Deep learning has been widely deployed for online ads systems to predict Click-Through Rate (CTR). Machine learning researchers and practitioners frequently retrain CTR models to test their new extracted features. However, the CTR model training often relies on a large number of raw input data logs. Hence, the feature extraction can take a significant proportion of the training time for an industrial-level CTR model. In this paper, we propose FeatureBox, a novel end-to-end training framework that pipelines the feature extraction and the training on GPU servers to save the intermediate I/O of the feature extraction. We rewrite computation-intensive feature extraction operators as GPU operators and leave the memory-intensive operator on CPUs. We introduce a layer-wise operator scheduling algorithm to schedule these heterogeneous operators. We present a light-weight GPU memory management algorithm that supports dynamic GPU memory allocation with minimal overhead. We experimentally evaluate FeatureBox and compare it with the previous in-production feature extraction framework on two real-world ads applications. The results confirm the effectiveness of our proposed method.
Object-oriented reinforcement learning (OORL) is a promising way to improve the sample efficiency and generalization ability over standard RL. Recent works that try to solve OORL tasks without additional feature engineering mainly focus on learning the object representations and then solving tasks via reasoning based on these object representations. However, none of these works tries to explicitly model the inherent similarity between different object instances of the same category. Objects of the same category should share similar functionalities; therefore, the category is the most critical property of an object. Following this insight, we propose a novel framework named Object-Category Aware Reinforcement Learning (OCARL), which utilizes the category information of objects to facilitate both perception and reasoning. OCARL consists of three parts: (1) Category-Aware Unsupervised Object Discovery (UOD), which discovers the objects as well as their corresponding categories; (2) Object-Category Aware Perception, which encodes the category information and is also robust to the incompleteness of (1) at the same time; (3) Object-Centric Modular Reasoning, which adopts multiple independent and object-category-specific networks when reasoning based on objects. Our experiments show that OCARL can improve both the sample efficiency and generalization in the OORL domain.
Smile veracity classification is a task of interpreting social interactions. Broadly, it distinguishes between spontaneous and posed smiles. Previous approaches used hand-engineered features from facial landmarks or considered raw smile videos in an end-to-end manner to perform smile classification tasks. Feature-based methods require intervention from human experts on feature engineering and heavy pre-processing steps. On the contrary, raw smile video inputs fed into end-to-end models bring more automation to the process with the cost of considering many redundant facial features (beyond landmark locations) that are mainly irrelevant to smile veracity classification. It remains unclear to establish discriminative features from landmarks in an end-to-end manner. We present a MeshSmileNet framework, a transformer architecture, to address the above limitations. To eliminate redundant facial features, our landmarks input is extracted from Attention Mesh, a pre-trained landmark detector. Again, to discover discriminative features, we consider the relativity and trajectory of the landmarks. For the relativity, we aggregate facial landmark that conceptually formats a curve at each frame to establish local spatial features. For the trajectory, we estimate the movements of landmark composed features across time by self-attention mechanism, which captures pairwise dependency on the trajectory of the same landmark. This idea allows us to achieve state-of-the-art performances on UVA-NEMO, BBC, MMI Facial Expression, and SPOS datasets.
Feature engineering is required to obtain better results for time series forecasting, and decomposition is a crucial one. One decomposition approach often cannot be used for numerous forecasting tasks since the standard time series decomposition lacks flexibility and robustness. Traditional feature selection relies heavily on preexisting domain knowledge, has no generic methodology, and requires a lot of labor. However, most time series prediction models based on deep learning typically suffer from interpretability issue, so the "black box" results lead to a lack of confidence. To deal with the above issues forms the motivation of the thesis. In the paper we propose TSDFNet as a neural network with self-decomposition mechanism and an attentive feature fusion mechanism, It abandons feature engineering as a preprocessing convention and creatively integrates it as an internal module with the deep model. The self-decomposition mechanism empowers TSDFNet with extensible and adaptive decomposition capabilities for any time series, users can choose their own basis functions to decompose the sequence into temporal and generalized spatial dimensions. Attentive feature fusion mechanism has the ability to capture the importance of external variables and the causality with target variables. It can automatically suppress the unimportant features while enhancing the effective ones, so that users do not have to struggle with feature selection. Moreover, TSDFNet is easy to look into the "black box" of the deep neural network by feature visualization and analyze the prediction results. We demonstrate performance improvements over existing widely accepted models on more than a dozen datasets, and three experiments showcase the interpretability of TSDFNet.
In this paper, we present Position-to-Structure Attention Transformers (PS-Former), a Transformer-based algorithm for 3D point cloud recognition. PS-Former deals with the challenge in 3D point cloud representation where points are not positioned in a fixed grid structure and have limited feature description (only 3D coordinates ($x, y, z$) for scattered points). Existing Transformer-based architectures in this domain often require a pre-specified feature engineering step to extract point features. Here, we introduce two new aspects in PS-Former: 1) a learnable condensation layer that performs point downsampling and feature extraction; and 2) a Position-to-Structure Attention mechanism that recursively enriches the structural information with the position attention branch. Compared with the competing methods, while being generic with less heuristics feature designs, PS-Former demonstrates competitive experimental results on three 3D point cloud tasks including classification, part segmentation, and scene segmentation.
The brain age has been proven to be a phenotype of relevance to cognitive performance and brain disease. Achieving accurate brain age prediction is an essential prerequisite for optimizing the predicted brain-age difference as a biomarker. As a comprehensive biological characteristic, the brain age is hard to be exploited accurately with models using feature engineering and local processing such as local convolution and recurrent operations that process one local neighborhood at a time. Instead, Vision Transformers learn global attentive interaction of patch tokens, introducing less inductive bias and modeling long-range dependencies. In terms of this, we proposed a novel network for learning brain age interpreting with global and local dependencies, where the corresponding representations are captured by Successive Permuted Transformer (SPT) and convolution blocks. The SPT brings computation efficiency and locates the 3D spatial information indirectly via continuously encoding 2D slices from different views. Finally, we collect a large cohort of 22645 subjects with ages ranging from 14 to 97 and our network performed the best among a series of deep learning methods, yielding a mean absolute error (MAE) of 2.855 in validation set, and 2.911 in an independent test set.
Feature transformation aims to extract a good representation (feature) space by mathematically transforming existing features. It is crucial to address the curse of dimensionality, enhance model generalization, overcome data sparsity, and expand the availability of classic models. Current research focuses on domain knowledge-based feature engineering or learning latent representations; nevertheless, these methods are not entirely automated and cannot produce a traceable and optimal representation space. When rebuilding a feature space for a machine learning task, can these limitations be addressed concurrently? In this extension study, we present a self-optimizing framework for feature transformation. To achieve a better performance, we improved the preliminary work by (1) obtaining an advanced state representation for enabling reinforced agents to comprehend the current feature set better; and (2) resolving Q-value overestimation in reinforced agents for learning unbiased and effective policies. Finally, to make experiments more convincing than the preliminary work, we conclude by adding the outlier detection task with five datasets, evaluating various state representation approaches, and comparing different training strategies. Extensive experiments and case studies show that our work is more effective and superior.
Twenty20 cricket, sometimes written Twenty-20, and often abbreviated to T20, is a short form of cricket. In a Twenty20 game the two teams of 11 players have a single innings each, which is restricted to a maximum of 20 overs. This version of cricket is especially unpredictable and is one of the reasons it has gained popularity over recent times. However, in this paper we try four different machine learning approaches for predicting the results of T20 Cricket Matches. Specifically we take in to account: previous performance statistics of the players involved in the competing teams, ratings of players obtained from reputed cricket statistics websites, clustering the players' with similar performance statistics and propose a novel method using an ELO based approach to rate players. We compare the performances of each of these feature engineering approaches by using different ML algorithms, including logistic regression, support vector machines, bayes network, decision tree, random forest.
The finance industry is producing an increasing amount of datasets that investment professionals can consider to be influential on the price of financial assets. These datasets were initially mainly limited to exchange data, namely price, capitalization and volume. Their coverage has now considerably expanded to include, for example, macroeconomic data, supply and demand of commodities, balance sheet data and more recently extra-financial data such as ESG scores. This broadening of the factors retained as influential constitutes a serious challenge for statistical modeling. Indeed, the instability of the correlations between these factors makes it practically impossible to identify the joint laws needed to construct scenarios. Fortunately, spectacular advances in Deep Learning field in recent years have given rise to GANs. GANs are a type of generative machine learning models that produce new data samples with the same characteristics as a training data distribution in an unsupervised way, avoiding data assumptions and human induced biases. In this work, we are exploring the use of GANs for synthetic financial scenarios generation. This pilot study is the result of a collaboration between Fujitsu and Advestis and it will be followed by a thorough exploration of the use cases that can benefit from the proposed solution. We propose a GANs-based algorithm that allows the replication of multivariate data representing several properties (including, but not limited to, price, market capitalization, ESG score, controversy score,. . .) of a set of stocks. This approach differs from examples in the financial literature, which are mainly focused on the reproduction of temporal asset price scenarios. We also propose several metrics to evaluate the quality of the data generated by the GANs. This approach is well fit for the generation of scenarios, the time direction simply arising as a subsequent (eventually conditioned) generation of data points drawn from the learned distribution. Our method will allow to simulate high dimensional scenarios (compared to $\lesssim10$ features currently employed in most recent use cases) where network complexity is reduced thanks to a wisely performed feature engineering and selection. Complete results will be presented in a forthcoming study.
Standardized datasets and benchmarks have spurred innovations in computer vision, natural language processing, multi-modal and tabular settings. We note that, as compared to other well researched fields, fraud detection has unique challenges: high-class imbalance, diverse feature types, frequently changing fraud patterns, and adversarial nature of the problem. Due to these, the modeling approaches evaluated on datasets from other research fields may not work well for the fraud detection. In this paper, we introduce Fraud Dataset Benchmark (FDB), a compilation of publicly available datasets catered to fraud detection FDB comprises variety of fraud related tasks, ranging from identifying fraudulent card-not-present transactions, detecting bot attacks, classifying malicious URLs, estimating risk of loan default to content moderation. The Python based library for FDB provides a consistent API for data loading with standardized training and testing splits. We demonstrate several applications of FDB that are of broad interest for fraud detection, including feature engineering, comparison of supervised learning algorithms, label noise removal, class-imbalance treatment and semi-supervised learning. We hope that FDB provides a common playground for researchers and practitioners in the fraud detection domain to develop robust and customized machine learning techniques targeting various fraud use cases.
Lateral Movement refers to methods by which threat actors gain initial access to a network and then progressively move through said network collecting key data about assets until they reach the ultimate target of their attack. Lateral Movement intrusions have become more intricate with the increasing complexity and interconnected nature of enterprise networks, and require equally sophisticated detection mechanisms to proactively detect such threats in near real-time at enterprise scale. In this paper, the authors propose a novel, lightweight method for Lateral Movement detection using user behavioral analysis and machine learning. Specifically, this paper introduces a novel methodology for cyber domain-specific feature engineering that identifies Lateral Movement behavior on a per-user basis. Furthermore, the engineered features have also been used to develop two supervised machine learning models for Lateral Movement identification that have demonstrably outperformed models previously seen in literature while maintaining robust performance on datasets with high class imbalance. The models and methodology introduced in this paper have also been designed in collaboration with security operators to be relevant and interpretable in order to maximize impact and minimize time to value as a cyber threat detection toolkit. The underlying goal of the paper is to provide a computationally efficient, domain-specific approach to near real-time Lateral Movement detection that is interpretable and robust to enterprise-scale data volumes and class imbalance.
Finger vein recognition is an emerging biometric recognition technology. Different from the other biometric features on the body surface, the venous vascular tissue of the fingers is buried deep inside the skin. Due to this advantage, finger vein recognition is highly stable and private. They are almost impossible to be stolen and difficult to interfere with by external conditions. Unlike the finger vein recognition methods based on traditional machine learning, the artificial neural network technique, especially deep learning, it without relying on feature engineering and have superior performance. To summarize the development of finger vein recognition based on artificial neural networks, this paper collects 149 related papers. First, we introduce the background of finger vein recognition and the motivation of this survey. Then, the development history of artificial neural networks and the representative networks on finger vein recognition tasks are introduced. The public datasets that are widely used in finger vein recognition are then described. After that, we summarize the related finger vein recognition tasks based on classical neural networks and deep neural networks, respectively. Finally, the challenges and potential development directions in finger vein recognition are discussed. To our best knowledge, this paper is the first comprehensive survey focusing on finger vein recognition based on artificial neural networks.
The popularity of automated machine learning (AutoML) tools in different domains has increased over the past few years. Machine learning (ML) practitioners use AutoML tools to automate and optimize the process of feature engineering, model training, and hyperparameter optimization and so on. Recent work performed qualitative studies on practitioners' experiences of using AutoML tools and compared different AutoML tools based on their performance and provided features, but none of the existing work studied the practices of using AutoML tools in real-world projects at a large scale. Therefore, we conducted an empirical study to understand how ML practitioners use AutoML tools in their projects. To this end, we examined the top 10 most used AutoML tools and their respective usages in a large number of open-source project repositories hosted on GitHub. The results of our study show 1) which AutoML tools are mostly used by ML practitioners and 2) the characteristics of the repositories that use these AutoML tools. Also, we identified the purpose of using AutoML tools (e.g. model parameter sampling, search space management, model evaluation/error-analysis, Data/ feature transformation, and data labeling) and the stages of the ML pipeline (e.g. feature engineering) where AutoML tools are used. Finally, we report how often AutoML tools are used together in the same source code files. We hope our results can help ML practitioners learn about different AutoML tools and their usages, so that they can pick the right tool for their purposes. Besides, AutoML tool developers can benefit from our findings to gain insight into the usages of their tools and improve their tools to better fit the users' usages and needs.
We developed an inherently interpretable multilevel Bayesian framework for representing variation in regression coefficients that mimics the piecewise linearity of ReLU-activated deep neural networks. We used the framework to formulate a survival model for using medical claims to predict hospital readmission and death that focuses on discharge placement, adjusting for confounding in estimating causal local average treatment effects. We trained the model on a 5% sample of Medicare beneficiaries from 2008 and 2011, based on their 2009--2011 inpatient episodes, and then tested the model on 2012 episodes. The model scored an AUROC of approximately 0.76 on predicting all-cause readmissions -- defined using official Centers for Medicare and Medicaid Services (CMS) methodology -- or death within 30-days of discharge, being competitive against XGBoost and a Bayesian deep neural network, demonstrating that one need-not sacrifice interpretability for accuracy. Crucially, as a regression model, we provide what blackboxes cannot -- the exact gold-standard global interpretation of the model, identifying relative risk factors and quantifying the effect of discharge placement. We also show that the posthoc explainer SHAP fails to provide accurate explanations.
Artificial Intelligence-based (AI) analysis of large, curated medical datasets is promising for providing early detection, faster diagnosis, and more effective treatment using low-power Electrocardiography (ECG) monitoring devices information. However, accessing sensitive medical data from diverse sources is highly restricted since improper use, unsafe storage, or data leakage could violate a person's privacy. This work uses a Federated Learning (FL) privacy-preserving methodology to train AI models over heterogeneous sets of high-definition ECG from 12-lead sensor arrays collected from six heterogeneous sources. We evaluated the capacity of the resulting models to achieve equivalent performance compared to state-of-the-art models trained in a Centralized Learning (CL) fashion. Moreover, we assessed the performance of our solution over Independent and Identical distributed (IID) and non-IID federated data. Our methodology involves machine learning techniques based on Deep Neural Networks and Long-Short-Term Memory models. It has a robust data preprocessing pipeline with feature engineering, selection, and data balancing techniques. Our AI models demonstrated comparable performance to models trained using CL, IID, and non-IID approaches. They showcased advantages in reduced complexity and faster training time, making them well-suited for cloud-edge architectures.
Over recent years, there has been a rapid development of deep learning (DL) in both industry and academia fields. However, finding the optimal hyperparameters of a DL model often needs high computational cost and human expertise. To mitigate the above issue, evolutionary computation (EC) as a powerful heuristic search approach has shown significant merits in the automated design of DL models, so-called evolutionary deep learning (EDL). This paper aims to analyze EDL from the perspective of automated machine learning (AutoML). Specifically, we firstly illuminate EDL from machine learning and EC and regard EDL as an optimization problem. According to the DL pipeline, we systematically introduce EDL methods ranging from feature engineering, model generation, to model deployment with a new taxonomy (i.e., what and how to evolve/optimize), and focus on the discussions of solution representation and search paradigm in handling the optimization problem by EC. Finally, key applications, open issues and potentially promising lines of future research are suggested. This survey has reviewed recent developments of EDL and offers insightful guidelines for the development of EDL.
Automatically estimating the complexity of texts for readers has a variety of applications, such as recommending texts with an appropriate complexity level to language learners or supporting the evaluation of text simplification approaches. In this paper, we present our submission to the Text Complexity DE Challenge 2022, a regression task where the goal is to predict the complexity of a German sentence for German learners at level B. Our approach relies on more than 220,000 pseudo-labels created from the German Wikipedia and other corpora to train Transformer-based models, and refrains from any feature engineering or any additional, labeled data. We find that the pseudo-label-based approach gives impressive results yet requires little to no adjustment to the specific task and therefore could be easily adapted to other domains and tasks.
Respiratory rate (RR) is an important biomarker as RR changes can reflect severe medical events such as heart disease, lung disease, and sleep disorders. Unfortunately, standard manual RR counting is prone to human error and cannot be performed continuously. This study proposes a method for continuously estimating RR, RRWaveNet. The method is a compact end-to-end deep learning model which does not require feature engineering and can use low-cost raw photoplethysmography (PPG) as input signal. RRWaveNet was tested subject-independently and compared to baseline in four datasets (BIDMC, CapnoBase, WESAD, and SensAI) and using three window sizes (16, 32, and 64 seconds). RRWaveNet outperformed current state-of-the-art methods with mean absolute errors at optimal window size of 1.66 \pm 1.01, 1.59 \pm 1.08, 1.92 \pm 0.96 and 1.23 \pm 0.61 breaths per minute for each dataset. In remote monitoring settings, such as in the WESAD and SensAI datasets, we apply transfer learning to improve the performance using two other ICU datasets as pretraining datasets, reducing the MAE by up to 21$\%$. This shows that this model allows accurate and practical estimation of RR on affordable and wearable devices. Our study also shows feasibility of remote RR monitoring in the context of telemedicine and at home.
KDD CUP 2022 proposes a time-series forecasting task on spatial dynamic wind power dataset, in which the participants are required to predict the future generation given the historical context factors. The evaluation metrics contain RMSE and MAE. This paper describes the solution of Team 88VIP, which mainly comprises two types of models: a gradient boosting decision tree to memorize the basic data patterns and a recurrent neural network to capture the deep and latent probabilistic transitions. Ensembling these models contributes to tackle the fluctuation of wind power, and training submodels targets on the distinguished properties in heterogeneous timescales of forecasting, from minutes to days. In addition, feature engineering, imputation techniques and the design of offline evaluation are also described in details. The proposed solution achieves an overall online score of -45.213 in Phase 3.
Fatal diseases, as Critical Health Episodes (CHEs), represent real dangers for patients hospitalized in Intensive Care Units. These episodes can lead to irreversible organ damage and death. Nevertheless, diagnosing them in time would greatly reduce their inconvenience. This study therefore focused on building a highly effective early warning system for CHEs such as Acute Hypotensive Episodes and Tachycardia Episodes. To facilitate the precocity of the prediction, a gap of one hour was considered between the observation periods (Observation Windows) and the periods during which a critical event can occur (Target Windows). The MIMIC II dataset was used to evaluate the performance of the proposed system. This system first includes extracting additional features using three different modes. Then, the feature selection process allowing the selection of the most relevant features was performed using the Mutual Information Gain feature importance. Finally, the high-performance predictive model LightGBM was used to perform episode classification. This approach called MIG-LightGBM was evaluated using five different metrics: Event Recall (ER), Reduced Precision (RP), average Anticipation Time (aveAT), average False Alarms (aveFA), and Event F1-score (EF1-score). A method is therefore considered highly efficient for the early prediction of CHEs if it exhibits not only a large aveAT but also a large EF1-score and a low aveFA. Compared to systems using Extreme Gradient Boosting, Support Vector Classification or Naive Bayes as a predictive model, the proposed system was found to be highly dominant. It also confirmed its superiority over the Layered Learning approach.