Detecting, analyzing, and predicting power outages is crucial for grid risk assessment and disaster mitigation. Numerous outages occur each year, exacerbated by extreme weather events such as hurricanes. Existing outage data are typically reported at the county level, limiting their spatial resolution and making it difficult to capture localized patterns. However, it offers excellent temporal granularity. In contrast, nighttime light satellite image data provides significantly higher spatial resolution and enables a more comprehensive spatial depiction of outages, enhancing the accuracy of assessing the geographic extent and severity of power loss after disaster events. However, these satellite data are only available on a daily basis. Integrating spatiotemporal visual and time-series data sources into a unified knowledge representation can substantially improve power outage detection, analysis, and predictive reasoning. In this paper, we propose GeoOutageKG, a multimodal knowledge graph that integrates diverse data sources, including nighttime light satellite image data, high-resolution spatiotemporal power outage maps, and county-level timeseries outage reports in the U.S. We describe our method for constructing GeoOutageKG by aligning source data with a developed ontology, GeoOutageOnto. Currently, GeoOutageKG includes over 10.6 million individual outage records spanning from 2014 to 2024, 300,000 NTL images spanning from 2012 to 2024, and 15,000 outage maps. GeoOutageKG is a novel, modular and reusable semantic resource that enables robust multimodal data integration. We demonstrate its use through multiresolution analysis of geospatiotemporal power outages. OSF Repository: https://doi.org/10.17605/OSF.IO/QVD8B GitHub Repository: https://purl.org/geooutagekg Ontology Documentation: https://purl.org/geooutageonto

We are developing geospatiotemporal predictive models

Spatiotemporal graph neural networks (STGNNs) have been adopted for predictive analysis in various scientific domains. Despite their promising performance, the dominance of big (geo)spatiotemporal data with large and heterogeneous dimensions raise computational challenges to effective adoption, generalization and fine-tuning of graph models. Moreover, continuous and high quality historical data may not always exist for such generalization. This paper proposes a framework that can co-evolve historical geospatial temporal datasets and an STGNN model by (1) incorporating elevation features optimized for water systems, and (2) integrating and interacting geospatial data discovery and graph learning with a "rehearsal" mechanism, that automatically generalize STGNNs to broader areas. The process divides spatiotemporal data into regional fragments with inferrable features, and iteratively (1) augment sparse training data in terms of feature similarity, (2) explore the augmented data by a trial "rehearsing" of the current model to decide a fraction of data to be adopted, over which a consistently good accuracy is observed, and (3) generalize STGNNs with promising regional data, ensured by rehearsal performance. This exploratory process hence learns to decide when and where to generalize STGNNs, for cost-effective generalization. Using real-world datasets, we experimentally verify the effectiveness and efficiency of our rehearsal framework.

Photovoltaic (PV) power stations have become an integral component to the global sustainable energy landscape. Accurately monitoring and estimating the performance of PV systems is critical to their feasibility for power generation and as a financial asset. One of the most challenging problems is to understand and estimate the long-term Performance Loss Rate (PLR) for large fleets of PV inverters. This paper introduces a novel Spatio-Temporal Graph Neural Network empowered, long-term Trend analysis system (ST-GTrend), to estimate PLR of PV systems at fleet-level. ST-GTrend nontrivially integrates spatio-temporal coherence and graph attention to separate PLR as a long-term 'aging' trend from multiple fluctuation terms in the PV input data, with a design that can easily scale to large PV sets with effective, multi-level parallel computation. (1) To cope with diverse degradation patterns in timeseries, ST-GTrend adopts a paralleled graph autoencoder array to extract aging and fluctuation terms simultaneously, and imposes flatness and smoothness regularizations to disentangle between aging and fluctuation. (2) For large PV systems, ST-GTrend enables a multi-level parallelization paradigm to scale the training and inference computation with a provable performance guarantee. ST-GTrend has been deployed in CRADLE, a scientific high performance computing infrastructure. We evaluated ST-GTrend with three real-world large-scale PV datasets, spanning a time period of 10 years. Our results show that ST-GTrend reduces MAPE and Euclidean distance-based errors on average by 34.74% and 33.66% of SOTA methods, and scales well to large PV sets. We also showcase that the advantages of ST-GTrend generalize for the need of long-term trend analysis in financial and economic data.

In this study, we conduct a detailed evaluation of machine learning and multimodal learning approaches in two distinct areas: a standard medical imaging benchmark and a novel material sciences benchmark. We utilize the CheXpert chest x-ray dataset for medical imaging and introduce a newly created Fluoropolymer Atomic Force Microscopy (AFM) dataset for material sciences. Both datasets are enhanced with additional images and binary metadata, encoded as one-hot vectors. We tested both pretrained and non-pretrained Convolutional Neural Network (CNN) models, such as ResNet50, ResNet101, DenseNet121, InceptionV3, and Xception, across different combinations of image and metadata inputs. Our results reveal that integrating multimodal data, including simple binary metadata, significantly enhances classification accuracy compared to conventional unimodal approaches or advanced MADDi models. This indicates the efficacy of multimodal learning in enriching data representation and boosting image classification performance. Notably, Xception models showed exceptional performance in CheXpert tests, and most models improved crystal structure predictions in AFM datasets. These insights set a new benchmark for performance and underscore the potential of multimodal learning in data-intensive applied science research.

The integration of the global Photovoltaic (PV) market with real time data-loggers has enabled large scale PV data analytical pipelines for power forecasting and reliability assessment of PV fleets. Nevertheless, the performance of PV data analysis depends on the quality of PV timeseries data. We propose a novel Spatio-Temporal Denoising Graph Autoencoder STD-GAE framework to impute missing PV Power Data. STD-GAE exploits temporal correlation, spatial coherence, and value dependencies from domain knowledge to recover missing data. It is empowered by two modules. (1) To cope with sparse yet various scenarios of missing data, STD-GAE incorporates a domain-knowledge aware data augmentation module to create plausible variations of missing data patterns. This generalizes STD-GAE to robust imputation over different seasons and environment. (2) STD-GAE nontrivially integrates spatiotemporal graph convolution layers and denoising autoencoder to improve the accuracy of imputation accuracy at PV fleet level. Experimental results on two PV datasets show that STD-GAE can achieve a gain of 43.14% in imputation accuracy and remains less sensitive to missing rate, different seasons, and missing scenarios, compared with state-of-the-art data imputation methods.

Using big data, distributed computing systems such as Apache Hadoop requires processing massive amount of data to support business and research applications. Thus, it is critical to ensure the cyber security of such systems. To better defend from advanced cyber attacks that pose threats to even well-protected enterprises, system-auditing based techniques have been adopted for monitoring system activities and assisting attack investigation. In this demo, we are building a system that collects system auditing logs from a big data system and performs data analysis to understand how system auditing can be used more effectively to assist attack investigation on big systems. We also built a demo application that detects unexpected file deletion and presents root causes for the deletion.