Pitch and meter are two fundamental music features for symbolic music generation tasks, where researchers usually choose different encoding methods depending on specific goals. However, the advantages and drawbacks of different encoding methods have not been frequently discussed. This paper presents a integrated analysis of the influence of two low-level feature, pitch and meter, on the performance of a token-based sequential music generation model. First, the commonly used MIDI number encoding and a less used class-octave encoding are compared. Second, an dense intra-bar metric grid is imposed to the encoded sequence as auxiliary features. Different complexity and resolutions of the metric grid are compared. For complexity, the single token approach and the multiple token approach are compared; for grid resolution, 0 (ablation), 1 (bar-level), 4 (downbeat-level) 12, (8th-triplet-level) up to 64 (64th-note-grid-level) are compared; for duration resolution, 4, 8, 12 and 16 subdivisions per beat are compared. All different encodings are tested on separately trained Transformer-XL models for a melody generation task. Regarding distribution similarity of several objective evaluation metrics to the test dataset, results suggest that the class-octave encoding significantly outperforms the taken-for-granted MIDI encoding on pitch-related metrics; finer grids and multiple-token grids improve the rhythmic quality, but also suffer from over-fitting at early training stage. Results display a general phenomenon of over-fitting from two aspects, the pitch embedding space and the test loss of the single-token grid encoding. From a practical perspective, we both demonstrate the feasibility and raise the concern of easy over-fitting problem of using smaller networks and lower embedding dimensions on the generation task. The findings can also contribute to futural models in terms of feature engineering.
Ranking algorithms in traditional search engines are powered by enormous training data sets that are meticulously engineered and curated by a centralized entity. Decentralized peer-to-peer (p2p) networks such as torrenting applications and Web3 protocols deliberately eschew centralized databases and computational architectures when designing services and features. As such, robust search-and-rank algorithms designed for such domains must be engineered specifically for decentralized networks, and must be lightweight enough to operate on consumer-grade personal devices such as a smartphone or laptop computer. We introduce G-Rank, an unsupervised ranking algorithm designed exclusively for decentralized networks. We demonstrate that accurate, relevant ranking results can be achieved in fully decentralized networks without any centralized data aggregation, feature engineering, or model training. Furthermore, we show that such results are obtainable with minimal data preprocessing and computational overhead, and can still return highly relevant results even when a user's device is disconnected from the network. G-Rank is highly modular in design, is not limited to categorical data, and can be implemented in a variety of domains with minimal modification. The results herein show that unsupervised ranking models designed for decentralized p2p networks are not only viable, but worthy of further research.
Cancer is a term that denotes a group of diseases caused by abnormal growth of cells that can spread in different parts of the body. According to the World Health Organization (WHO), cancer is the second major cause of death after cardiovascular diseases. Gene expression can play a fundamental role in the early detection of cancer, as it is indicative of the biochemical processes in tissue and cells, as well as the genetic characteristics of an organism. Deoxyribonucleic Acid (DNA) microarrays and Ribonucleic Acid (RNA)- sequencing methods for gene expression data allow quantifying the expression levels of genes and produce valuable data for computational analysis. This study reviews recent progress in gene expression analysis for cancer classification using machine learning methods. Both conventional and deep learning-based approaches are reviewed, with an emphasis on the ap-plication of deep learning models due to their comparative advantages for identifying gene patterns that are distinctive for various types of cancers. Relevant works that employ the most commonly used deep neural network architectures are covered, including multi-layer perceptrons, convolutional, recurrent, graph, and transformer networks. This survey also presents an overview of the data collection methods for gene expression analysis and lists important datasets that are commonly used for supervised machine learning for this task. Furthermore, reviewed are pertinent techniques for feature engineering and data preprocessing that are typically used to handle the high dimensionality of gene expression data, caused by a large number of genes present in data samples. The paper concludes with a discussion of future research directions for machine learning-based gene expression analysis for cancer classification.
This paper presents our solution for the Real Robot Challenge (RRC) III, a competition featured in the NeurIPS 2022 Competition Track, aimed at addressing dexterous robotic manipulation tasks through learning from pre-collected offline data. Participants were provided with two types of datasets for each task: expert and mixed datasets with varying skill levels. While the simplest offline policy learning algorithm, Behavioral Cloning (BC), performed remarkably well when trained on expert datasets, it outperformed even the most advanced offline reinforcement learning (RL) algorithms. However, BC's performance deteriorated when applied to mixed datasets, and the performance of offline RL algorithms was also unsatisfactory. Upon examining the mixed datasets, we observed that they contained a significant amount of expert data, although this data was unlabeled. To address this issue, we proposed a semi-supervised learning-based classifier to identify the underlying expert behavior within mixed datasets, effectively isolating the expert data. To further enhance BC's performance, we leveraged the geometric symmetry of the RRC arena to augment the training dataset through mathematical transformations. In the end, our submission surpassed that of all other participants, even those who employed complex offline RL algorithms and intricate data processing and feature engineering techniques.
Data-driven intelligent computational design (DICD) is a research hotspot emerged under the context of fast-developing artificial intelligence. It emphasizes on utilizing deep learning algorithms to extract and represent the design features hidden in historical or fabricated design process data, and then learn the combination and mapping patterns of these design features for the purposes of design solution retrieval, generation, optimization, evaluation, etc. Due to its capability of automatically and efficiently generating design solutions and thus supporting human-in-the-loop intelligent and innovative design activities, DICD has drawn the attentions from both academic and industrial fields. However, as an emerging research subject, there are still many unexplored issues that limit the development and application of DICD, such as specific dataset building, engineering design related feature engineering, systematic methods and techniques for DICD implementation in the entire product design process, etc. In this regard, a systematic and operable road map for DICD implementation from full-process perspective is established, including a general workflow for DICD project planning, an overall framework for DICD project implementation, the computing mechanisms for DICD implementation, key enabling technologies for detailed DICD implementation, and three application scenarios of DICD. The road map reveals the common mechanisms and calculation principles of existing DICD researches, and thus it can provide systematic guidance for the possible DICD applications that have not been explored.
Many parameters of interest in economics and other social sciences depend on generated regressors. Examples in economics include structural parameters in models with endogenous variables estimated by control functions and in models with sample selection, treatment effect estimation with propensity score matching, and marginal treatment effects. More recently, Machine Learning (ML) generated regressors are becoming ubiquitous for these and other applications such as imputation with missing regressors, dimension reduction, including autoencoders, learned proxies, confounders and treatments, and for feature engineering with unstructured data, among others. We provide the first general method for valid inference with regressors generated from ML. Inference with generated regressors is complicated by the very complex expression for influence functions and asymptotic variances. Additionally, ML-generated regressors may lead to large biases in downstream inferences. To address these problems, we propose Automatic Locally Robust/debiased GMM estimators in a general three-step setting with ML-generated regressors. We illustrate our results with treatment effects and counterfactual parameters in the partially linear and nonparametric models with ML-generated regressors. We provide sufficient conditions for the asymptotic normality of our debiased GMM estimators and investigate their finite-sample performance through Monte Carlo simulations.
Electrocardiography (ECG) signals can be considered as multi-variable time-series. The state-of-the-art ECG data classification approaches, based on either feature engineering or deep learning techniques, treat separately spectral and time domains in machine learning systems. No spectral-time domain communication mechanism inside the classifier model can be found in current approaches, leading to difficulties in identifying complex ECG forms. In this paper, we proposed a novel deep learning model named Spectral Cross-domain neural network (SCDNN) with a new block called Soft-adaptive threshold spectral enhancement (SATSE), to simultaneously reveal the key information embedded in spectral and time domains inside the neural network. More precisely, the domain-cross information is captured by a general Convolutional neural network (CNN) backbone, and different information sources are merged by a self-adaptive mechanism to mine the connection between time and spectral domains. In SATSE, the knowledge from time and spectral domains is extracted via the Fast Fourier Transformation (FFT) with soft trainable thresholds in modified Sigmoid functions. The proposed SCDNN is tested with several classification tasks implemented on the public ECG databases \textit{PTB-XL} and \textit{MIT-BIH}. SCDNN outperforms the state-of-the-art approaches with a low computational cost regarding a variety of metrics in all classification tasks on both databases, by finding appropriate domains from the infinite spectral mapping. The convergence of the trainable thresholds in the spectral domain is also numerically investigated in this paper. The robust performance of SCDNN provides a new perspective to exploit knowledge across deep learning models from time and spectral domains. The repository can be found: https://github.com/DL-WG/SCDNN-TS
Cardiac image segmentation is a powerful tool in regard to diagnostics and treatment of cardiovascular diseases. Purely feature-based detection of anatomical structures like the mitral valve is a laborious task due to specifically required feature engineering and is especially challenging in echocardiograms, because of their inherently low contrast and blurry boundaries between some anatomical structures. With the publication of further annotated medical datasets and the increase in GPU processing power, deep learning-based methods in medical image segmentation became more feasible in the past years. We propose a fully automatic detection method for mitral valve hinge points, which uses a U-Net based deep neural net to segment cardiac chambers in echocardiograms in a first step, and subsequently extracts the mitral valve hinge points from the resulting segmentations in a second step. Results measured with this automatic detection method were compared to reference coordinate values, which with median absolute hinge point coordinate errors of 1.35 mm for the x- (15-85 percentile range: [0.3 mm; 3.15 mm]) and 0.75 mm for the y- coordinate (15-85 percentile range: [0.15 mm; 1.88 mm]).
Automatic term extraction (ATE) is a Natural Language Processing (NLP) task that eases the effort of manually identifying terms from domain-specific corpora by providing a list of candidate terms. As units of knowledge in a specific field of expertise, extracted terms are not only beneficial for several terminographical tasks, but also support and improve several complex downstream tasks, e.g., information retrieval, machine translation, topic detection, and sentiment analysis. ATE systems, along with annotated datasets, have been studied and developed widely for decades, but recently we observed a surge in novel neural systems for the task at hand. Despite a large amount of new research on ATE, systematic survey studies covering novel neural approaches are lacking. We present a comprehensive survey of deep learning-based approaches to ATE, with a focus on Transformer-based neural models. The study also offers a comparison between these systems and previous ATE approaches, which were based on feature engineering and non-neural supervised learning algorithms.
Machine learning models work better when curated features are provided to them. Feature engineering methods have been usually used as a preprocessing step to obtain or build a proper feature set. In late years, autoencoders (a specific type of symmetrical neural network) have been widely used to perform representation learning, proving their competitiveness against classical feature engineering algorithms. The main obstacle in the use of autoencoders is finding a good architecture, a process that most experts confront manually. An automated autoencoder architecture search procedure, based on evolutionary methods, is proposed in this paper. The methodology is tested against nine heterogeneous data sets. The obtained results show the ability of this approach to find better architectures, able to concentrate most of the useful information in a minimized coding, in a reduced time.
Through the generalization of deep learning, the research community has addressed critical challenges in the network security domain, like malware identification and anomaly detection. However, they have yet to discuss deploying them on Internet of Things (IoT) devices for day-to-day operations. IoT devices are often limited in memory and processing power, rendering the compute-intensive deep learning environment unusable. This research proposes a way to overcome this barrier by bypassing feature engineering in the deep learning pipeline and using raw packet data as input. We introduce a feature engineering-less machine learning (ML) process to perform malware detection on IoT devices. Our proposed model, "Feature engineering-less-ML (FEL-ML)," is a lighter-weight detection algorithm that expends no extra computations on "engineered" features. It effectively accelerates the low-powered IoT edge. It is trained on unprocessed byte-streams of packets. Aside from providing better results, it is quicker than traditional feature-based methods. FEL-ML facilitates resource-sensitive network traffic security with the added benefit of eliminating the significant investment by subject matter experts in feature engineering.
Data-driven modeling is an approach in energy systems modeling that has been gaining popularity. In data-driven modeling, machine learning methods such as linear regression, neural networks or decision-tree based methods are being applied. While these methods do not require domain knowledge, they are sensitive to data quality. Therefore, improving data quality in a dataset is beneficial for creating machine learning-based models. The improvement of data quality can be implemented through preprocessing methods. A selected type of preprocessing is feature engineering, which focuses on evaluating and improving the quality of certain features inside the dataset. Feature engineering methods include methods such as feature creation, feature expansion, or feature selection. In this work, a Python framework containing different feature engineering methods is presented. This framework contains different methods for feature creation, expansion and selection; in addition, methods for transforming or filtering data are implemented. The implementation of the framework is based on the Python library scikit-learn. The framework is demonstrated on a case study of a use case from energy demand prediction. A data-driven model is created including selected feature engineering methods. The results show an improvement in prediction accuracy through the engineered features.
Learning feature interactions is the key to success for the large-scale CTR prediction and recommendation. In practice, handcrafted feature engineering usually requires exhaustive searching. In order to reduce the high cost of human efforts in feature engineering, researchers propose several deep neural networks (DNN)-based approaches to learn the feature interactions in an end-to-end fashion. However, existing methods either do not learn both vector-wise interactions and bit-wise interactions simultaneously, or fail to combine them in a controllable manner. In this paper, we propose a new model, xDeepInt, based on a novel network architecture called polynomial interaction network (PIN) which learns higher-order vector-wise interactions recursively. By integrating subspace-crossing mechanism, we enable xDeepInt to balance the mixture of vector-wise and bit-wise feature interactions at a bounded order. Based on the network architecture, we customize a combined optimization strategy to conduct feature selection and interaction selection. We implement the proposed model and evaluate the model performance on three real-world datasets. Our experiment results demonstrate the efficacy and effectiveness of xDeepInt over state-of-the-art models. We open-source the TensorFlow implementation of xDeepInt: https://github.com/yanyachen/xDeepInt.
The application of deep learning to non-stationary temporal datasets can lead to overfitted models that underperform under regime changes. In this work, we propose a modular machine learning pipeline for ranking predictions on temporal panel datasets which is robust under regime changes. The modularity of the pipeline allows the use of different models, including Gradient Boosting Decision Trees (GBDTs) and Neural Networks, with and without feature engineering. We evaluate our framework on financial data for stock portfolio prediction, and find that GBDT models with dropout display high performance, robustness and generalisability with reduced complexity and computational cost. We then demonstrate how online learning techniques, which require no retraining of models, can be used post-prediction to enhance the results. First, we show that dynamic feature projection improves robustness by reducing drawdown in regime changes. Second, we demonstrate that dynamical model ensembling based on selection of models with good recent performance leads to improved Sharpe and Calmar ratios of out-of-sample predictions. We also evaluate the robustness of our pipeline across different data splits and random seeds with good reproducibility.
This study presents a general machine learning framework to estimate the traffic-measurement-level experience rate at given throughput values in the form of a Key Performance Indicator for the cells on base stations across various cities, using busy-hour counter data, and several technical parameters together with the network topology. Relying on feature engineering techniques, scores of additional predictors are proposed to enhance the effects of raw correlated counter values over the corresponding targets, and to represent the underlying interactions among groups of cells within nearby spatial locations effectively. An end-to-end regression modeling is applied on the transformed data, with results presented on unseen cities of varying sizes.
Tool flank wear monitoring can minimize machining downtime costs while increasing productivity and product quality. In some industrial applications, only a limited level of tool wear is allowed to attain necessary tolerances. It may become challenging to monitor a limited level of tool wear in the data collected from the machine due to the other components, such as the flexible vibrations of the machine, dominating the measurement signals. In this study, a tool wear monitoring technique to predict limited levels of tool wear from the spindle motor current and dynamometer measurements is presented. High-frequency spindle motor current data is collected with an industrial edge device while the cutting forces and torque are measured with a rotary dynamometer in drilling tests for a selected number of holes. Feature engineering is conducted to identify the statistical features of the measurement signals that are most sensitive to small changes in tool wear. A neural network based on the long short-term memory (LSTM) architecture is developed to predict tool flank wear from the measured spindle motor current and dynamometer signals. It is demonstrated that the proposed technique predicts tool flank wear with good accuracy and high computational efficiency. The proposed technique can easily be implemented in an industrial edge device as a real-time predictive maintenance application to minimize the costs due to manufacturing downtime and tool underuse or overuse.
Automated Feature Engineering (AFE) refers to automatically generate and select optimal feature sets for downstream tasks, which has achieved great success in real-world applications. Current AFE methods mainly focus on improving the effectiveness of the produced features, but ignoring the low-efficiency issue for large-scale deployment. Therefore, in this work, we propose a generic framework to improve the efficiency of AFE. Specifically, we construct the AFE pipeline based on reinforcement learning setting, where each feature is assigned an agent to perform feature transformation \com{and} selection, and the evaluation score of the produced features in downstream tasks serve as the reward to update the policy. We improve the efficiency of AFE in two perspectives. On the one hand, we develop a Feature Pre-Evaluation (FPE) Model to reduce the sample size and feature size that are two main factors on undermining the efficiency of feature evaluation. On the other hand, we devise a two-stage policy training strategy by running FPE on the pre-evaluation task as the initialization of the policy to avoid training policy from scratch. We conduct comprehensive experiments on 36 datasets in terms of both classification and regression tasks. The results show $2.9\%$ higher performance in average and 2x higher computational efficiency comparing to state-of-the-art AFE methods.
The arise of cutting-edge technologies and services such as XR promise to change the concepts of how day-to-day things are done. At the same time, the appearance of modern and decentralized architectures approaches has given birth to a new generation of mobile networks such as 5G, as well as outlining the roadmap for B5G and posterior. These networks are expected to be the enablers for bringing to life the Metaverse and other futuristic approaches. In this sense, this work presents an ML-based (Machine Learning) framework that allows the estimation of service Key Quality Indicators (KQIs). For this, only information reachable to operators is required, such as statistics and configuration parameters from these networks. This strategy prevents operators from avoiding intrusion into the user data and guaranteeing privacy. To test this proposal, 360-Video has been selected as a use case of Virtual Reality (VR), from which specific KQIs are estimated such as video resolution, frame rate, initial startup time, throughput, and latency, among others. To select the best model for each KQI, a search grid with a cross-validation strategy has been used to determine the best hyperparameter tuning. To boost the creation of each KQI model, feature engineering techniques together with cross-validation strategies have been used. The performance is assessed using MAE (Mean Average Error) and the prediction time. The outcomes point out that KNR (K-Near Neighbors) and RF (Random Forest) are the best algorithms in combination with Feature Selection techniques. Likewise, this work will help as a baseline for E2E-Quality-of-Experience-based network management working in conjunction with network slicing, virtualization, and MEC, among other enabler technologies.
A methodology is proposed, which addresses the caveat that line-of-sight emission spectroscopy presents in that it cannot provide spatially resolved temperature measurements in nonhomogeneous temperature fields. The aim of this research is to explore the use of data-driven models in measuring temperature distributions in a spatially resolved manner using emission spectroscopy data. Two categories of data-driven methods are analyzed: (i) Feature engineering and classical machine learning algorithms, and (ii) end-to-end convolutional neural networks (CNN). In total, combinations of fifteen feature groups and fifteen classical machine learning models, and eleven CNN models are considered and their performances explored. The results indicate that the combination of feature engineering and machine learning provides better performance than the direct use of CNN. Notably, feature engineering which is comprised of physics-guided transformation, signal representation-based feature extraction and Principal Component Analysis is found to be the most effective. Moreover, it is shown that when using the extracted features, the ensemble-based, light blender learning model offers the best performance with RMSE, RE, RRMSE and R values of 64.3, 0.017, 0.025 and 0.994, respectively. The proposed method, based on feature engineering and the light blender model, is capable of measuring nonuniform temperature distributions from low-resolution spectra, even when the species concentration distribution in the gas mixtures is unknown.
This paper aims to improve the Warping Planer Object Detection Network (WPOD-Net) using feature engineering to increase accuracy. What problems are solved using the Warping Object Detection Network using feature engineering? More specifically, we think that it makes sense to add knowledge about edges in the image to enhance the information for determining the license plate contour of the original WPOD-Net model. The Sobel filter has been selected experimentally and acts as a Convolutional Neural Network layer, the edge information is combined with the old information of the original network to create the final embedding vector. The proposed model was compared with the original model on a set of data that we collected for evaluation. The results are evaluated through the Quadrilateral Intersection over Union value and demonstrate that the model has a significant improvement in performance.