Search

Estimating depth from RGB images is a long-standing ill-posed problem, which has been explored for decades by the computer vision, graphics, and machine learning communities. Among the existing techniques, stereo matching remains one of the most widely used in the literature due to its strong connection to the human binocular system. Traditionally, stereo-based depth estimation has been addressed through matching hand-crafted features across multiple images. Despite the extensive amount of research, these traditional techniques still suffer in the presence of highly textured areas, large uniform regions, and occlusions. Motivated by their growing success in solving various 2D and 3D vision problems, deep learning for stereo-based depth estimation has attracted a growing interest from the community, with more than 150 papers published in this area between 2014 and 2019. This new generation of methods has demonstrated a significant leap in performance, enabling applications such as autonomous driving and augmented reality. In this paper, we provide a comprehensive survey of this new and continuously growing field of research, summarize the most commonly used pipelines, and discuss their benefits and limitations. In retrospect of what has been achieved so far, we also conjecture what the future may hold for deep learning-based stereo for depth estimation research. [ABSTRACT FROM AUTHOR]

Accurate land valuation is necessary for tax purposes, land resources allocation, real estate management and urban development and planning. Since various factors from different domains affect land prices through non-linear relationships, automating the land valuation process on a large scale is a complex task. Advanced technologies in big data analysis and artificial intelligence have demonstrated superior capabilities in knowledge extraction in such cases. Accordingly, this paper develops and compares the performance of four Automated Valuation Models (AVMs) based on machine learning and deep learning techniques utilizing physical, geographical, socio-economic, environmental, legal and planning factors in Melbourne Metropolitan, Australia. According to the results, the eXtreme Gradient Boosting (XGBoost) method outperforms other algorithms of Support Vector Regression (SVR), random forest and Deep Neural Network (DNN). This method has achieved the coefficient of determination (R2) of 0.862, Mean Absolute Percentage Error (MAPE) of 0.139, and normalized Root Mean Square Error (nRMSE) of 0.281. The achieved high accuracy is due to incorporating a wide range of driving factors and applying innovative feature selection and hyperparameter tuning procedures evaluating various possible feature sets and hyperparameters. Accordingly, this paper can contribute to research, governmental and industry-based activities in terms of developing AVMs for mass land valuation. • Automated land valuation models using machine learning and deep learning • Comprehensive feature set and hybrid feature selection process for robust model performance • Spatial mapping of land values in large urban areas • Contribution to fair housing policies and urban development by accurate land valuation • Practical insights for housing affordability strategies in terms of land management [ABSTRACT FROM AUTHOR]

In past few decades, there was a tremendous enhancement in natural disaster and their effects on economy and population. An adverse events like floods, wildfires, cyclones, earthquakes, tsunamis, etc., are regarded as a natural disaster once it strikes the vulnerable population areas. An early tracking of susceptibility areas and immediate tracking of affected areas might help in facilitating rescue and early warnings to the public. To achieve autonomous natural disaster prediction, this paper makes use of current developments in remote sensing, which speed up the availability of aerial/satellite data and are reinforced by progress in the computing sector. The aerial/satellite imageries are employed for acquiring the data from areas of Queensland-Australia that are more prone to natural disaster in an eagle-eye perspective. Since there were several techniques employed so far for the automatic prediction of natural disaster susceptibilities, there were some limitations like reduced rate of accuracy and so on. So as to overcome these limitations, deep learning based automated process is employed for predicting the natural disaster areas and probability of event occurrences. The main intention of the work is to detect the natural disaster occurrence from the sensed data which aids in providing warning to public and to safeguard them by taking necessary actions. Initially, remote sensing data is pre-processed and the features are extracted using Adaptive linear Internal embedding algorithm-based feature extraction (ALIE-FE). The extracted features are selected using Recursive Wrapper-based feature subset selection. To estimate best fitness function and to enhance the prediction accuracy, the optimization process is carried using Bio-Inspired Squirrel Search Optimization algorithm (BI-SSOA). Finally, the classification is carried by means of Deep learning based Multi-layer Alex Net classifier (DBMLA) approach. The simulation is carried and the outcomes attained are estimated for predicting flood susceptibility and wildfire susceptibility. The proposed BISSOA-DBMLA offers sensitivity of 98%, specificity of 99%, and TSS of 97%. The proposed system offers 98.99% classification accuracy. Accuracy, sensitivity, specificity, TSS, and area under the curve (AUC) are used to evaluate the efficacy of the suggested system in light of the achieved results from other approaches. To demonstrate the efficacy of the suggested mechanism, the achieved results are compared with those of current approaches. [ABSTRACT FROM AUTHOR]

Although traditional physical models have been used to analyse groundwater systems, the emergence of novel machine learning models can improve the accuracy of the predictions. Deep learning has been prominent in environmental and climate change problems. In this paper, we present a framework for utilising deep learning models to predict groundwater levels based on nearby streamflow and rainfall data. We address the missing data problem using a Bayesian linear regression model within the deep learning framework. Our deep learning framework utilises models such as long-short term memory (LSTM) networks and convolutional neural networks (CNN) for multi-step ahead time series prediction. We examine the fluctuations in groundwater levels at various boreholes located near Middle Creek in New South Wales, Australia. We use the National Collaborative Research Infrastructure Strategy (NCRIS) groundwater database and utilise Bayesian linear regression to impute missing data. We investigate the accuracy of the selected models for individual and regional basins and univariate and multivariate strategies. Our results show that the LSTM-based regional model with multivariate strategy using rainfall data provided the best accuracy. • We predict groundwater levels based on stream flow and rainfall data. • Our framework uses Bayesian data imputation and deep learning for multi-step prediction. • We examine the groundwater fluctuations at various boreholes located in Australia. • Results show a fine-tuned regional model with rainfall data yielded the best accuracy. • We find that LSTM models outperformed CNNs using the multivariate approach. [ABSTRACT FROM AUTHOR]

Power grid damage and blackouts are increasing with climate change. Load forecasting methods that integrate climate resilience are therefore essential to facilitate timely and accurate network reconfiguration during periods of extreme stress. Our paper proposes a generalised Wildfire Resilient Load Forecasting Model (WRLFM) to predict electricity load based on operational data of a Distribution Network (DN) in Australia during wildfire seasons in 2015–2020. We demonstrate that load forecasting during wildfire seasons is more challenging than during non-wildfire seasons, motivating an imperative need to improve forecast performance during wildfire seasons. To develop the robust WRLFM, comprehensive comparative analyses were conducted to determine proper Machine Learning (ML) forecast structures and methods for incorporating multiple factors. Bi-directional Gated Recurrent Unit (Bi-GRU) and Vision Transformer (ViT) were selected as they performed the best among all 13 recently trending ML methods. Multi-factors were incorporated to contribute to forecast performance, including input sequence structures, calendar information, flexible correlation-based temperature conditions, and categorical Fire Weather Index (FWI). High-resolution categorical FWI was used to build a forecasting model with climate resilience for the first time, significantly enhancing the average stability of forecast performances by 42%. A sensitivity analysis compared data set patterns and model performances during wildfire and non-wildfire seasons. The improvement rate of load forecasting performance during wildfire seasons was more than two times greater than in non-wildfire seasons. This indicates the significance and effectiveness of applying the WRLFM to improve forecast accuracy under extreme weather risks. Overall, the WRLFM reduces the Mean Absolute Percentage Error (MAPE) of the forecast by 14.37% and 20.86% for Bi-GRU and ViT-based models, respectively, achieving an average forecast MAPE of around 3%. • Novel climate-resilient load forecast model for use during extreme wildfire seasons. • Bi-GRU and ViT outperform 11 recently trending machine learning models. • Categorical FWI with high resolution enhances forecast stability by 42%. • Optimised calendar, temperature and FWI input reduce errors by 17.6%. • Forecast improvement during wildfires is twice that in non-wildfire seasons. [ABSTRACT FROM AUTHOR]

The peak period of an energy-generating wave is one of the most important parameters that describe the spectral shape of the oceanic wave, as this indicates the duration for which the waves prevail with respect to their maximum extractable energy. In this paper, a half-hourly peak wave energy period (T P) forecast model is constructed using a suite of statistically significant lagged inputs based on the partial auto-correlation function with an extreme learning machine model developed and its predictive utility is benchmarked against deep learning models, i.e., convolutional neural network (CNN/CovNet) and recurrent neural network (RNN) models and other traditional M5tree, Conditional Maximization based Multiple Linear Regression (MLR-ECM) and MLR models. The objective model (ELM) vs. the comparison models (CNN, RNN, M5tree, MLR-ECM, and MLR) were trained and validated independently on the test dataset obtained from coastal zones of eastern Australia that have a high potential for implementation of wave energy generation systems. The outcomes ascertain that the ELM model can generate significantly accurate predictions of the half-hourly peak wave energy period, providing a good level of accuracy relative to deep learning models in selected coastal study zones. The study establishes the practical usefulness of the ELM model as being a noteworthy methodology for the applications in renewable and sustainable energy resource management systems. • A real-time peak wave energy period forecasting model is constructed using ELM model. • The advance ELM is benchmarked against a deep learning and other traditional models. • The advance ELM generates better predictions than deep learning models in this region. • ELM model provides significant energy management implications for ocean waves. [ABSTRACT FROM AUTHOR]

Local and state governments depend on small area population forecasts to make important decisions concerning the development of local infrastructure and services. Despite their importance, current methods often produce highly inaccurate forecasts. Recent years have witnessed promising developments in time series forecasting using Machine Learning across a wide range of social and economic variables. However, limited work has been undertaken to investigate the potential application of Machine Learning methods in demography, particularly for small area population forecasting. In this paper we describe the development of two Long-Short Term Memory network architectures for small area populations. We employ the Keras Tuner to select layer unit numbers, vary the window width of input data, and apply a double training and validation regime which supports work with short time series and prioritises later sequence values for forecasts. These methods are transferable and can be applied to other data sets. Retrospective small area population forecasts for Australia were created for the periods 2006–16 and 2011–16. Model performance was evaluated against actual data and two benchmark methods (LIN/EXP and CSP-VSG). We also evaluated the impact of constraining small area population forecasts to an independent national forecast. Forecast accuracy was influenced by jump-off year, constraining, area size, and remoteness. The LIN/EXP model was the best performing method for the 2011-based forecasts whilst deep learning methods performed best for the 2006-based forecasts, including significant improvements in the accuracy of 10 year forecasts. However, benchmark methods were consistently more accurate for more remote areas and for those with populations <5000. • A novel deep learning method for small area forecasts is presented. • Forecast accuracy was influenced by jump-off year, constraining, area size, and remoteness. • Deep learning methods performed best for the 5 and 10 year 2006-based forecasts. • The LIN/EXP model was the best performing method for the 5 year 2011-based forecasts. • Benchmark methods were consistently more accurate for more remote areas and for those with populations <5000. [ABSTRACT FROM AUTHOR]

Predicting electricity demand data is considered an essential task in decisions taking, and establishing new infrastructure in the power generation network. To deliver a high-quality electricity demand prediction, this paper proposes a hybrid combination technique, based on a deep learning model of Convolutional Neural Networks and Echo State Networks, named as CESN. Daily electricity demand data from four sites (Roderick, Rocklea, Hemmant and Carpendale), located in Southeast Queensland, Australia, have been used to develop the proposed hybrid prediction model. The study also analyzes five other machine learning-based models (support vector regression, multilayer perceptron, extreme gradient boosting, deep neural network, and Light Gradient Boosting) to compare and evaluate the outcomes of the proposed deep learning approach. The results obtained in the experimental study showed that the proposed hybrid deep learning model is able to obtain the highest performance compared to other existing models developed for daily electricity demand data forecasting. Based on the statistical approaches utilized in this study, the proposed hybrid approach presents the highest prediction accuracy among the compared models. The obtained results showed that the proposed hybrid deep learning algorithm is an excellent and accurate electricity demand forecasting method, which outperformed the state of the art algorithms that are currently used in this problem. [Display omitted] • A Hybrid deep learning model is proposed for daily electricity demand prediction. • Convolutional Neural Networks and Echo State Networks are integrated into a single forecasting algorithm. • The proposed deep learning-based approach is excellent in electricity demand prediction. • The proposed algorithm has practical use in inter-daily electricity demand prediction. [ABSTRACT FROM AUTHOR]

Modern power systems are experiencing the challenge of high uncertainty with the increasing penetration of renewable energy resources and the electrification of heating systems. In this paradigm shift, understanding electricity users' demand is of utmost value to the retailers, aggregators, and policymakers. However, behind-the-meter (BTM) equipment and appliances at the household level are unknown to the other stakeholders mainly due to privacy concerns and tight regulations. In this paper, we seek to identify residential consumers based on their BTM equipment, mainly rooftop photovoltaic (PV) systems and electric heating, using imported/purchased energy data from utility meters. To solve this problem with an interpretable, fast, secure, and maintainable solution, we propose an integrated method called Interpretable Refined Motifs And binary Classification (IRMAC). The proposed method comprises a novel shape-based pattern extraction technique, called Refined Motif (RM) discovery, and a single-neuron classifier. The first part extracts a sub-pattern from the long time series considering the frequency of occurrences, average dissimilarity, and time dynamics while emphasising specific times with annotated distances. The second part identifies users' types with linear complexity while preserving the transparency of the algorithms. With the real data from Australia and Denmark, the proposed method is tested and verified in identifying PV owners and electrical heating system users. The performance of the IRMAC is studied and compared with various state-of-the-art methods. The proposed method reached an accuracy of 96% in identifying rooftop PV users and 94.4% in identifying electric heating users, which is comparable to the best solution based on deep learning, while the speed of the inference model is a thousand times faster. Last but not least, the proposed method is a transparent algorithm, which can tackle the concerns regarding the agnostic decision-making process when policies prohibit some machine learning methods. [ABSTRACT FROM AUTHOR]

More timely and accurate air quality forecasting could contribute to better public health protection and air pollution prevention. Particulates are a significant indicator for measuring the degree of air pollution. This paper reports on research to model an early warning tool for coarse particulates when assessing the impact of the 12 satellite-derived and ground-based meteorological pollutants out of 30 pollutants considered using hourly Australian data from January 2018–December 2020. A one-dimensional convolutional neural network (CNN) was integrated with a one-directional fully gated recurrent unit (GRU) to forecast consecutive hours' air quality. The CNN model acts as a spatial feature extractor, whereas the new generation GRU makes it computationally efficient. The resultant hybrid ' CNN-GRU' is then comprehensively benchmarked outperforming an ensemble of six other deep learning models. The proposed model's efficacy is indicated at the four most air polluted Australian postcodes in the testing phase. A detailed error analysis with visual and statistical metrics for air quality forecasting ascertains the proposed model's countermeasure to reduce harm and loss. The practical tool is immensely beneficial and can be widely deployed to the regions of public health concern where air pollution is a significant hazard. [Display omitted] • A novel early warning hybrid PM10 forecasting framework designed for Australia. • 30 satellite and ground-based meteorological pollutants were considered with 7 modeling approaches. • CNN integrated with GRU outperformed the other models. • Early warning AI model show potential in the health, environment sector, and decision-making. [ABSTRACT FROM AUTHOR]

This paper presents a new hybrid approach for Global Solar Radiation (GSR) prediction problems, based on deep learning approaches. Predictive models are useful ploys in solar energy industries to optimize the performance of photovoltaic power systems. Specifically, in this work we develop a new 4-phase hybrid CXGBRFR framework, which includes a deep learning Convolutional Neural Network, an Extreme Gradient Boosting with Random Forest Regression, and a Harris Hawks Optimization for initial feature selection. The proposed system has been tested in a problem of daily GSR prediction at six solar farms in Australia. Data from three global climate models (GCM) (CSIRO-BOM ACCESS1-0, MOHC Hadley-GEM2-CC and MRI MRI-CGCM3) have been considered as predictive (input) variables for the proposed approach. The variables from these GCMs contain enough information to obtain an accurate prediction of the GSR at each solar farm. The performance of the proposed approach is compared against different deep and shallows learning approaches: Deep Belief Network, Deep Neural Network, Artificial Neural Network, Extreme Learning Machine and Multivariate Auto-Regressive Spline models. We show that the proposed approach exhibits an excellent performance in GSR prediction, against all alternative approaches in all solar farms considered. [Display omitted] • We develop hybrid deep learning model to predict solar radiation • We optimize the algorithm with Harris Hawks, gradient boosting and random forest • We adopt global climate model and observed predictors as learning model inputs • We build an expert system with relevant knowledge in context of solar application • Our methods benefit solar photovoltaic power monitoring in solar engineering problems. [ABSTRACT FROM AUTHOR]

Background: The paper introduces a deep learning-based approach for real-time detection and insights generation about one of the most prevalent chronic conditions in Australia - Pollen allergy. The popular social media platform is used for data collection as cost-effective and unobtrusive alternative for public health monitoring to complement the traditional survey-based approaches.Methods: The data was extracted from Twitter based on pre-defined keywords (i.e. 'hayfever' OR 'hay fever') throughout the period of 6 months, covering the high pollen season in Australia. The following deep learning architectures were adopted in the experiments: CNN, RNN, LSTM and GRU. Both default (GloVe) and domain-specific (HF) word embeddings were used in training the classifiers. Standard evaluation metrics (i.e. Accuracy, Precision and Recall) were calculated for the results validation. Finally, visual correlation with weather variables was performed.Results: The neural networks-based approach was able to correctly identify the implicit mentions of the symptoms and treatments, even unseen previously (accuracy up to 87.9% for GRU with GloVe embeddings of 300 dimensions).Conclusions: The system addresses the shortcomings of the conventional machine learning techniques with manual feature-engineering that prove limiting when exposed to a wide range of non-standard expressions relating to medical concepts. The case-study presented demonstrates an application of 'black-box' approach to the real-world problem, along with its internal workings demonstration towards more transparent, interpretable and reproducible decision-making in health informatics domain. [ABSTRACT FROM AUTHOR]