Recent years have witnessed the great success of self-supervised learning (SSL) in recommendation systems. However, SSL recommender models are likely to suffer from spurious correlations, leading to poor generalization. To mitigate spurious correlations, existing work usually pursues ID-based SSL recommendation or utilizes feature engineering to identify spurious features. Nevertheless, ID-based SSL approaches sacrifice the positive impact of invariant features, while feature engineering methods require high-cost human labeling. To address the problems, we aim to automatically mitigate the effect of spurious correlations. This objective requires to 1) automatically mask spurious features without supervision, and 2) block the negative effect transmission from spurious features to other features during SSL. To handle the two challenges, we propose an invariant feature learning framework, which first divides user-item interactions into multiple environments with distribution shifts and then learns a feature mask mechanism to capture invariant features across environments. Based on the mask mechanism, we can remove the spurious features for robust predictions and block the negative effect transmission via mask-guided feature augmentation. Extensive experiments on two datasets demonstrate the effectiveness of the proposed framework in mitigating spurious correlations and improving the generalization abilities of SSL models. The code is available at https://github.com/Linxyhaha/IFL.
Any organization needs to improve their products, services, and processes. In this context, engaging with customers and understanding their journey is essential. Organizations have leveraged various techniques and technologies to support customer engagement, from call centres to chatbots and virtual agents. Recently, these systems have used Machine Learning (ML) and Natural Language Processing (NLP) to analyze large volumes of customer feedback and engagement data. The goal is to understand customers in context and provide meaningful answers across various channels. Despite multiple advances in Conversational Artificial Intelligence (AI) and Recommender Systems (RS), it is still challenging to understand the intent behind customer questions during the customer journey. To address this challenge, in this paper, we study and analyze the recent work in Conversational Recommender Systems (CRS) in general and, more specifically, in chatbot-based CRS. We introduce a pipeline to contextualize the input utterances in conversations. We then take the next step towards leveraging reverse feature engineering to link the contextualized input and learning model to support intent recognition. Since performance evaluation is achieved based on different ML models, we use transformer base models to evaluate the proposed approach using a labelled dialogue dataset (MSDialogue) of question-answering interactions between information seekers and answer providers.
Estimating causal effects has become an integral part of most applied fields. In this work we consider the violation of the classical no-interference assumption with units connected by a network. For tractability, we consider a known network that describes how interference may spread. Unlike previous work the radius (and intensity) of the interference experienced by a unit is unknown and can depend on different (local) sub-networks and the assigned treatments. We study estimators for the average direct treatment effect on the treated in such a setting under additive treatment effects. We establish rates of convergence and distributional results. The proposed estimators considers all possible radii for each (local) treatment assignment pattern. In contrast to previous work, we approximate the relevant network interference patterns that lead to good estimates of the interference. To handle feature engineering, a key innovation is to propose the use of synthetic treatments to decouple the dependence. We provide simulations, an empirical illustration and insights for the general study of interference.
With the increasing enrichment and development of the financial derivatives market, the frequency of transactions is also faster and faster. Due to human limitations, algorithms and automatic trading have recently become the focus of discussion. In this paper, we propose a bidirectional LSTM neural network based on an attention mechanism, which is based on two popular assets, gold and bitcoin. In terms of Feature Engineering, on the one hand, we add traditional technical factors, and at the same time, we combine time series models to develop factors. In the selection of model parameters, we finally chose a two-layer deep learning network. According to AUC measurement, the accuracy of bitcoin and gold is 71.94% and 73.03% respectively. Using the forecast results, we achieved a return of 1089.34% in two years. At the same time, we also compare the attention Bi-LSTM model proposed in this paper with the traditional model, and the results show that our model has the best performance in this data set. Finally, we discuss the significance of the model and the experimental results, as well as the possible improvement direction in the future.
Full electronic automation in stock exchanges has recently become popular, generating high-frequency intraday data and motivating the development of near real-time price forecasting methods. Machine learning algorithms are widely applied to mid-price stock predictions. Processing raw data as inputs for prediction models (e.g., data thinning and feature engineering) can primarily affect the performance of the prediction methods. However, researchers rarely discuss this topic. This motivated us to propose three novel modelling strategies for processing raw data. We illustrate how our novel modelling strategies improve forecasting performance by analyzing high-frequency data of the Dow Jones 30 component stocks. In these experiments, our strategies often lead to statistically significant improvement in predictions. The three strategies improve the F1 scores of the SVM models by 0.056, 0.087, and 0.016, respectively.
Choosing which properties of the data to use as input to multivariate decision algorithms -- a.k.a. feature selection -- is an important step in solving any problem with machine learning. While there is a clear trend towards training sophisticated deep networks on large numbers of relatively unprocessed inputs (so-called automated feature engineering), for many tasks in physics, sets of theoretically well-motivated and well-understood features already exist. Working with such features can bring many benefits, including greater interpretability, reduced training and run time, and enhanced stability and robustness. We develop a new feature selection method based on Distance Correlation (DisCo), and demonstrate its effectiveness on the tasks of boosted top- and $W$-tagging. Using our method to select features from a set of over 7,000 energy flow polynomials, we show that we can match the performance of much deeper architectures, by using only ten features and two orders-of-magnitude fewer model parameters.
Tasks critical to enterprise profitability, such as customer churn prediction, fraudulent account detection or customer lifetime value estimation, are often tackled by models trained on features engineered from customer data in tabular format. Application-specific feature engineering adds development, operationalization and maintenance costs over time. Recent advances in representation learning present an opportunity to simplify and generalize feature engineering across applications. When applying these advancements to tabular data researchers deal with data heterogeneity, variations in customer engagement history or the sheer volume of enterprise datasets. In this paper, we propose a novel approach to encode tabular data containing customer transactions, purchase history and other interactions into a generic representation of a customer's association with the business. We then evaluate these embeddings as features to train multiple models spanning a variety of applications. CASPR, Customer Activity Sequence-based Prediction and Representation, applies Transformer architecture to encode activity sequences to improve model performance and avoid bespoke feature engineering across applications. Our experiments at scale validate CASPR for both small and large enterprise applications.
We present MargNet, a deep learning-based classifier for identifying stars, quasars and compact galaxies using photometric parameters and images from the Sloan Digital Sky Survey (SDSS) Data Release 16 (DR16) catalogue. MargNet consists of a combination of Convolutional Neural Network (CNN) and Artificial Neural Network (ANN) architectures. Using a carefully curated dataset consisting of 240,000 compact objects and an additional 150,000 faint objects, the machine learns classification directly from the data, minimising the need for human intervention. MargNet is the first classifier focusing exclusively on compact galaxies and performs better than other methods to classify compact galaxies from stars and quasars, even at fainter magnitudes. This model and feature engineering in such deep learning architectures will provide greater success in identifying objects in the ongoing and upcoming surveys, such as Dark Energy Survey (DES) and images from the Vera C. Rubin Observatory.
The electroencephalogram (EEG) is one of the most precious technologies to understand the happenings inside our brain and further understand our body's happenings. Automatic prediction of oncoming seizures using the EEG signals helps the doctors and clinical experts and reduces their workload. This paper proposes an end-to-end deep learning algorithm to fully automate seizure prediction's laborious task without any heavy pre-processing on the EEG data or feature engineering. The proposed deep learning network is a blend of signal processing and deep learning pipeline, which automates the seizure prediction framework using the EEG signals. This proposed model was evaluated on an open EEG dataset, CHB-MIT. The network achieved an average sensitivity of 97.746\text{\%} and a false positive rate (FPR) of 0.2373 per hour.
Supervised deep learning is most commonly applied to difficult problems defined on large and often extensively curated datasets. Here we demonstrate the ability of deep representation learning to address problems of classification and regression from small and poorly formed tabular datasets by encoding input information as abstracted sequences composed of a fixed number of characters per input field. We find that small models have sufficient capacity for approximation of various functions and achieve record classification benchmark accuracy. Such models are shown to form useful embeddings of various input features in their hidden layers, even if the learned task does not explicitly require knowledge of those features. These models are also amenable to input attribution, allowing for an estimation of the importance of each input element to the model output as well as of which inputs features are effectively embedded in the model. We present a proof-of-concept for the application of small language models to mixed tabular data without explicit feature engineering, cleaning, or preprocessing, relying on the model to perform these tasks as part of the representation learning process.
The emergence of pre-trained language models (PLMs) has shown great success in many Natural Language Processing (NLP) tasks including text classification. Due to the minimal to no feature engineering required when using these models, PLMs are becoming the de facto choice for any NLP task. However, for domain-specific corpora (e.g., financial, legal, and industrial), fine-tuning a pre-trained model for a specific task has shown to provide a performance improvement. In this paper, we compare the performance of four different PLMs on three public domain-free datasets and a real-world dataset containing domain-specific words, against a simple SVM linear classifier with TFIDF vectorized text. The experimental results on the four datasets show that using PLMs, even fine-tuned, do not provide significant gain over the linear SVM classifier. Hence, we recommend that for text classification tasks, traditional SVM along with careful feature engineering can pro-vide a cheaper and superior performance than PLMs.
The events of recent years have highlighted the importance of telemedicine solutions which could potentially allow remote treatment and diagnosis. Relatedly, Computational Paralinguistics, a unique subfield of Speech Processing, aims to extract information about the speaker and form an important part of telemedicine applications. In this work, we focus on two paralinguistic problems: mask detection and breathing state prediction. Solutions developed for these tasks could be invaluable and have the potential to help monitor and limit the spread of a virus like COVID-19. The current state-of-the-art methods proposed for these tasks are ensembles based on deep neural networks like ResNets in conjunction with feature engineering. Although these ensembles can achieve high accuracy, they also have a large footprint and require substantial computational power reducing portability to devices with limited resources. These drawbacks also mean that the previously proposed solutions are infeasible to be used in a telemedicine system due to their size and speed. On the other hand, employing lighter feature-engineered systems can be laborious and add further complexity making them difficult to create a deployable system quickly. This work proposes an ensemble-based automatic feature selection method to enable the development of fast and memory-efficient systems. In particular, we propose an output-gradient-based method to discover essential features using large, well-performing ensembles before training a smaller one. In our experiments, we observed considerable (25-32%) reductions in inference times using neural network ensembles based on output-gradient-based features. Our method offers a simple way to increase the speed of the system and enable real-time usage while maintaining competitive results with larger-footprint ensemble using all spectral features.
Large Language Models (LLMs) have shown impressive results on a variety of text understanding tasks. Search queries though pose a unique challenge, given their short-length and lack of nuance or context. Complicated feature engineering efforts do not always lead to downstream improvements as their performance benefits may be offset by increased complexity of knowledge distillation. Thus, in this paper we make the following contributions: (1) We demonstrate that Retrieval Augmentation of queries provides LLMs with valuable additional context enabling improved understanding. While Retrieval Augmentation typically increases latency of LMs (thus hurting distillation efficacy), (2) we provide a practical and effective way of distilling Retrieval Augmentation LLMs. Specifically, we use a novel two-stage distillation approach that allows us to carry over the gains of retrieval augmentation, without suffering the increased compute typically associated with it. (3) We demonstrate the benefits of the proposed approach (QUILL) on a billion-scale, real-world query understanding system resulting in huge gains. Via extensive experiments, including on public benchmarks, we believe this work offers a recipe for practical use of retrieval-augmented query understanding.
In this paper, we compare the performances of traditional machine learning models using feature engineering and word vectors and the state-of-the-art language model BERT using word embeddings on three datasets. We also consider the time and cost efficiency of feature engineering compared to BERT. From our results we conclude that the use of the BERT model was only worth the time and cost trade-off for one of the three datasets we used for comparison, where the BERT model significantly outperformed any kind of traditional classifier that uses feature vectors, instead of embeddings. Using the BERT model for the other datasets only achieved an increase of 0.03 and 0.05 of accuracy and F1 score respectively, which could be argued makes its use not worth the time and cost of GPU.
Recent work has shown that neural feature- and representation-learning, e.g. BERT, achieves superior performance over traditional manual feature engineering based approaches, with e.g. SVMs, in translationese classification tasks. Previous research did not show $(i)$ whether the difference is because of the features, the classifiers or both, and $(ii)$ what the neural classifiers actually learn. To address $(i)$, we carefully design experiments that swap features between BERT- and SVM-based classifiers. We show that an SVM fed with BERT representations performs at the level of the best BERT classifiers, while BERT learning and using handcrafted features performs at the level of an SVM using handcrafted features. This shows that the performance differences are due to the features. To address $(ii)$ we use integrated gradients and find that $(a)$ there is indication that information captured by hand-crafted features is only a subset of what BERT learns, and $(b)$ part of BERT's top performance results are due to BERT learning topic differences and spurious correlations with translationese.
Atmospheric processes involve both space and time. This is why human analysis of atmospheric imagery can often extract more information from animated loops of image sequences than from individual images. Automating such an analysis requires the ability to identify spatio-temporal patterns in image sequences which is a very challenging task, because of the endless possibilities of patterns in both space and time. In this paper we review different concepts and techniques that are useful to extract spatio-temporal context specifically for meteorological applications. In this survey we first motivate the need for these approaches in meteorology using two applications, solar forecasting and detecting convection from satellite imagery. Then we provide an overview of many different concepts and techniques that are helpful for the interpretation of meteorological image sequences, such as (1) feature engineering methods to strengthen the desired signal in the input, using meteorological knowledge, classic image processing, harmonic analysis and topological data analysis (2) explain how different convolution filters (2D/3D/LSTM-convolution) can be utilized strategically in convolutional neural network architectures to find patterns in both space and time (3) discuss the powerful new concept of 'attention' in neural networks and the powerful abilities it brings to the interpretation of image sequences (4) briefly survey strategies from unsupervised, self-supervised and transfer learning to reduce the need for large labeled datasets. We hope that presenting an overview of these tools - many of which are underutilized - will help accelerate progress in this area.
To ensure reliability, power transformers are monitored for partial discharge (PD) events, which are symptoms of transformer failure. Since failures can have catastrophic cascading consequences, it is critical to preempt them as early as possible. Our goal is to classify PDs as corona, floating, particle, or void, to gain an understanding of the failure location. Using phase resolved PD signal data, we create a small set of features, which can be used to classify PDs with high accuracy. This set of features consists of the total magnitude, the maximum magnitude, and the length of the longest empty band. These features represent the entire signal and not just a single phase, so the feature set has a fixed size and is easily comprehensible. With both Random Forest and SVM classification methods, we attain a 99% classification accuracy, which is significantly higher than classification using phase based feature sets such as phase magnitude. Furthermore, we develop a stacking ensemble to combine several classification models, resulting in a superior model that outperforms existing methods in both accuracy and variance.
Ransomware has appeared as one of the major global threats in recent days. The alarming increasing rate of ransomware attacks and new ransomware variants intrigue the researchers to constantly examine the distinguishing traits of ransomware and refine their detection strategies. Application Programming Interface (API) is a way for one program to collaborate with another; API calls are the medium by which they communicate. Ransomware uses this strategy to interact with the OS and makes a significantly higher number of calls in different sequences to ask for taking action. This research work utilizes the frequencies of different API calls to detect and classify ransomware families. First, a Web-Crawler is developed to automate collecting the Windows Portable Executable (PE) files of 15 different ransomware families. By extracting different frequencies of 68 API calls, we develop our dataset in the first phase of the two-phase feature engineering process. After selecting the most significant features in the second phase of the feature engineering process, we deploy six Supervised Machine Learning models: Na"ive Bayes, Logistic Regression, Random Forest, Stochastic Gradient Descent, K-Nearest Neighbor, and Support Vector Machine. Then, the performances of all the classifiers are compared to select the best model. The results reveal that Logistic Regression can efficiently classify ransomware into their corresponding families securing 99.15% overall accuracy. Finally, instead of relying on the 'Black box' characteristic of the Machine Learning models, we present the post-hoc analysis of our best-performing model using 'SHapley Additive exPlanations' or SHAP values to ascertain the transparency and trustworthiness of the model's prediction.
Two-stage robust optimization problems constitute one of the hardest optimization problem classes. One of the solution approaches to this class of problems is K-adaptability. This approach simultaneously seeks the best partitioning of the uncertainty set of scenarios into K subsets, and optimizes decisions corresponding to each of these subsets. In general case, it is solved using the K-adaptability branch-and-bound algorithm, which requires exploration of exponentially-growing solution trees. To accelerate finding high-quality solutions in such trees, we propose a machine learning-based node selection strategy. In particular, we construct a feature engineering scheme based on general two-stage robust optimization insights that allows us to train our machine learning tool on a database of resolved B&B trees, and to apply it as-is to problems of different sizes and/or types. We experimentally show that using our learned node selection strategy outperforms a vanilla, random node selection strategy when tested on problems of the same type as the training problems, also in case the K-value or the problem size differs from the training ones.
Nowadays, differential privacy (DP) has become a well-accepted standard for privacy protection, and deep neural networks (DNN) have been immensely successful in machine learning. The combination of these two techniques, i.e., deep learning with differential privacy, promises the privacy-preserving release of high-utility models trained with sensitive data such as medical records. A classic mechanism for this purpose is DP-SGD, which is a differentially private version of the stochastic gradient descent (SGD) optimizer commonly used for DNN training. Subsequent approaches have improved various aspects of the model training process, including noise decay schedule, model architecture, feature engineering, and hyperparameter tuning. However, the core mechanism for enforcing DP in the SGD optimizer remains unchanged ever since the original DP-SGD algorithm, which has increasingly become a fundamental barrier limiting the performance of DP-compliant machine learning solutions. Motivated by this, we propose DPIS, a novel mechanism for differentially private SGD training that can be used as a drop-in replacement of the core optimizer of DP-SGD, with consistent and significant accuracy gains over the latter. The main idea is to employ importance sampling (IS) in each SGD iteration for mini-batch selection, which reduces both sampling variance and the amount of random noise injected to the gradients that is required to satisfy DP. Integrating IS into the complex mathematical machinery of DP-SGD is highly non-trivial. DPIS addresses the challenge through novel mechanism designs, fine-grained privacy analysis, efficiency enhancements, and an adaptive gradient clipping optimization. Extensive experiments on four benchmark datasets, namely MNIST, FMNIST, CIFAR-10 and IMDb, demonstrate the superior effectiveness of DPIS over existing solutions for deep learning with differential privacy.