSM approach can serve as a useful tool, informed by historical records, to facilitate engineering applications.
more »
« less
Estimating the Spread of Wildland Fires via Evidence-Based Information Fusion
This article presents a new evidential reasoning approach for estimating the state of an evolving wildfire in real time. Given assumed terrain information and localized wind velocity distribution parameters, a probabilistic representation (i.e., the belief state) of a wildfire is forecast across the spatiotemporal domain through a compilation of fire spread simulations. The forecast is updated through information fusion based on observations provided by: 1) embedded temperature sensors and 2) mobile vision agents that are advantageously directed toward locations of information extraction based on the current state estimate. This combination of uncertain sources is performed under the evidence-based Dempster’s rule of combination and is then used to enact sensor reconfiguration based on the updated estimate. This research finds that the evidential belief combination vastly outperforms the standard forecasting approach (where no sensor data are incorporated) in the presence of imprecise environmental parameters.
more »
« less
- Award ID(s):
- 2132798
- PAR ID:
- 10484158
- Publisher / Repository:
- IEEE
- Date Published:
- Journal Name:
- IEEE Transactions on Control Systems Technology
- Volume:
- 31
- Issue:
- 2
- ISSN:
- 1063-6536
- Page Range / eLocation ID:
- 511 to 526
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Deep neural networks (DNNs) have been shown to perform well on exclusive, multi-class classification tasks. However, when different classes have similar visual features, it becomes challenging for human annotators to differentiate them. This scenario necessitates the use of composite class labels. In this paper, we propose a novel framework called Hyper-Evidential Neural Network (HENN) that explicitly models predictive uncertainty due to composite class labels in training data in the context of the belief theory called Subjective Logic (SL). By placing a grouped Dirichlet distribution on the class probabilities, we treat predictions of a neural network as parameters of hyper-subjective opinions and learn the network that collects both single and composite evidence leading to these hyper-opinions by a deterministic DNN from data. We introduce a new uncertainty type called vagueness originally designed for hyper-opinions in SL to quantify composite classification uncertainty for DNNs. Our results demonstrate that HENN outperforms its state-of-the-art counterparts based on four image datasets. The code and datasets are available at: https://github.com/ Hugo101/HyperEvidentialNN.more » « less
-
Guruswami, Venkatesan (Ed.)In decentralized finance ("DeFi"), automated market makers (AMMs) enable traders to programmatically exchange one asset for another. Such trades are enabled by the assets deposited by liquidity providers (LPs). The goal of this paper is to characterize and interpret the optimal (i.e., profit-maximizing) strategy of a monopolist liquidity provider, as a function of that LP’s beliefs about asset prices and trader behavior. We introduce a general framework for reasoning about AMMs based on a Bayesian-like belief inference framework, where LPs maintain an asset price estimate, which is updated by incorporating traders' price estimates. In this model, the market maker (i.e., LP) chooses a demand curve that specifies the quantity of a risky asset to be held at each dollar price. Traders arrive sequentially and submit a price bid that can be interpreted as their estimate of the risky asset price; the AMM responds to this submitted bid with an allocation of the risky asset to the trader, a payment that the trader must pay, and a revised internal estimate for the true asset price. We define an incentive-compatible (IC) AMM as one in which a trader’s optimal strategy is to submit its true estimate of the asset price, and characterize the IC AMMs as those with downward-sloping demand curves and payments defined by a formula familiar from Myerson’s optimal auction theory. We generalize Myerson’s virtual values, and characterize the profit-maximizing IC AMM. The optimal demand curve generally has a jump that can be interpreted as a "bid-ask spread," which we show is caused by a combination of adverse selection risk (dominant when the degree of information asymmetry is large) and monopoly pricing (dominant when asymmetry is small). This work opens up new research directions into the study of automated exchange mechanisms from the lens of optimal auction theory and iterative belief inference, using tools of theoretical computer science in a novel way.more » « less
-
Abstract Robust quantification of predictive uncertainty is a critical addition needed for machine learning applied to weather and climate problems to improve the understanding of what is driving prediction sensitivity. Ensembles of machine learning models provide predictive uncertainty estimates in a conceptually simple way but require multiple models for training and prediction, increasing computational cost and latency. Parametric deep learning can estimate uncertainty with one model by predicting the parameters of a probability distribution but does not account for epistemic uncertainty. Evidential deep learning, a technique that extends parametric deep learning to higher-order distributions, can account for both aleatoric and epistemic uncertainties with one model. This study compares the uncertainty derived from evidential neural networks to that obtained from ensembles. Through applications of the classification of winter precipitation type and regression of surface-layer fluxes, we show evidential deep learning models attaining predictive accuracy rivaling standard methods while robustly quantifying both sources of uncertainty. We evaluate the uncertainty in terms of how well the predictions are calibrated and how well the uncertainty correlates with prediction error. Analyses of uncertainty in the context of the inputs reveal sensitivities to underlying meteorological processes, facilitating interpretation of the models. The conceptual simplicity, interpretability, and computational efficiency of evidential neural networks make them highly extensible, offering a promising approach for reliable and practical uncertainty quantification in Earth system science modeling. To encourage broader adoption of evidential deep learning, we have developed a new Python package, Machine Integration and Learning for Earth Systems (MILES) group Generalized Uncertainty for Earth System Science (GUESS) (MILES-GUESS) (https://github.com/ai2es/miles-guess), that enables users to train and evaluate both evidential and ensemble deep learning. Significance StatementThis study demonstrates a new technique, evidential deep learning, for robust and computationally efficient uncertainty quantification in modeling the Earth system. The method integrates probabilistic principles into deep neural networks, enabling the estimation of both aleatoric uncertainty from noisy data and epistemic uncertainty from model limitations using a single model. Our analyses reveal how decomposing these uncertainties provides valuable insights into reliability, accuracy, and model shortcomings. We show that the approach can rival standard methods in classification and regression tasks within atmospheric science while offering practical advantages such as computational efficiency. With further advances, evidential networks have the potential to enhance risk assessment and decision-making across meteorology by improving uncertainty quantification, a longstanding challenge. This work establishes a strong foundation and motivation for the broader adoption of evidential learning, where properly quantifying uncertainties is critical yet lacking.more » « less
-
This EDI data package contains instructional materials necessary to teach Macrosystems EDDIE Module 7: Using Data to Improve Ecological Forecasts, a ~3-hour educational module for undergraduates. Ecological forecasting is an emerging approach that provides an estimate of the future state of an ecological system with uncertainty, allowing society to prepare for changes in important ecosystem services. To be useful for management, ecological forecasts need to be both accurate enough for managers to be able to rely on them for decision-making and include a representation of forecast uncertainty, so managers can properly interpret the probability of future events. To improve forecast accuracy, forecasts can be updated with observational data once they become available, a process known as data assimilation. Recent improvements in environmental sensor technology and an increase in the number of sensors deployed in ecosystems have increased the availability of data for assimilation to develop and improve forecasts for natural resource management. In this module, students will explore how assimilating data with different amounts of observation uncertainty and at different temporal frequencies affects forecasts of lake water quality, using data from the U.S. National Ecological Observatory Network (NEON). The flexible, three-part (A-B-C) structure of this module makes it adaptable to a range of student levels and course structures. There are two versions of the module: an R Shiny application which does not require students to code, and an RMarkdown version which requires students to read and alter R code to complete module activities. The R Shiny application is published to shinyapps.io and is available at the following link: https://macrosystemseddie.shinyapps.io/module7/. GitHub repositories are available for both the R Shiny (https://github.com/MacrosystemsEDDIE/module7) and RMarkdown versions (https://github.com/MacrosystemsEDDIE/module7_R) of the module, and both code repositories have been published with DOIs to Zenodo (R Shiny version at DOI 10.5281/zenodo.10903839 and RMarkdown version at DOI 10.5281/zenodo.10909589). Readers are referred to the module landing page for additional information (https://serc.carleton.edu/eddie/teaching_materials/modules/module7.html).more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1109/TCST.2022.3183645
