The task of developing a machine learning (ML) model for a particular problem is inherently open-ended, and there is an unbounded set of possible solutions. Steps of the ML development pipeline, such as feature engineering, loss function specification, data imputation, and dimensionality reduction, require the engineer to consider an extensive and often infinite array of possibilities. Successfully identifying high-performing solutions for an unfamiliar dataset or problem requires a mix of mathematical prowess and creativity applied towards inventing and repurposing novel ML methods. Here, we explore the feasibility of hosting crowdsourced ML challenges to facilitate a breadth-first exploration of open-ended research problems, thereby expanding the search space of problem solutions beyond what a typical ML team could viably investigate. We develop MLPro, a system which combines the notion of open-ended ML coding problems with the concept of an automatic online code judging platform. To conduct a pilot evaluation of this paradigm, we crowdsource several open-ended ML challenges to ML and data science practitioners. We describe results from two separate challenges. We find that for sufficiently unconstrained and complex problems, many experts submit similar solutions, but some experts provide unique solutions which outperform the "typical" solution class. We suggest that automated expert crowdsourcing systems such as MLPro have the potential to accelerate ML engineering creativity.
Input features play a crucial role in DNN-based recommender systems with thousands of categorical and continuous fields from users, items, contexts, and interactions. Noisy features and inappropriate embedding dimension assignments can deteriorate the performance of recommender systems and introduce unnecessary complexity in model training and online serving. Optimizing the input configuration of DNN models, including feature selection and embedding dimension assignment, has become one of the essential topics in feature engineering. However, in existing industrial practices, feature selection and dimension search are optimized sequentially, i.e., feature selection is performed first, followed by dimension search to determine the optimal dimension size for each selected feature. Such a sequential optimization mechanism increases training costs and risks generating suboptimal input configurations. To address this problem, we propose a differentiable neural input razor (i-Razor) that enables joint optimization of feature selection and dimension search. Concretely, we introduce an end-to-end differentiable model to learn the relative importance of different embedding regions of each feature. Furthermore, a flexible pruning algorithm is proposed to achieve feature filtering and dimension derivation simultaneously. Extensive experiments on two large-scale public datasets in the Click-Through-Rate (CTR) prediction task demonstrate the efficacy and superiority of i-Razor in balancing model complexity and performance.
We present a study of the potential for Convolutional Neural Networks (CNNs) to enable separation of astrophysical transients from image artifacts, a task known as "real-bogus" classification without requiring a template subtracted (or difference) image which requires a computationally expensive process to generate, involving image matching on small spatial scales in large volumes of data. Using data from the Dark Energy Survey, we explore the use of CNNs to (1) automate the "real-bogus" classification, (2) reduce the computational costs of transient discovery. We compare the efficiency of two CNNs with similar architectures, one that uses "image triplets" (templates, search, and difference image) and one that takes as input the template and search only. We measure the decrease in efficiency associated with the loss of information in input finding that the testing accuracy is reduced from 96% to 91.1%. We further investigate how the latter model learns the required information from the template and search by exploring the saliency maps. Our work (1) confirms that CNNs are excellent models for "real-bogus" classification that rely exclusively on the imaging data and require no feature engineering task; (2) demonstrates that high-accuracy (> 90%) models can be built without the need to construct difference images, but some accuracy is lost. Since once trained, neural networks can generate predictions at minimal computational costs, we argue that future implementations of this methodology could dramatically reduce the computational costs in the detection of transients in synoptic surveys like Rubin Observatory's Legacy Survey of Space and Time by bypassing the Difference Image Analysis entirely.
Contact tracing is a method used by public health organisations to try prevent the spread of infectious diseases in the community. Traditionally performed by manual contact tracers, more recently the use of apps have been considered utilising phone sensor data to determine the distance between two phones. In this paper, we investigate the development of machine learning approaches to determine the distance between two mobile phone devices using Bluetooth Low Energy, sensory data and meta data. We use TableNet architecture and feature engineering to improve on the existing state of the art (total nDCF 0.21 vs 2.08), significantly outperforming existing models.
This paper describes our submission to SemEval-2022 Task 6 on sarcasm detection and its five subtasks for English and Arabic. Sarcasm conveys a meaning which contradicts the literal meaning, and it is mainly found on social networks. It has a significant role in understanding the intention of the user. For detecting sarcasm, we used deep learning techniques based on transformers due to its success in the field of Natural Language Processing (NLP) without the need for feature engineering. The datasets were taken from tweets. We created new datasets by augmenting with external data or by using word embeddings and repetition of instances. Experiments were done on the datasets with different types of preprocessing because it is crucial in this task. The rank of our team was consistent across four subtasks (fourth rank in three subtasks and sixth rank in one subtask); whereas other teams might be in the top ranks for some subtasks but rank drastically less in other subtasks. This implies the robustness and stability of the models and the techniques we used.
Deep Learning has revolutionized various fields, including Computer Vision, Natural Language Processing, as well as Biomedical research. Within the field of neuroscience, specifically in electrophysiological neuroimaging, researchers are starting to explore leveraging deep learning to make predictions on their data without extensive feature engineering. The availability of large-scale datasets is a crucial aspect of allowing the experimentation of Deep Learning models. We are publishing the first large-scale clinical EEG dataset that simplifies data access and management for Deep Learning. This dataset contains eyes-closed EEG data prepared from a collection of 1,574 juvenile participants from the Healthy Brain Network. We demonstrate a use case integrating this framework, and discuss why providing such neuroinformatics infrastructure to the community is critical for future scientific discoveries.
The hypothesis here states that neural network algorithms such as Multi-layer Perceptron (MLP) have higher accuracy in differentiating malicious and semi-structured phishing URLs. Compared to classical machine learning algorithms such as Logistic Regression and Multinomial Naive Bayes, the classical algorithms rely heavily on substantial corpus data training and machine learning experts' domain knowledge to perform complex feature engineering. MLP could perform non-linear separable multi-classes classification and focus less on corpus feature training. In addition, backpropagation weight adjustment could learn which features are more important in differentiating phishing from other attack types.
Automated essay scoring plays an important role in judging students' language abilities in education. Traditional approaches use handcrafted features to score and are time-consuming and complicated. Recently, neural network approaches have improved performance without any feature engineering. Unlike other natural language processing tasks, only a small number of datasets are publicly available for automated essay scoring, and the size of the dataset is not sufficiently large. Considering that the performance of a neural network is closely related to the size of the dataset, the lack of data limits the performance improvement of the automated essay scoring model. In this paper, we proposed a method to increase the number of essay-score pairs using back-translation and score adjustment and applied it to the Automated Student Assessment Prize dataset for augmentation. We evaluated the effectiveness of the augmented data using models from prior work. In addition, performance was evaluated in a model using long short-term memory, which is widely used for automated essay scoring. The performance of the models was improved by using augmented data to train the models.
The Automunge open source python library platform for tabular data pre-processing automates feature engineering data transformations of numerical encoding and missing data infill to received tidy data on bases fit to properties of columns in a designated train set for consistent and efficient application to subsequent data pipelines such as for inference, where transformations may be applied to distinct columns in "family tree" sets with generations and branches of derivations. Included in the library of transformations are methods to extract structure from bounded categorical string sets by way of automated string parsing, in which comparisons between entries in the set of unique values are parsed to identify character subset overlaps which may be encoded by appended columns of boolean overlap detection activations or by replacing string entries with identified overlap partitions. Further string parsing options, which may also be applied to unbounded categoric sets, include extraction of numeric substring partitions from entries or search functions to identify presence of specified substring partitions. The aggregation of these methods into "family tree" sets of transformations are demonstrated for use to automatically extract structure from categoric string compositions in relation to the set of entries in a column, such as may be applied to prepare categoric string set encodings for machine learning without human intervention.
Mainstream practice in machine learning with tabular data may take for granted that any feature engineering beyond scaling for numeric sets is superfluous in context of deep neural networks. This paper will offer arguments for potential benefits of extended encodings of numeric streams in deep learning by way of a survey of options for numeric transformations as available in the Automunge open source python library platform for tabular data pipelines, where transformations may be applied to distinct columns in "family tree" sets with generations and branches of derivations. Automunge transformation options include normalization, binning, noise injection, derivatives, and more. The aggregation of these methods into family tree sets of transformations are demonstrated for use to present numeric features to machine learning in multiple configurations of varying information content, as may be applied to encode numeric sets of unknown interpretation. Experiments demonstrate the realization of a novel generalized solution to data augmentation by noise injection for tabular learning, as may materially benefit model performance in applications with underserved training data.
Vital node identification is the problem of finding nodes of highest importance in complex networks. This problem has crucial applications in various contexts such as viral marketing or controlling the propagation of virus or rumours in real-world networks. Existing approaches for vital node identification mainly focus on capturing the importance of a node through a mathematical expression which directly relates structural properties of the node to its vitality. Although these heuristic approaches have achieved good performance in practice, they have weak adaptability, and their performance is limited to specific settings and certain dynamics. Inspired by the power of machine learning models for efficiently capturing different types of patterns and relations, we propose a machine learning-based, data driven approach for vital node identification. The main idea is to train the model with a small portion of the graph, say 0.5% of the nodes, and do the prediction on the rest of the nodes. The ground-truth vitality for the train data is computed by simulating the SIR diffusion method starting from the train nodes. We use collective feature engineering where each node in the network is represented by incorporating elements of its connectivity, degree and extended coreness. Several machine learning models are trained on the node representations, but the best results are achieved by a Support Vector Regression machine with RBF kernel. The empirical results confirms that the proposed model outperforms state-of-the-art models on a selection of datasets, while it also shows more adaptability to changes in the dynamics parameters.
People undergoing neuromuscular dysfunctions and amputated limbs require automatic prosthetic appliances. In developing such prostheses, the precise detection of brain motor actions is imperative for the Grasp-and-Lift (GAL) tasks. Because of the low-cost and non-invasive essence of Electroencephalography (EEG), it is widely preferred for detecting motor actions during the controls of prosthetic tools. This article has automated the hand movement activity viz GAL detection method from the 32-channel EEG signals. The proposed pipeline essentially combines preprocessing and end-to-end detection steps, eliminating the requirement of hand-crafted feature engineering. Preprocessing action consists of raw signal denoising, using either Discrete Wavelet Transform (DWT) or highpass or bandpass filtering and data standardization. The detection step consists of Convolutional Neural Network (CNN)- or Long Short Term Memory (LSTM)-based model. All the investigations utilize the publicly available WAY-EEG-GAL dataset, having six different GAL events. The best experiment reveals that the proposed framework achieves an average area under the ROC curve of 0.944, employing the DWT-based denoising filter, data standardization, and CNN-based detection model. The obtained outcome designates an excellent achievement of the introduced method in detecting GAL events from the EEG signals, turning it applicable to prosthetic appliances, brain-computer interfaces, robotic arms, etc.
The Quality of Experience (QoE) based service management remains key for successful provisioning of multimedia services in next-generation networks such as 5G/6G, which requires proper tools for quality monitoring, prediction and resource management where machine learning (ML) can play a crucial role. In this paper, we provide a tutorial on the development and deployment of the QoE measurement and prediction solutions for video streaming services based on supervised learning ML models. Firstly, we provide a detailed pipeline for developing and deploying supervised learning-based video streaming QoE prediction models which covers several stages including data collection, feature engineering, model optimization and training, testing and prediction and evaluation. Secondly, we discuss the deployment of the ML model for the QoE prediction/measurement in the next generation networks (5G/6G) using network enabling technologies such as Software-Defined Networking (SDN), Network Function Virtualization (NFV) and Mobile Edge Computing (MEC) by proposing reference architecture. Thirdly, we present a comparative study of the state-of-the-art supervised learning ML models for QoE prediction of video streaming applications based on multiple performance metrics.
Time series forecasting is fundamental for various use cases in different domains such as energy systems and economics. Creating a forecasting model for a specific use case requires an iterative and complex design process. The typical design process includes the five sections (1) data pre-processing, (2) feature engineering, (3) hyperparameter optimization, (4) forecasting method selection, and (5) forecast ensembling, which are commonly organized in a pipeline structure. One promising approach to handle the ever-growing demand for time series forecasts is automating this design process. The present paper, thus, analyzes the existing literature on automated time series forecasting pipelines to investigate how to automate the design process of forecasting models. Thereby, we consider both Automated Machine Learning (AutoML) and automated statistical forecasting methods in a single forecasting pipeline. For this purpose, we firstly present and compare the proposed automation methods for each pipeline section. Secondly, we analyze the automation methods regarding their interaction, combination, and coverage of the five pipeline sections. For both, we discuss the literature, identify problems, give recommendations, and suggest future research. This review reveals that the majority of papers only cover two or three of the five pipeline sections. We conclude that future research has to holistically consider the automation of the forecasting pipeline to enable the large-scale application of time series forecasting.
Along with the flourish of the information age, massive amounts of data are generated day by day. Due to the large-scale and high-dimensional characteristics of these data, it is often difficult to achieve better decision-making in practical applications. Therefore, an efficient big data analytics method is urgently needed. For feature engineering, feature selection seems to be an important research content in which is anticipated to select "excellent" features from candidate ones. Different functions can be realized through feature selection, such as dimensionality reduction, model effect improvement, and model performance improvement. In many classification tasks, researchers found that data seem to be usually close to each other if they are from the same class; thus, local compactness is of great importance for the evaluation of a feature. In this manuscript, we propose a fast unsupervised feature selection method, named as, Compactness Score (CSUFS), to select desired features. To demonstrate the efficiency and accuracy, several data sets are chosen with extensive experiments being performed. Later, the effectiveness and superiority of our method are revealed through addressing clustering tasks. Here, the performance is indicated by several well-known evaluation metrics, while the efficiency is reflected by the corresponding running time. As revealed by the simulation results, our proposed algorithm seems to be more accurate and efficient compared with existing algorithms.
Emotion recognition using EEG has been widely studied to address the challenges associated with affective computing. Using manual feature extraction methods on EEG signals results in sub-optimal performance by the learning models. With the advancements in deep learning as a tool for automated feature engineering, in this work, a hybrid of manual and automatic feature extraction methods has been proposed. The asymmetry in different brain regions is captured in a 2D vector, termed the AsMap, from the differential entropy features of EEG signals. These AsMaps are then used to extract features automatically using a convolutional neural network model. The proposed feature extraction method has been compared with differential entropy and other feature extraction methods such as relative asymmetry, differential asymmetry and differential caudality. Experiments are conducted using the SJTU emotion EEG dataset and the DEAP dataset on different classification problems based on the number of classes. Results obtained indicate that the proposed method of feature extraction results in higher classification accuracy, outperforming the other feature extraction methods. The highest classification accuracy of 97.10% is achieved on a three-class classification problem using the SJTU emotion EEG dataset. Further, this work has also assessed the impact of window size on classification accuracy.
In this work, we focus on low-resource dependency parsing for multiple languages. Several strategies are tailored to enhance performance in low-resource scenarios. While these are well-known to the community, it is not trivial to select the best-performing combination of these strategies for a low-resource language that we are interested in, and not much attention has been given to measuring the efficacy of these strategies. We experiment with 5 low-resource strategies for our ensembled approach on 7 Universal Dependency (UD) low-resource languages. Our exhaustive experimentation on these languages supports the effective improvements for languages not covered in pretrained models. We show a successful application of the ensembled system on a truly low-resource language Sanskrit. The code and data are available at: https://github.com/Jivnesh/SanDP
Past research has demonstrated that the explicit use of protected attributes in machine learning can improve both performance and fairness. Many machine learning algorithms, however, cannot directly process categorical attributes, such as country of birth or ethnicity. Because protected attributes frequently are categorical, they must be encoded as features that can be input to a chosen machine learning algorithm, e.g.\ support vector machines, gradient boosting decision trees or linear models. Thereby, encoding methods influence how and what the machine learning algorithm will learn, affecting model performance and fairness. This work compares the accuracy and fairness implications of the two most well-known encoding methods: \emph{one-hot encoding} and \emph{target encoding}. We distinguish between two types of induced bias that may arise from these encoding methods and may lead to unfair models. The first type, \textit{irreducible bias}, is due to direct group category discrimination, and the second type, \textit{reducible bias}, is due to the large variance in statistically underrepresented groups. We investigate the interaction between categorical encodings and target encoding regularization methods that reduce unfairness. Furthermore, we consider the problem of intersectional unfairness that may arise when machine learning best practices improve performance measures by encoding several categorical attributes into a high-cardinality feature.
Capturing high-frequency data concerning the condition of complex systems, e.g. by acoustic monitoring, has become increasingly prevalent. Such high-frequency signals typically contain time dependencies ranging over different time scales and different types of cyclic behaviors. Processing such signals requires careful feature engineering, particularly the extraction of meaningful time-frequency features. This can be time-consuming and the performance is often dependent on the choice of parameters. To address these limitations, we propose a deep learning framework for learnable wavelet packet transforms, enabling to learn features automatically from data and optimise them with respect to the defined objective function. The learned features can be represented as a spectrogram, containing the important time-frequency information of the dataset. We evaluate the properties and performance of the proposed approach by evaluating its improved spectral leakage and by applying it to an anomaly detection task for acoustic monitoring.
Since frequent severe droughts are lengthening the dry season in the Amazon Rainforest, it is important to detect wildfires promptly and forecast possible spread for effective suppression response. Current wildfire detection models are not versatile enough for the low-technology conditions of South American hot spots. This deep learning study first trains a Fully Convolutional Neural Network on Landsat 8 images of Ecuador and the Galapagos, using Green and Short-wave Infrared bands to predict pixel-level binary fire masks. This model achieves a 0.962 validation F2 score and a 0.932 F2 score on test data from Guyana and Suriname. Afterward, image segmentation is conducted on the Cirrus band using K-Means Clustering to simplify continuous pixel values into three discrete classes representing differing degrees of cirrus cloud contamination. Three additional Convolutional Neural Networks are trained to conduct a sensitivity analysis measuring the effect of simplified features on model accuracy and train time. The Experimental model trained on the segmented cirrus images provides a statistically significant decrease in train time compared to the Control model trained on raw cirrus images, without compromising binary accuracy. This proof of concept reveals that feature engineering can improve the performance of wildfire detection models by lowering computational expense.