AI, Volume 2, Issue 4 (December 2021) – 16 articles

Natural language understanding technologies play an essential role in automatically solving math word problems. In the process of machine understanding Chinese math word problems, comma disambiguation, which is associated with a class imbalance binary learning problem, is addressed as a valuable instrument to transform the problem statement of math word problems into structured representation. Aiming to resolve this problem, we employed the synthetic minority oversampling technique (SMOTE) and random forests to comma classification after their hyperparameters were jointly optimized. We propose a strict measure to evaluate the performance of deployed comma classification models on comma disambiguation in math word problems. To verify the effectiveness of random forest classifiers with SMOTE on comma disambiguation, we conducted two-stage experiments on two datasets with a collection of evaluation measures. Experimental results showed that random forest classifiers were significantly superior to baseline methods in Chinese comma disambiguation. The SMOTE algorithm with optimized hyperparameter settings based on the categorical distribution of different datasets is preferable, instead of with its default values. For practitioners, we suggest that hyperparameters of a classification models be optimized again after parameter settings of SMOTE have been changed. Full article

Structured Query Language (SQL) is commonly used in Relational Database Management Systems (RDBMS) and is currently one of the most popular data definition and manipulation languages. Its core functionality is implemented, with only some minor variations, throughout all RDBMS products. It is an effective tool in the process of managing and querying data in relational databases. This paper describes a method to effectively automate the conversion of a data query from a Natural Language Query (NLQ) to Structured Query Language (SQL) with Online Analytical Processing (OLAP) cube data warehouse objects. To obtain or manipulate the data from relational databases, the user must be familiar with SQL and must also write an appropriate and valid SQL statement. However, users who are not familiar with SQL are unable to obtain relevant data through relational databases. To address this, we propose a Natural Language Processing (NLP) model to convert an NLQ into an SQL query. This allows novice users to obtain the required data without having to know any complicated SQL details. The model is also capable of handling complex queries using the OLAP cube technique, which allows data to be pre-calculated and stored in a multi-dimensional and ready-to-use format. A multi-dimensional cube (hypercube) is used to connect with the NLP interface, thereby eliminating long-running data queries and enabling self-service business intelligence. The study demonstrated how the use of hypercube technology helps to increase the system response speed and the ability to process very complex query sentences. The system achieved impressive performance in terms of NLP and the accuracy of generating different query sentences. Using OLAP hypercube technology, the study achieved distinguished results compared to previous studies in terms of the speed of the response of the model to NLQ analysis, the generation of complex SQL statements, and the dynamic display of the results. As a plan for future work, it is recommended to use infinite-dimension (n-D) cubes instead of 4-D cubes to enable ingesting as much data as possible in a single object and to facilitate the execution of query statements that may be too complex in query interfaces running in a data warehouse. The study demonstrated how the use of hypercube technology helps to increase system response speed and process very complex query sentences. Full article

Artificial intelligence (AI) is fundamentally transforming smart buildings by increasing energy efficiency and operational productivity, improving life experience, and providing better healthcare services. Sudden Infant Death Syndrome (SIDS) is an unexpected and unexplained death of infants under one year old. Previous research reports that sleeping on the back can significantly reduce the risk of SIDS. Existing sensor-based wearable or touchable monitors have serious drawbacks such as inconvenience and false alarm, so they are not attractive in monitoring infant sleeping postures. Several recent studies use a camera, portable electronics, and AI algorithm to monitor the sleep postures of infants. However, there are two major bottlenecks that prevent AI from detecting potential baby sleeping hazards in smart buildings. In order to overcome these bottlenecks, in this work, we create a complete dataset containing 10,240 day and night vision samples, and use post-training weight quantization to solve the huge memory demand problem. Experimental results verify the effectiveness and benefits of our proposed idea. Compared with the state-of-the-art AI algorithms in the literature, the proposed method reduces memory footprint by at least 89%, while achieving a similar high detection accuracy of about 90%. Our proposed AI algorithm only requires 6.4 MB of memory space, while other existing AI algorithms for sleep posture detection require 58.2 MB to 275 MB of memory space. This comparison shows that the memory is reduced by at least 9 times without sacrificing the detection accuracy. Therefore, our proposed memory-efficient AI algorithm has great potential to be deployed and to run on edge devices, such as micro-controllers and Raspberry Pi, which have low memory footprint, limited power budget, and constrained computing resources. Full article

The concept of searching and localizing vehicles from live traffic videos based on descriptive textual input has yet to be explored in the scholarly literature. Endowing Intelligent Transportation Systems (ITS) with such a capability could help solve crimes on roadways. One major impediment to the advancement of fine-grain vehicle recognition models is the lack of video testbench datasets with annotated ground truth data. Additionally, to the best of our knowledge, no metrics currently exist for evaluating the robustness and performance efficiency of a vehicle recognition model on live videos and even less so for vehicle search and localization models. In this paper, we address these challenges by proposing V-Localize, a novel artificial intelligence framework for vehicle search and continuous localization captured from live traffic videos based on input textual descriptions. An efficient hashgraph algorithm is introduced to compute valid target information from textual input. This work further introduces two novel datasets to advance AI research in these challenging areas. These datasets include (a) the most diverse and large-scale Vehicle Color Recognition (VCoR) dataset with 15 color classes—twice as many as the number of color classes in the largest existing such dataset—to facilitate finer-grain recognition with color information; and (b) a Vehicle Recognition in Video (VRiV) dataset, a first of its kind video testbench dataset for evaluating the performance of vehicle recognition models in live videos rather than still image data. The VRiV dataset will open new avenues for AI researchers to investigate innovative approaches that were previously intractable due to the lack of annotated traffic vehicle recognition video testbench dataset. Finally, to address the gap in the field, five novel metrics are introduced in this paper for adequately accessing the performance of vehicle recognition models in live videos. Ultimately, the proposed metrics could also prove intuitively effective at quantitative model evaluation in other video recognition applications. T One major advantage of the proposed vehicle search and continuous localization framework is that it could be integrated in ITS software solution to aid law enforcement, especially in critical cases such as of amber alerts or hit-and-run incidents. Full article

Sports climbing has grown as a competitive sport over the last decades. This has leading to an increasing interest in guaranteeing the safety of the climber. In particular, operational errors, caused by the belayer, are one of the major issues leading to severe injuries. The objective of this study is to analyze and predict the severity of a pendulum fall based on the movement information from the belayer alone. Therefore, the impact force served as a reference. It was extracted using an Inertial Measurement Unit (IMU) on the climber. Additionally, another IMU was attached to the belayer, from which several hand-crafted features were explored. As this led to a high dimensional feature space, dimension reduction techniques were required to improve the performance. We were able to predict the impact force with a median error of about 4.96%. Pre-defined windows as well as the applied feature dimension reduction techniques allowed for a meaningful interpretation of the results. The belayer was able to reduce the impact force, which is acting onto the climber, by over 30%. So, a monitoring system in a training center could improve the skills of a belayer and hence alleviate the severity of the injuries. Full article

Accurate and timely traffic forecasting is an important task for the realization of urban smart traffic. The random occurrence of social events such as traffic accidents will make traffic prediction particularly difficult. At the same time, most of the existing prediction methods rely on prior knowledge to obtain traffic maps and the obtained map structure cannot be guaranteed to be accurate for the current learning task. In addition, traffic data is highly non-linear and long-term dependent, so it is more difficult to achieve accurate prediction. In response to the above problems, this paper proposes a new integrated unified architecture for traffic prediction based on heterogeneous graph attention network combined with residual-time-series convolutional network, which is called HGA-ResTCN. First, the heterogeneous graph attention is used to capture the changes in the relationship between the traffic graph nodes caused by social events, so as to learn the link weights between the target node and its neighbor nodes; at the same time, by introducing the timing of residual links convolutional network to capture the long-term dependence of complex traffic data. These two models are integrated into a unified framework to learn in an end-to-end manner. Through testing on real-world data sets, the results show that the accuracy of the model in this paper is better than other proposed baselines. Full article

Ambient Assisted Living (AAL) technologies are being developed which could assist elderly people to live healthy and active lives. These technologies have been used to monitor people’s daily exercises, consumption of calories and sleep patterns, and to provide coaching interventions to foster positive behaviour. Speech and audio processing can be used to complement such AAL technologies to inform interventions for healthy ageing by analyzing speech data captured in the user’s home. However, collection of data in home settings presents challenges. One of the most pressing challenges concerns how to manage privacy and data protection. To address this issue, we proposed a low cost system for recording disguised speech signals which can protect user identity by using pitch shifting. The disguised speech so recorded can then be used for training machine learning models for affective behaviour monitoring. Affective behaviour could provide an indicator of the onset of mental health issues such as depression and cognitive impairment, and help develop clinical tools for automatically detecting and monitoring disease progression. In this article, acoustic features extracted from the non-disguised and disguised speech are evaluated in an affect recognition task using six different machine learning classification methods. The results of transfer learning from non-disguised to disguised speech are also demonstrated. We have identified sets of acoustic features which are not affected by the pitch shifting algorithm and also evaluated them in affect recognition. We found that, while the non-disguised speech signal gives the best Unweighted Average Recall (UAR) of 80.01%, the disguised speech signal only causes a slight degradation of performance, reaching 76.29%. The transfer learning from non-disguised to disguised speech results in a reduction of UAR (65.13%). However, feature selection improves the UAR (68.32%). This approach forms part of a large project which includes health and wellbeing monitoring and coaching. Full article

Interpretability is becoming increasingly important for predictive model analysis. Unfortunately, as remarked by many authors, there is still no consensus regarding this notion. The goal of this paper is to propose the definition of a score that allows for quickly comparing interpretable algorithms. This definition consists of three terms, each one being quantitatively measured with a simple formula: predictivity, stability and simplicity. While predictivity has been extensively studied to measure the accuracy of predictive algorithms, stability is based on the Dice-Sorensen index for comparing two rule sets generated by an algorithm using two independent samples. The simplicity is based on the sum of the lengths of the rules derived from the predictive model. The proposed score is a weighted sum of the three terms mentioned above. We use this score to compare the interpretability of a set of rule-based algorithms and tree-based algorithms for the regression case and for the classification case. Full article

One of the most important open challenges in climate science is downscaling. It is a procedure that allows making predictions at local scales, starting from climatic field information available at large scale. Recent advances in deep learning provide new insights and modeling solutions to tackle downscaling-related tasks by automatically learning the coarse-to-fine grained resolution mapping. In particular, deep learning models designed for super-resolution problems in computer vision can be exploited because of the similarity between images and climatic fields maps. For this reason, a new architecture tailored for statistical downscaling (SD), named MSG-GAN-SD, has been developed, allowing interpretability and good stability during training, due to multi-scale gradient information. The proposed architecture, based on a Generative Adversarial Network (GAN), was applied to downscale ERA-Interim 2-m temperature fields, from 83.25 to 13.87 km resolution, covering the EURO-CORDEX domain within the 1979–2018 period. The training process involves seasonal and monthly dataset arrangements, in addition to different training strategies, leading to several models. Furthermore, a model selection framework is introduced in order to mathematically select the best models during the training. The selected models were then tested on the 2015–2018 period using several metrics to identify the best training strategy and dataset arrangement, which finally produced several evaluation maps. This work is the first attempt to use the MSG-GAN architecture for statistical downscaling. The achieved results demonstrate that the models trained on seasonal datasets performed better than those trained on monthly datasets. This study presents an accurate and cost-effective solution that is able to perform downscaling of 2 m temperature climatic maps. Full article

Comparing two sets of documents to identify new topics is useful in many applications, like discovering trending topics from sets of scientific papers, emerging topic detection in microblogs, and interpreting sentiment variations in Twitter. In this paper, the main topic-modeling-based approaches to address this task are examined to identify limitations and necessary enhancements. To overcome these limitations, we introduce two separate frameworks to discover emerging topics through a filtered latent Dirichlet allocation (filtered-LDA) model. The model acts as a filter that identifies old topics from a timestamped set of documents, removes all documents that focus on old topics, and keeps documents that discuss new topics. Filtered-LDA also genuinely reduces the chance of using keywords from old topics to represent emerging topics. The final stage of the filter uses multiple topic visualization formats to improve human interpretability of the filtered topics, and it presents the most-representative document for each topic. Full article

Object detection is one of the vital and challenging tasks of computer vision. It supports a wide range of applications in real life, such as surveillance, shipping, and medical diagnostics. Object detection techniques aim to detect objects of certain target classes in a given image and assign each object to a corresponding class label. These techniques proceed differently in network architecture, training strategy and optimization function. In this paper, we focus on animal species detection as an initial step to mitigate the negative impacts of wildlife–human and wildlife–vehicle encounters in remote wilderness regions and on highways. Our goal is to provide a summary of object detection techniques based on R-CNN models, and to enhance the performance of detecting animal species in accuracy and speed, by using four different R-CNN models and a deformable convolutional neural network. Each model is applied on three wildlife datasets, results are compared and analyzed by using four evaluation metrics. Based on the evaluation, an animal species detection system is proposed. Full article

Solving Constraint Optimization Problems (COPs) can be dramatically simplified by boundary estimation, that is providing tight boundaries of cost functions. By feeding a supervised Machine Learning (ML) model with data composed of the known boundaries and extracted features of COPs, it is possible to train the model to estimate the boundaries of a new COP instance. In this paper, we first give an overview of the existing body of knowledge on ML for Constraint Programming (CP), which learns from problem instances. Second, we introduce a boundary estimation framework that is applied as a tool to support a CP solver. Within this framework, different ML models are discussed and evaluated regarding their suitability for boundary estimation, and countermeasures to avoid unfeasible estimations that avoid the solver finding an optimal solution are shown. Third, we present an experimental study with distinct CP solvers on seven COPs. Our results show that near-optimal boundaries can be learned for these COPs with only little overhead. These estimated boundaries reduce the objective domain size by 60-88% and can help the solver find near-optimal solutions early during the search. Full article

Recent developments in machine learning and deep learning have led to the use of multiple algorithms to make better predictions. Surgical units in hospitals allocate their resources for day surgeries based on the number of elective patients, which is mostly disrupted by emergency surgeries. Sixteen different models were constructed for this comparative study, including four simple and twelve hybrid models for predicting the demand for endocrinology, gastroenterology, vascular, urology, and pediatric surgical units. The four simple models used were seasonal autoregressive integrated moving average (SARIMA), support vector regression (SVR), multilayer perceptron (MLP), and long short-term memory (LSTM). The twelve hybrid models used were a combination of any two of the above-mentioned simple models, namely, SARIMA–SVR, SVR–SARIMA, SARIMA–MLP, MLP–SARIMA, SARIMA–LSTM, LSTM–SARIMA, SVR–MLP, MLP–SVR, SVR–LSTM, LSTM–SVR, MLP–LSTM, and LSTM–MLP. Data from the period 2012–2018 were used to build and test the models for each surgical unit. The results indicated that, in some cases, the simple LSTM model outperformed the others while, in other cases, there was a need for hybrid models. This shows that surgical units are unique in nature and need separate models for predicting their corresponding surgical volumes. Full article

Makeup can disguise facial features, which results in degradation in the performance of many facial-related analysis systems, including face recognition, facial landmark characterisation, aesthetic quantification and automated age estimation methods. Thus, facial makeup is likely to directly affect several real-life applications such as cosmetology and virtual cosmetics recommendation systems, security and access control, and social interaction. In this work, we conduct a comparative study and design automated facial makeup detection systems leveraging multiple learning schemes from a single unconstrained photograph. We have investigated and studied the efficacy of deep learning models for makeup detection incorporating the use of transfer learning strategy with semi-supervised learning using labelled and unlabelled data. First, during the supervised learning, the VGG16 convolution neural network, pre-trained on a large dataset, is fine-tuned on makeup labelled data. Secondly, two unsupervised learning methods, which are self-learning and convolutional auto-encoder, are trained on unlabelled data and then incorporated with supervised learning during semi-supervised learning. Comprehensive experiments and comparative analysis have been conducted on 2479 labelled images and 446 unlabelled images collected from six challenging makeup datasets. The obtained results reveal that the convolutional auto-encoder merged with supervised learning gives the best makeup detection performance achieving an accuracy of 88.33% and area under ROC curve of 95.15%. The promising results obtained from conducted experiments reveal and reflect the efficiency of combining different learning strategies by harnessing labelled and unlabelled data. It would also be advantageous to the beauty industry to develop such computational intelligence methods. Full article

Cryptocurrency is a new sort of asset that has emerged as a result of the advancement of financial technology and it has created a big opportunity for researches. Cryptocurrency price forecasting is difficult due to price volatility and dynamism. Around the world, there are hundreds of cryptocurrencies that are used. This paper proposes three types of recurrent neural network (RNN) algorithms used to predict the prices of three types of cryptocurrencies, namely Bitcoin (BTC), Litecoin (LTC), and Ethereum (ETH). The models show excellent predictions depending on the mean absolute percentage error (MAPE). Results obtained from these models show that the gated recurrent unit (GRU) performed better in prediction for all types of cryptocurrency than the long short-term memory (LSTM) and bidirectional LSTM (bi-LSTM) models. Therefore, it can be considered the best algorithm. GRU presents the most accurate prediction for LTC with MAPE percentages of 0.2454%, 0.8267%, and 0.2116% for BTC, ETH, and LTC, respectively. The bi-LSTM algorithm presents the lowest prediction result compared with the other two algorithms as the MAPE percentages are: 5.990%, 6.85%, and 2.332% for BTC, ETH, and LTC, respectively. Overall, the prediction models in this paper represent accurate results close to the actual prices of cryptocurrencies. The importance of having these models is that they can have significant economic ramifications by helping investors and traders to pinpoint cryptocurrency sales and purchasing. As a plan for future work, a recommendation is made to investigate other factors that might affect the prices of cryptocurrency market such as social media, tweets, and trading volume. Full article

Continuous action spaces impose a serious challenge for reinforcement learning agents. While several off-policy reinforcement learning algorithms provide a universal solution to continuous control problems, the real challenge lies in the fact that different actuators feature different response functions due to wear and tear (in mechanical systems) and fatigue (in biomechanical systems). In this paper, we propose enhancing the actor-critic reinforcement learning agents by parameterising the final layer in the actor network. This layer produces the actions to accommodate the behaviour discrepancy of different actuators under different load conditions during interaction with the environment. To achieve this, the actor is trained to learn the tuning parameter controlling the activation layer (e.g., Tanh and Sigmoid). The learned parameters are then used to create tailored activation functions for each actuator. We ran experiments on three OpenAI Gym environments, i.e., Pendulum-v0, LunarLanderContinuous-v2, and BipedalWalker-v2. Results showed an average of 23.15% and 33.80% increase in total episode reward of the LunarLanderContinuous-v2 and BipedalWalker-v2 environments, respectively. There was no apparent improvement in Pendulum-v0 environment but the proposed method produces a more stable actuation signal compared to the state-of-the-art method. The proposed method allows the reinforcement learning actor to produce more robust actions that accommodate the discrepancy in the actuators’ response functions. This is particularly useful for real life scenarios where actuators exhibit different response functions depending on the load and the interaction with the environment. This also simplifies the transfer learning problem by fine-tuning the parameterised activation layers instead of retraining the entire policy every time an actuator is replaced. Finally, the proposed method would allow better accommodation to biological actuators (e.g., muscles) in biomechanical systems. Full article