More Like this

1. Data-Driven Energy and Population Estimation for Real-Time City-Wide Energy Footprinting

   https://doi.org/10.1145/3360322.3360847  

   Wei, Peter; Jiang, Xiaofan (November 2019, ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation)

   Energy footprinting has the potential to raise awareness of energy consumption and lead to energy saving behavior. However, current methods are largely restricted to single buildings; these methods require energy and occupancy monitoring sensor deployments, which can be expensive and difficult to deploy at scale. Further, current methods for estimating energy consumption and population cannot provide fine enough temporal or spatial granularity for a reasonable personal energy footprint estimate. In this work, we present CityEnergy, a data-driven system for city-wide estimation of personal energy footprints. CityEnergy takes advantage of existing sensing infrastructure and data sources in urban cities to provide energy and population estimates at the building level, even in built environments that do not have existing or accessible energy or population data. 

   more » « less
2. Statistical and Machine Learning Approaches for Electrical Energy Forecasting

   https://doi.org/10.1002/widm.70033  

   Machado, Solange; Zhu, Xingquan (September 2025, WIREs Data Mining and Knowledge Discovery)

   ABSTRACT With renewable energy being aggressively integrated into the grid, energy supplies are becoming vulnerable to weather and the environment, and are often incapable of meeting population demands at a large scale if not accurately predicted for energy planning. Understanding consumers' power demands ahead of time and the influences of weather on consumption and generation can help producers generate effective power management plans to support the target demand. In addition to the high correlation with the environment, consumers' behaviors also cause non‐stationary characteristics of energy data, which is the main challenge for energy prediction. In this survey, we perform a review of the literature on prediction methods in the energy field. So far, most of the available research encompasses one type of generation or consumption. There is no research approaching prediction in the energy sector as a whole and its correlated features. We propose to address the energy prediction challenges from both consumption and generation sides, encompassing techniques from statistical to machine learning techniques. We also summarize the work related to energy prediction, electricity measurements, challenges related to energy consumption and generation, energy forecasting methods, and real‐world energy forecasting resources, such as datasets and software solutions for energy prediction. This article is categorized under:Application Areas > Industry Specific ApplicationsTechnologies > PredictionTechnologies > Machine Learning 

   more » « less
3. Chiplet-GAN: Chiplet-based Accelerator Design for Scalable Generative Adversarial Network Inference

   Chen, Yuechen; Louri, Ahmed; Lombardi, Fabrizio; Liu, Shanshan (August 2024, IEEE Circuits and System)

   Generative adversarial networks (GANs) have emerged as a powerful solution for generating synthetic data when the availability of large, labeled training datasets is limited or costly in large-scale machine learning systems. Recent advancements in GAN models have extended their applications across diverse domains, including medicine, robotics, and content synthesis. These advanced GAN models have gained recognition for their excellent accuracy by scaling the model. However, existing accelerators face scalability challenges when dealing with large-scale GAN models. As the size of GAN models increases, the demand for computation and communication resources during inference continues to grow. To address this scalability issue, this article proposes Chiplet-GAN, a chiplet-based accelerator design for GAN inference. Chiplet-GAN enables scalability by adding more chiplets to the system, thereby supporting the scaling of computation capabilities. To handle the increasing communication demand as the system and model scale, a novel interconnection network with adaptive topology and passive/active network links is developed to provide adequate communication support for Chiplet-GAN. Coupled with workload partition and allocation algorithms, Chiplet-GAN reduces execution time and energy consumption for GAN inference workloads as both model and chiplet-system scales. Evaluation results using various GAN models show the effectiveness of Chiplet-GAN. On average, compared to GANAX, SpAtten, and Simba, the Chiplet-GAN reduces execution time and energy consumption by 34% and 21%, respectively. Furthermore, as the system scales for large-scale GAN model inference, Chiplet-GAN achieves reductions in execution time of up to 63% compared to the Simba, a chiplet-based accelerator. 

   more » « less
4. Chiplet-GAN: Chiplet-Based Accelerator Design for Scalable Generative Adversarial Network Inference [Feature]

   Yuechen, Chen; Ahmed, Louri; Fabrizio, Lombardi; Shanshan, Liu (August 2024, IEEE circuits and systems magazine)

   Generative adversarial networks (GANs) have emerged as a powerful solution for generating synthetic data when the availability of large, labeled training datasets is limited or costly in large-scale machine learning systems. Recent advancements in GAN models have extended their applications across diverse domains, including medicine, robotics, and content synthesis. These advanced GAN models have gained recognition for their excellent accuracy by scaling the model. However, existing accelerators face scalability challenges when dealing with large-scale GAN models. As the size of GAN models increases, the demand for computation and communication resources during inference continues to grow. To address this scalability issue, this article proposes Chiplet-GAN, a chiplet-based accelerator design for GAN inference. Chiplet-GAN enables scalability by adding more chiplets to the system, thereby supporting the scaling of computation capabilities. To handle the increasing communication demand as the system and model scale, a novel interconnection network with adaptive topology and passive/active network links is developed to provide adequate communication support for Chiplet-GAN. Coupled with workload partition and allocation algorithms, Chiplet-GAN reduces execution time and energy consumption for GAN inference workloads as both model and chiplet-system scales. Evaluation results using various GAN models show the effectiveness of Chiplet-GAN. On average, compared to GANAX, SpAtten, and Simba, the Chiplet-GAN reduces execution time and energy consumption by 34% and 21%, respectively. Furthermore, as the system scales for large-scale GAN model inference, Chiplet-GAN achieves reductions in execution time of up to 63% compared to the Simba, a chiplet-based accelerator. 

   more » « less
5. Culture Process and the Interpretation of Radiocarbon Data

   https://doi.org/10.1017/RDC.2017.124  

   Freeman, Jacob; Byers, David A; Robinson, Erick; Kelly, Robert L (April 2018, Radiocarbon)

   Abstract Over the last decade, archaeologists have turned to large radiocarbon ( 14 C) data sets to infer prehistoric population size and change. An outstanding question concerns just how direct of an estimate 14 C dates are for human populations. In this paper we propose that 14 C dates are a better estimate of energy consumption, rather than an unmediated, proportional estimate of population size. We use a parametric model to describe the relationship between population size, economic complexity and energy consumption in human societies, and then parametrize the model using data from modern contexts. Our results suggest that energy consumption scales sub-linearly with population size, which means that the analysis of a large 14 C time-series has the potential to misestimate rates of population change and absolute population size. Energy consumption is also an exponential function of economic complexity. Thus, the 14 C record could change semi-independent of population as complexity grows or declines. Scaling models are an important tool for stimulating future research to tease apart the different effects of population and social complexity on energy consumption, and explain variation in the forms of 14 C date time-series in different regions. 

   more » « less