skip to main content


Title: DeepExtrema: A Deep Learning Approach for Forecasting Block Maxima in Time Series Data

Accurate forecasting of extreme values in time series is critical due to the significant impact of extreme events on human and natural systems. This paper presents DeepExtrema, a novel framework that combines a deep neural network (DNN) with generalized extreme value (GEV) distribution to forecast the block maximum value of a time series. Implementing such a network is a challenge as the framework must preserve the inter-dependent constraints among the GEV model parameters even when the DNN is initialized. We describe our approach to address this challenge and present an architecture that enables both conditional mean and quantile prediction of the block maxima. The extensive experiments performed on both real-world and synthetic data demonstrated the superiority of DeepExtrema compared to other baseline methods.

 
more » « less
Award ID(s):
2006633
NSF-PAR ID:
10358675
Author(s) / Creator(s):
; ; ; ;
Date Published:
Journal Name:
Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence
Page Range / eLocation ID:
2980 to 2986
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Forecasting the block maxima of a future time window is a challenging task due to the difficulty in inferring the tail distribution of a target variable. As the historical observations alone may not be sufficient to train robust models to predict the block maxima, domain-driven process models are often available in many scientific domains to supplement the observation data and improve the forecast accuracy. Unfortunately, coupling the historical observations with process model outputs is a challenge due to their disparate temporal coverage. This paper presents Self-Recover, a deep learning framework to predict the block maxima of a time window by employing self-supervised learning to address the varying temporal data coverage problem. Specifically Self-Recover uses a combination of contrastive and generative self-supervised learning schemes along with a denoising autoencoder to impute the missing values. The framework also combines representations of the historical observations with process model outputs via a residual learning approach and learns the generalized extreme value (GEV) distribution characterizing the block maxima values. This enables the framework to reliably estimate the block maxima of each time window along with its confidence interval. Extensive experiments on real-world datasets demonstrate the superiority of Self-Recover compared to other state-of-the-art forecasting methods.

     
    more » « less
  2. Abstract

    We develop a space‐time Bayesian hierarchical modeling (BHM) framework for two flood risk attributes—seasonal daily maximum flow and the number of events that exceed a threshold during a season (NEETM)—at a suite of gauge locations on a river network. The model uses generalized extreme value (GEV) and Poisson distributions as marginals for these flood attributes with non‐stationary parameters. The rate parameters of the Poisson distribution and location, scale, and shape parameters of the GEV are modeled as linear functions of suitable covariates. Gaussian copulas are applied to capture the spatial dependence. The best covariates are selected using the Watanabe‐Akaike information criterion (WAIC). The modeling framework results in the posterior distribution of the flood attributes at all the gauges and various lead times. We demonstrate the utility of this modeling framework to forecast the flood risk attributes during the summer peak monsoon season (July‐August) at five gauges in the Narmada River basin (NRB) of West‐Central India for several lead times (0–3 months). As potential covariates, we consider climate indices such as El Niño–Southern Oscillation (ENSO), the Indian Ocean Dipole (IOD), and the Pacific Warm Pool Region (PWPR) from antecedent seasons, which have shown strong teleconnections with the Indian monsoon. We also include new indices related to the East Pacific and West Indian Ocean regions depending on the lead times. We show useful long lead skill from this modeling approach which has a strong potential to enable robust risk‐based flood mitigation and adaptation strategies 3 months before flood occurrences.

     
    more » « less
  3. Abstract

    In this study, we examine extremes of atmospheric water balance components through analyses of annual maxima of precipitable water and water vapor transport. Our analyses are grounded in Extreme Value Theory, using the Generalized Extreme Value (GEV) distribution as a platform for assessing water balance extremes. Annual maxima of atmospheric water balance terms are computed from North American Regional Reanalysis (NARR) fields for the 40‐year period extending from 1979 to 2018 on a grid of approximately 0.3‐degree resolution. We assess nonstationarities in the annual maximum time series through tests for monotonic trends. Estimates of the location, scale, and shape parameters for the Generalized Extreme Value (GEV) distribution are used to examine the spatial variability of water balance extremes. We focus on estimates of the GEV shape parameter, which dictates the “thickness” of the upper tail of the distribution. Estimates of the GEV shape parameter for precipitable water generally point to bounded distributions, but clusters of unbounded, thick‐tailed distributions are linked to exceptionally large record values of precipitable water associated with tropical cyclones in the Gulf of Mexico and Atlantic. Larger regions of “thick‐tailed” distributions are found for integrated water vapor transport (IVT). Non‐stationary GEV models are used to examine the impacts of trends on extremes of the atmospheric water balance. Mixtures of rare events associated with tropical cyclones and extratropical cyclones play a central role in analyses of water balance extremes.

     
    more » « less
  4. Extreme storm surges can overwhelm many coastal flooding protection measures in place and cause severe damages to private communities, public infrastructure, and natural ecosystems. In the US Mid-Atlantic, a highly developed and commercially active region, coastal flooding is one of the most significant natural hazards and a year-round threat from both tropical and extra-tropical cyclones. Mean sea levels and high-tide flood frequency has increased significantly in recent years, and major storms are projected to increase into the foreseeable future. We estimate extreme surges using hourly water level data and harmonic analysis for 1980–2019 at 12 NOAA tide gauges in and around the Delaware and Chesapeake Bays. Return levels (RLs) are computed for 1.1, 3, 5, 10, 25, 50, and 100-year return periods using stationary extreme value analysis on detrended skew surges. Two traditional approaches are investigated, Block Maxima fit to General Extreme Value distribution and Points-Over-Threshold fit to Generalized Pareto distribution, although with two important enhancements. First, the GEV r -largest order statistics distribution is used; a modified version of the GEV distribution that allows for multiple maximum values per year. Second, a systematic procedure is used to select the optimum value for r (for the BM/GEVr approach) and the threshold (for the POT/GP approach) at each tide gauge separately. RLs have similar magnitudes and spatial patterns from both methods, with BM/GEVr resulting in generally larger 100-year and smaller 1.1-year RLs. Maximum values are found at the Lewes (Delaware Bay) and Sewells Point (Chesapeake Bay) tide gauges, both located in the southwest region of their respective bays. Minimum values are found toward the central bay regions. In the Delaware Bay, the POT/GP approach is consistent and results in narrower uncertainty bands whereas the results are mixed for the Chesapeake. Results from this study aim to increase reliability of projections of extreme water levels due to extreme storms and ultimately help in long-term planning of mitigation and implementation of adaptation measures. 
    more » « less
  5. It is appealing but challenging to achieve real-time deep neural network (DNN) inference on mobile devices because even the powerful modern mobile devices are considered “resource-constrained” when executing large-scale DNNs. It necessitates the sparse model inference via weight pruning, i.e., DNN weight sparsity, and it is desirable to design a new DNN weight sparsity scheme that can facilitate real-time inference on mobile devices while preserving a high sparse model accuracy. This paper designs a novel mobile inference acceleration framework GRIM that is General to both convolutional neural networks (CNNs) and recurrent neural networks (RNNs) and that achieves Real-time execution and high accuracy, leveraging fine-grained structured sparse model Inference and compiler optimizations for Mobiles. We start by proposing a new fine-grained structured sparsity scheme through the Block-based Column-Row (BCR) pruning. Based on this new fine-grained structured sparsity, our GRIM framework consists of two parts: (a) the compiler optimization and code generation for real-time mobile inference; and (b) the BCR pruning optimizations for determining pruning hyperparameters and performing weight pruning. We compare GRIM with Alibaba MNN, TVM, TensorFlow-Lite, a sparse implementation based on CSR, PatDNN, and ESE (a representative FPGA inference acceleration framework for RNNs), and achieve up to 14.08× speedup. 
    more » « less