Neural networks (NNs) have been shown to be competitive against state-of-the-art feature engineering and random forest (RF) classification of periodic variable stars. Although previous work utilising NNs has made use of periodicity by period folding multiple-cycle time-series into a single cycle -- from time-space to phase-space -- no approach to date has taken advantage of the fact that network predictions should be invariant to the initial phase of the period-folded sequence. Initial phase is exogenous to the physical origin of the variability and should thus be factored out. Here, we present cyclic-permutation invariant networks, a novel class of NNs for which invariance to phase shifts is guaranteed through polar coordinate convolutions, which we implement by means of "Symmetry Padding." Across three different datasets of variable star light curves, we show that two implementations of the cyclic-permutation invariant network: the iTCN and the iResNet, consistently outperform non-invariant baselines and reduce overall error rates by between 4% to 22%. Over a 10-class OGLE-III sample, the iTCN/iResNet achieves an average per-class accuracy of 93.4%/93.3%, compared to RNN/RF accuracies of 70.5%/89.5% in a recent study using the same data. Finding improvement on a non-astronomy benchmark, we suggest that the methodology introduced here should also be applicable to a wide range of science domains where periodic data abounds due to physical symmetries.
Bounding is one of the important gaits in quadrupedal locomotion for negotiating obstacles. The authors proposed an effective approach that can learn robust bounding gaits more efficiently despite its large variation in dynamic body movements. The authors first pretrained the neural network (NN) based on data from a robot operated by conventional model based controllers, and then further optimised the pretrained NN via deep reinforcement learning (DRL). In particular, the authors designed a reward function considering contact points and phases to enforce the gait symmetry and periodicity, which improved the bounding performance. The NN based feedback controller was learned in the simulation and directly deployed on the real quadruped robot Jueying Mini successfully. A variety of environments are presented both indoors and outdoors with the authors approach. The authors approach shows efficient computing and good locomotion results by the Jueying Mini quadrupedal robot bounding over uneven terrain.
Existing approaches for replay and synthetic speech detection still lack generalizability to unseen spoofing attacks. This work proposes to leverage a novel model structure, so-called Res2Net, to improve the anti-spoofing countermeasure's generalizability. Res2Net mainly modifies the ResNet block to enable multiple feature scales. Specifically, it splits the feature maps within one block into multiple channel groups and designs a residual-like connection across different channel groups. Such connection increases the possible receptive fields, resulting in multiple feature scales. This multiple scaling mechanism significantly improves the countermeasure's generalizability to unseen spoofing attacks. It also decreases the model size compared to ResNet-based models. Experimental results show that the Res2Net model consistently outperforms ResNet34 and ResNet50 by a large margin in both physical access (PA) and logical access (LA) of the ASVspoof 2019 corpus. Moreover, integration with the squeeze-and-excitation (SE) block can further enhance performance. For feature engineering, we investigate the generalizability of Res2Net combined with different acoustic features, and observe that the constant-Q transform (CQT) achieves the most promising performance in both PA and LA scenarios. Our best single system outperforms other state-of-the-art single systems in both PA and LA of the ASVspoof 2019 corpus.
In this paper, I present churn prediction techniques that have been released so far. Churn prediction is used in the fields of Internet services, games, insurance, and management. However, since it has been used intensively to increase the predictability of various industry/academic fields, there is a big difference in its definition and utilization. In this paper, I collected the definitions of churn used in the fields of business administration, marketing, IT, telecommunications, newspapers, insurance and psychology, and described their differences. Based on this, I classified and explained churn loss, feature engineering, and prediction models. Our study can be used to select the definition of churn and its associated models suitable for the service field that researchers are most interested in by integrating fragmented churn studies in industry/academic fields.
Eye movements are intricate and dynamic biosignals that contain a wealth of cognitive information about the subject. However, these are ambiguous signals and therefore require meticulous feature engineering to be used by machine learning algorithms. We instead propose to learn feature vectors of eye movements in a self-supervised manner. We adopt a contrastive learning approach and propose a set of data transformations that encourage a deep neural network to discern salient and granular gaze patterns. This paper presents a novel experiment utilizing six eye-tracking data sets despite different data specifications and experimental conditions. We assess the learned features on biometric tasks with only a linear classifier, achieving 84.6% accuracy on a mixed dataset, and up to 97.3% accuracy on a single dataset. Our work advances the state of machine learning for eye movements and provides insights into a general representation learning method not only for eye movements but also for similar biosignals.
Huge amounts of textual conversations occur online every day, where multiple conversations take place concurrently. Interleaved conversations lead to difficulties in not only following the ongoing discussions but also extracting relevant information from simultaneous messages. Conversation disentanglement aims to separate intermingled messages into detached conversations. However, existing disentanglement methods rely mostly on handcrafted features that are dataset specific, which hinders generalization and adaptability. In this work, we propose an end-to-end online framework for conversation disentanglement that avoids time-consuming domain-specific feature engineering. We design a novel way to embed the whole utterance that comprises timestamp, speaker, and message text, and proposes a custom attention mechanism that models disentanglement as a pointing problem while effectively capturing inter-utterance interactions in an end-to-end fashion. We also introduce a joint-learning objective to better capture contextual information. Our experiments on the Ubuntu IRC dataset show that our method achieves state-of-the-art performance in both link and conversation prediction tasks.
There has been empirical evidence of presence of non-linearity and chaos at the level of single neurons in biological neural networks. The properties of chaotic neurons inspires us to employ them in artificial learning systems. Here, we propose a Neurochaos Learning (NL) architecture, where the neurons used to extract features from data are 1D chaotic maps. ChaosFEX+SVM, an instance of this NL architecture, is proposed as a hybrid combination of chaos and classical machine learning algorithm. We formally prove that a single layer of NL with a finite number of 1D chaotic neurons satisfies the Universal Approximation Theorem with an exact value for the number of chaotic neurons needed to approximate a discrete real valued function with finite support. This is made possible due to the topological transitivity property of chaos and the existence of uncountably infinite number of dense orbits for the chosen 1D chaotic map. The chaotic neurons in NL get activated under the presence of an input stimulus (data) and output a chaotic firing trajectory. From such chaotic firing trajectories of individual neurons of NL, we extract Firing Time, Firing Rate, Energy and Entropy that constitute ChaosFEX features. These ChaosFEX features are then fed to a Support Vector Machine with linear kernel for classification. The effectiveness of chaotic feature engineering performed by NL (ChaosFEX+SVM) is demonstrated for synthetic and real world datasets in the low and high training sample regimes. Specifically, we consider the problem of classification of genome sequences of SARS-CoV-2 from other coronaviruses (SARS-CoV-1, MERS-CoV and others). With just one training sample per class for 1000 random trials of training, we report an average macro F1-score > 0.99 for the classification of SARS-CoV-2 from SARS-CoV-1 genome sequences. Robustness of ChaosFEX features to additive noise is also demonstrated.
Malware detection and classification is a challenging problem and an active area of research. Traditional machine learning methods depend almost entirely on the ability to extract a set of discriminative features into which characterize malware. However, this feature engineering process is very time consuming. On the contrary, deep learning methods replace manual feature engineering by a system that performs both feature extraction and classification from raw data at once. Despite that, a major shortfall of these methods is their inhability to consider multiple disparate sources of information when performing classification, leading them to perform poorly when compared to multimodal approaches. In this work, we introduce Orthrus, a new bimodal approach to categorize malware into families based on deep learning. Orthrus combines two modalities of data: (1) the byte sequence representing the malware’s binary content, and (2) the assembly language instructions extracted from the assembly language source code of malware, and performs automatic feature learning and classification with a convolutional neural network. The idea is to benefit from multiple feature types to reflect malware’s characteristics. The experiments carried on the Microsoft Malware Classification Challenge dataset show that our proposed solution achieves higher classification performance than deep learning approaches in the literature and n-gram based methods.
Feature engineering, a crucial step of machine learning, aims to extract useful features from raw data to improve data quality. In recent years, great efforts have been devoted to Automated Feature Engineering (AutoFE) to replace expensive human labor. However, existing methods are computationally demanding due to treating AutoFE as a coarse-grained black-box optimization problem over a discrete space. In this work, we propose an efficient gradient-based method called DIFER to perform differentiable automated feature engineering in a continuous vector space. DIFER selects potential features based on evolutionary algorithm and leverages an encoder-predictor-decoder controller to optimize existing features. We map features into the continuous vector space via the encoder, optimize the embedding along the gradient direction induced by the predicted score, and recover better features from the optimized embedding by the decoder. Extensive experiments on classification and regression datasets demonstrate that DIFER can significantly improve the performance of various machine learning algorithms and outperform current state-of-the-art AutoFE methods in terms of both efficiency and performance.
Objective: Many recent studies have suggested that brain deformation resulting from a head impact is linked to the corresponding clinical outcome, such as mild traumatic brain injury (mTBI). Even though several finite element (FE) head models have been developed and validated to calculate brain deformation based on impact kinematics, the clinical application of these FE head models is limited due to the time-consuming nature of FE simulations. This work aims to accelerate the process of brain deformation calculation and thus improve the potential for clinical applications. Methods: We propose a deep learning head model with a five-layer deep neural network and feature engineering, and trained and tested the model on 1803 total head impacts from a combination of head model simulations and on-field college football and mixed martial arts impacts. Results: The proposed deep learning head model can calculate the maximum principal strain for every element in the entire brain in less than 0.001s (with an average root mean squared error of 0.025, and with a standard deviation of 0.002 over twenty repeats with random data partition and model initialization). The contributions of various features to the predictive power of the model were investigated, and it was noted that the features based on angular acceleration were found to be more predictive than the features based on angular velocity. Conclusion: Trained using the dataset of 1803 head impacts, this model can be applied to various sports in the calculation of brain strain with accuracy, and its applicability can even further be extended by incorporating data from other types of head impacts. Significance: In addition to the potential clinical application in real-time brain deformation monitoring, this model will help researchers estimate the brain strain from a large number of head impacts more efficiently than using FE models.
Time series forecasting is a challenging task with applications in a wide range of domains. Auto-regression is one of the most common approaches to address these problems. Accordingly, observations are modelled by multiple regression using their past lags as predictor variables. We investigate the extension of auto-regressive processes using statistics which summarise the recent past dynamics of time series. The result of our research is a novel framework called VEST, designed to perform feature engineering using univariate and numeric time series automatically. The proposed approach works in three main steps. First, recent observations are mapped onto different representations. Second, each representation is summarised by statistical functions. Finally, a filter is applied for feature selection. We discovered that combining the features generated by VEST with auto-regression significantly improves forecasting performance. We provide evidence using 90 time series with high sampling frequency. VEST is publicly available online.
Large software organizations handle many customer support issues every day in the form of bug reports, feature requests, and general misunderstandings as submitted by customers. Strategies to gather, analyze, and negotiate requirements are complemented by efforts to manage customer input after products have been deployed. For the latter, support tickets are key in allowing customers to submit their issues, bug reports, and feature requests. Whenever insufficient attention is given to support issues, there is a chance customers will escalate their issues, and escalation to management is time-consuming and expensive, especially for large organizations managing hundreds of customers and thousands of support tickets. This thesis provides a step towards simplifying the job for support analysts and managers, particularly in predicting the risk of escalating support tickets. In a field study at our large industrial partner, IBM, a design science methodology was employed to characterize the support process and data available to IBM analysts in managing escalations. Through iterative cycles of design and evaluation, support analysts' expert knowledge about their customers was translated into features of a support ticket model to be implemented into a Machine Learning model to predict support ticket escalations. The Machine Learning model was trained and evaluated on over 2.5 million support tickets and 10,000 escalations, obtaining a recall of 79.9% and an 80.8% reduction in the workload for support analysts looking to identify support tickets at risk of escalation. The features developed in the Support Ticket Model are designed to serve as a starting place for organizations interested in implementing the model to predict support ticket escalations, and for future researchers to build on to advance research in Escalation Prediction.
Understanding and keeping the customer happy is a central tenet of requirements engineering. Strategies to gather, analyze, and negotiate requirements are complemented by efforts to manage customer input after products have been deployed. For the latter, support tickets are key in allowing customers to submit their issues, bug reports, and feature requests. If insufficient attention is given to support issues, however, their escalation to management becomes time-consuming and expensive, especially for large organizations managing hundreds of customers and thousands of support tickets. Our work provides a step towards simplifying the job of support analysts and managers, particularly in predicting the risk of escalating support tickets. In a field study at our large industrial partner, IBM, we used a design science research methodology to characterize the support process and data available to IBM analysts in managing escalations. We then implemented these features into a machine learning model to predict support ticket escalations. We trained and evaluated our machine learning model on over 2.5 million support tickets and 10,000 escalations, obtaining a recall of 87.36% and an 88.23% reduction in the workload for support analysts looking to identify support tickets at risk of escalation. Finally, in addition to these research evaluation activities, we compared the performance of our support ticket model with that of a model developed with no feature engineering; the support ticket model features outperformed the non-engineered model. The artifacts created in this research are designed to serve as a starting place for organizations interested in predicting support ticket escalations, and for future researchers to build on to advance research in escalation prediction.
Credit card fraud has emerged as major problem in the electronic payment sector. In this survey, we study data-driven credit card fraud detection particularities and several machine learning methods to address each of its intricate challenges with the goal to identify fraudulent transactions that have been issued illegitimately on behalf of the rightful card owner. In particular, we first characterize a typical credit card detection task: the dataset and its attributes, the metric choice along with some methods to handle such unbalanced datasets. These questions are the entry point of every credit card fraud detection problem. Then we focus on dataset shift (sometimes called concept drift), which refers to the fact that the underlying distribution generating the dataset evolves over times: For example, card holders may change their buying habits over seasons and fraudsters may adapt their strategies. This phenomenon may hinder the usage of machine learning methods for real world datasets such as credit card transactions datasets. Afterwards we highlights different approaches used in order to capture the sequential properties of credit card transactions. These approaches range from feature engineering techniques (transactions aggregations for example) to proper sequence modeling methods such as recurrent neural networks (LSTM) or graphical models (hidden markov models).
Despite the widely reported success of embedding-based machine learning methods on natural language processing tasks, the use of more easily interpreted engineered features remains common in fields such as cognitive impairment (CI) detection. Manually engineering features from noisy text is time and resource consuming, and can potentially result in features that do not enhance model performance. To combat this, we describe a new approach to feature engineering that leverages sequential machine learning models and domain knowledge to predict which features help enhance performance. We provide a concrete example of this method on a standard data set of CI speech and demonstrate that CI classification accuracy improves by 2.3% over a strong baseline when using features produced by this method. This demonstration provides an ex-ample of how this method can be used to assist classification in fields where interpretability is important, such as health care.
Myocardial infarction (MI), or commonly known as heart attack, is a life-threatening health problem worldwide from which 32.4 million people suffer each year. Early diagnosis and treatment of MI are crucial to prevent further heart tissue damages or death. The earliest and most reliable sign of ischemia is regional wall motion abnormality (RWMA) of the affected part of the ventricular muscle. Echocardiography can easily, inexpensively, and non-invasively exhibit the RWMA. In this article, we introduce a three-phase approach for early MI detection in low-quality echocardiography: 1) segmentation of the entire left ventricle (LV) wall using a state-of-the-art deep learning model, 2) analysis of the segmented LV wall by feature engineering, and 3) early MI detection. The main contributions of this study are highly accurate segmentation of the LV wall from low-quality echocardiography, pseudo labeling approach for ground-truth formation of the unannotated LV wall, and the first public echocardiographic dataset (HMC-QU)* for MI detection. Furthermore, the outputs of the proposed approach can significantly help cardiologists for a better assessment of the LV wall characteristics. The proposed approach has achieved 95.72% sensitivity and 99.58% specificity for the LV wall segmentation, and 85.97% sensitivity, 74.03% specificity, and 86.85% precision for MI detection on the HMC-QU dataset. *The benchmark HMC-QU dataset is publicly shared at the repository https://www.kaggle.com/aysendegerli/hmcqu-dataset
Smiles play a vital role in the understanding of social interactions within different communities, and reveal the physical state of mind of people in both real and deceptive ways. Several methods have been proposed to recognize spontaneous and posed smiles. All follow a feature-engineering based pipeline requiring costly pre-processing steps such as manual annotation of face landmarks, tracking, segmentation of smile phases, and hand-crafted features. The resulting computation is expensive, and strongly dependent on pre-processing steps. We investigate an end-to-end deep learning model to address these problems, the first end-to-end model for spontaneous and posed smile recognition. Our fully automated model is fast and learns the feature extraction processes by training a series of convolution and ConvLSTM layer from scratch. Our experiments on four datasets demonstrate the robustness and generalization of the proposed model by achieving state-of-the-art performances.
A physics-informed neural network is developed to solve conductive heat transfer partial differential equation (PDE), along with convective heat transfer PDEs as boundary conditions (BCs), in manufacturing and engineering applications where parts are heated in ovens. Since convective coefficients are typically unknown, current analysis approaches based on trial and error finite element (FE) simulations are slow. The loss function is defined based on errors to satisfy PDE, BCs and initial condition. An adaptive normalizing scheme is developed to reduce loss terms simultaneously. In addition, theory of heat transfer is used for feature engineering. The predictions for 1D and 2D cases are validated by comparing with FE results. It is shown that using engineered features, heat transfer beyond the training zone can be predicted. Trained model allows for fast evaluation of a range of BCs to develop feedback loops, realizing Industry 4.0 concept of active manufacturing control based on sensor data.
An estimated 180 papers focusing on deep learning and EHR were published between 2010 and 2018. Despite the common workflow structure appearing in these publications, no trusted and verified software framework exists, forcing researchers to arduously repeat previous work. In this paper, we propose Cardea, an extensible open-source automated machine learning framework encapsulating common prediction problems in the health domain and allows users to build predictive models with their own data. This system relies on two components: Fast Healthcare Interoperability Resources (FHIR) -- a standardized data structure for electronic health systems -- and several AUTOML frameworks for automated feature engineering, model selection, and tuning. We augment these components with an adaptive data assembler and comprehensive data- and model- auditing capabilities. We demonstrate our framework via 5 prediction tasks on MIMIC-III and Kaggle datasets, which highlight Cardea's human competitiveness, flexibility in problem definition, extensive feature generation capability, adaptable automatic data assembler, and its usability.
Business Processes, i.e., a set of coordinated tasks and activities to achieve a business goal, and their continuous improvements are key to the operation of any organization. In banking, business processes are increasingly dynamic as various technologies have made dynamic processes more prevalent. For example, customer segmentation, i.e., the process of grouping related customers based on common activities and behaviors, could be a data-driven and knowledge-intensive process. In this paper, we present an intelligent data-driven pipeline composed of a set of processing elements to move customers' data from one system to another, transforming the data into the contextualized data and knowledge along the way. The goal is to present a novel intelligent customer segmentation process which automates the feature engineering, i.e., the process of using (banking) domain knowledge to extract features from raw data via data mining techniques, in the banking domain. We adopt a typical scenario for analyzing customer transaction records, to highlight how the presented approach can significantly improve the quality of risk-based customer segmentation in the absence of feature engineering.As result, our proposed method is able to achieve accuracy of 91% compared to classical approaches in terms of detecting, identifying and classifying transaction to the right classification.