A Survey on Deep Learning Pouyanfar, Samira; Sadiq, Saad; Yan, Yilin ...
ACM computing surveys,
09/2019, Letnik:
51, Številka:
5
Journal Article
Recenzirano
The field of machine learning is witnessing its golden era as deep learning slowly becomes the leader in this domain. Deep learning uses multiple layers to represent the abstractions of data to build ...computational models. Some key enabler deep learning algorithms such as generative adversarial networks, convolutional neural networks, and model transfers have completely changed our perception of information processing. However, there exists an aperture of understanding behind this tremendously fast-paced domain, because it was never previously represented from a multiscope perspective. The lack of core understanding renders these powerful methods as black-box machines that inhibit development at a fundamental level. Moreover, deep learning has repeatedly been perceived as a silver bullet to all stumbling blocks in machine learning, which is far from the truth. This article presents a comprehensive review of historical and recent state-of-the-art approaches in visual, audio, and text processing; social network analysis; and natural language processing, followed by the in-depth analysis on pivoting and groundbreaking advances in deep learning applications. It was also undertaken to review the issues faced in deep learning such as unsupervised learning, black-box models, and online learning and to illustrate how these challenges can be transformed into prolific future research avenues.
More than one billion people will face water scarcity within the next ten years due to climate change and unsustainable water usage, and this number is only expected to grow exponentially in the ...future. At current water use rates, supply-side demand management is no longer an effective way to combat water scarcity. Instead, many municipalities and water agencies are looking to demand-side solutions to prevent major water loss. While changing conservation behavior is one demand-based strategy, there is a growing movement toward the adoption of water conservation technology as a way to solve water resource depletion. Installing technology into one’s household requires additional costs and motivation, creating a gap between the overall potential households that could adopt this technology, and how many actually do. This study identified and modeled a variety of demographic and household characteristics, social network influence, and external factors such as water price and rebate policy to see their effect on residential water conservation technology adoption. Using Agent-based Modeling and data obtained from the City of Miami Beach, the coupled effects of these factors were evaluated to examine the effectiveness of different pathways towards the adoption of more water conservation technologies. The results showed that income growth and water pricing structure, more so than any of the demographic or building characteristics, impacted household adoption of water conservation technologies. The results also revealed that the effectiveness of rebate programs depends on conservation technology cost and the affluence of the community. Rebate allocation did influence expensive technology adoption, with the potential to increase the adoption rate by 50%. Additionally, social network connections were shown to have an impact on the rate of adoption independent of price strategy or rebate status. These findings will lead the way for municipalities and other water agencies to more strategically implement interventions to encourage household technology adoption based on the characteristics of their communities.
Occupant behavior is a significant contributor to energy waste in buildings. This research introduces an advanced smartphone application, developed based on the theoretical underpinnings of ...situational awareness theory, to effectively implement multi-method and personalized intervention to encourage energy conservation behaviors of building occupants. The new smart application provides several innovative features, such as energy saving points, customized feedback, and visualized user interface, which are implemented in the application to support multi-method interventions. The application was created using the Java language for Android devices. With the use of the Android platform, the app takes advantage of hardware technology from the user’s mobile device. Measurement of occupancy behavior is accomplished by making use of the device’s positional sensors. Orientation and geomagnetic field sensors serve to provide an accurate location of an occupant inside the building. The application can determine energy waste in a zone by using occupancy behavior. Moreover, the application offers real-time and projected future energy consumption based on occupants’ behaviors. This novel feature can significantly improve communication that can lead to prompt action for building energy reduction. Results show how the app can compile raw data on energy behavior and make it easy to understand for the user through the use of visuals and statistical algorithms.
Using new technologies to maintain, construct, and reuse naturally created products like asphalt, soils, and water can reserve the environment (Baqersad et. al 2017, 2016). The objective of this ...study was to specify and model the behavior of households regarding the installation of water conservation technology and evaluate strategies that could potentially increase water conservation technology adoption at the household level. In particular, this study created an agent-based modeling framework in order to understand various factors and dynamic behaviors affecting the adoption of water conservation technology by households. The model captures various demographic characteristics, household attributes, social network influence, and pricing policies; and then evaluates their influence simultaneously on household decisions in adoption of water conservation technology. The application of the proposed simulation model was demonstrated in a case study of the City of Miami Beach. The simulation results identified the intersectional effects of various factors in household water conservation technology adoption and also investigated the scenario landscape of the adoptions that can inform policy formulation and planning.
Coastal water supply infrastructure systems are exposed to saltwater intrusion exacerbated by sea-level rise stressors. To enable assessing the long-term resilience of these systems to the impact of ...sea-level rise, this study developed a novel hazards-humans-infrastructure nexus framework that enables the integrated modeling of stochastic processes of hazard scenarios, decision-theoretic elements of adaptation planning processes of utility agencies, and dynamic processes of water supply infrastructure performance. Using the proposed framework and data collected from South Miami-Dade service area, a multi-agent simulation model was created to conduct exploratory assessments of the long-term resilience of water supply infrastructure under various sea-level rise scenarios and adaptation approaches. The results showed the capability of the proposed model for scenario landscape generation to discover robust adaptation pathways for enhanced infrastructure resilience under uncertainty. The analysis results could provide actionable scientific information to water infrastructure managers to improve their adaptation planning and investment decision-making processes.
•Climatic hazard stressors are the most significant determinant of resilience in coastal water infrastructure systems.•Adaptive planning approach would increase the likelihood of achieving greater resilience.•Reactive planning approaches are unable to fully close the resilience gap induced by uncertainty.•Robust adaptation decision-making would enhance the long-term resilience of infrastructure systems.
The particle drag force coefficient is a critical variable in modeling multiphase energy systems. Even though various empirical drag models with limited conditions have been published before, ...developing a general drag model is still an important topic. Existing analytical and modeling techniques of particle drag are incapable of supporting the application’s complexity due to their limitations to very specific conditions. This paper proposes the Drag Coefficient Correlation-aided Deep Neural Network (DCC-DNN) architecture to predict the particle drag force coefficient from various single-particle experimental data. Beyond sphericity and Reynolds number, the proposed approach includes an expanded set of features supported by the literature. Simultaneously, model regularization and meta-learning help train a generalized and more reliable drag model, despite the limited data available and the variance exhibited in individual single-particle studies. The presented model applies to spherical and non-spherical particles, providing much-needed generality and reliability for industrial applications.
Display omitted
•Modeling Drag on Non-Spherical Particles Using DNN Models.•The model was trained using over 3000 experimental data points.•DNN aided by correlations using Stack generalization and Mixture of Expert (MoE).•DNN performance supersedes all the previous drag correlations.•Advanced shape features such as lengthwise and crosswise sphericity are considered.
Pictures or videos captured from a low-altitude aircraft or an unmanned aerial vehicle are a fast and cost-effective way to survey the affected scene for the quick and precise assessment of a ...catastrophic event's impacts and damages. Using advanced techniques, such as deep learning, it is now possible to automate the description of disaster scenes and identify features in captured images or recorded videos to gain situational awareness. However, building a large-scale, high-quality dataset with annotated disaster-related features for supervised model training is time-consuming and costly. In this article, we propose a weakly supervised approach to train a deep neural network on low-altitude imagery with highly imbalanced and noisy crowd-sourced labels. We further make use of the rich spatiotemporal data obtained from the pictures and its sequence information to enhance the model's performance during training via label propagation. Our approach achieves the highest score among all the submitted runs in the TRECVID2020 Disaster Scene Description and Indexing (DSDI) Challenge, indicating its superior capabilities in retrieving disaster-related video clips compared to other proposed methods.
From the start, the airline industry has remarkably connected countries all over the world through rapid long-distance transportation, helping people overcome geographic barriers. Consequently, this ...has ushered in substantial economic growth, both nationally and internationally. The airline industry produces vast amounts of data, capturing a diverse set of information about their operations, including data related to passengers, freight, flights, and much more. Analyzing air travel data can advance the understanding of airline market dynamics, allowing companies to provide customized, efficient, and safe transportation services. Due to big data challenges in such a complex environment, the benefits of drawing insights from the air travel data in the airline industry have not yet been fully explored. This article aims to survey various components and corresponding proposed data analysis methodologies that have been identified as essential to the inner workings of the airline industry. We introduce existing data sources commonly used in the papers surveyed and summarize their availability. Finally, we discuss several potential research directions to better harness airline data in the future. We anticipate this study to be used as a comprehensive reference for both members of the airline industry and academic scholars with an interest in airline research.
In this paper, we present a storm surge flooding animation system using Three-Dimensional (3D) visualization of real life Geographic Information System (GIS) data. Putting together ground elevation ...with building information provided by Open Street Maps (OSM), we can recreate real life cities (e.g., South Miami Beach in this paper) in a 3D environment. The 3D terrain development and visualization are done with the aid of the game engine Unity. With this tool, learning about storm surge and hurricanes can be an interactive experience. Moreover, since the system more closely portrays a real life environment, visualizing the effects of storm surge can help users study past hurricane disasters as well as possible forecasted hurricane events. For an immersive experience, we connect the system with an Integrated Computer Augmented Virtual Environment (I-CAVE) to give users the capability of navigation through the terrain in a human-scale view.
Academic literature search is a vital step of every research project, especially in the face of the increasingly rapid growth of scientific knowledge. Semantic academic literature search is an ...approach to scientific article retrieval and ranking using concepts in an attempt to address well-known deficiencies of keyword-based search. The difficulty of semantic search, however, is that it requires significant knowledge engineering, often in the form of conceptual ontologies tailored to a particular scientific domain. It also requires non-trivial tuning, in the form of domain-specific term and concepts weights. As part of an ongoing project seeking to build a domain-specific semantic search system, we present an ontology-based supervised concept learning approach for the biogeochemical scientific literature. We first discuss the creation of a dataset of scientific articles in the biogeochemical domain annotated using the Environment Ontology (ENVO). Next we present a supervised machine learning classifier-a random decision forest-that uses a distinctive set of features to learn ENVO concepts and then label and index scientific articles at the sentence level. Finally, we evaluate our approach against two baseline methods, keyword-based and bag-of-words, achieving an overall performance of 0.76 F_1 measure, an improvement of approximately 50%.