Both in the domains of Feature Selection and Interpretable AI, there exists a desire to `rank' features based on their importance. Such feature importance rankings can then be used to either: (1) reduce the dataset size or (2) interpret the Machine Learning model. In the literature, however, such Feature Rankers are not evaluated in a systematic, consistent way. Many papers have a different way of arguing which feature importance ranker works best. This paper fills this gap, by proposing a new evaluation methodology. By making use of synthetic datasets, feature importance scores can be known beforehand, allowing more systematic evaluation. To facilitate large-scale experimentation using the new methodology, a benchmarking framework was built in Python, called fseval. The framework allows running experiments in parallel and distributed over machines on HPC systems. By integrating with an online platform called Weights and Biases, charts can be interactively explored on a live dashboard. The software was released as open-source software, and is published as a package on the PyPi platform. The research concludes by exploring one such large-scale experiment, to find the strengths and weaknesses of the participating algorithms, on many fronts.

How to sample high quality negative instances from unlabeled data, i.e., negative sampling, is important for training implicit collaborative filtering and contrastive learning models. Although previous studies have proposed some approaches to sample informative instances, few has been done to discriminating false negative from true negative for unbiased negative sampling. On the basis of our order relation analysis of negatives' scores, we first derive the class conditional density of true negatives and that of false negatives. We next design a Bayesian classifier for negative classification, from which we define a model-agnostic posterior probability estimate of an instance being true negative as a quantitative negative signal measure. We also propose a Bayesian optimal sampling rule to sample high-quality negatives. The proposed Bayesian Negative Sampling (BNS) algorithm has a linear time complexity. Experimental studies validate the superiority of BNS over the peers in terms of better sampling quality and better recommendation performance.

This article was published as a part of the Data Science Blogathon. Now, we enter the secret sauce: CM_Norm adjusts the colour-bar, such that its point of origin is equal to the accuracy expected for a random prediction. Essentially, the "naive-prediction accuracy" is our "point of origin" because a model which predicts worse than a coin-flip, is not a helpful model to begin with (hence the name: "coin-flip confusion-matrix"). In other words, we are interested in a models "excess performance", rather than its "absolute" error rates. To give two examples: For 3 different classes, the "point of origin", of the colour-bar, would be set at 1/3, or for 10 classes it would be set at 1/10.

Unlike linear models and SVM (see Part 1), some machine learning models are really complex to learn from their mathematical formulation. Fortunately, they can be understood by following a step-by-step process they execute on a small dummy dataset. This way, you can uncover machine learning models under the hood without the "math bottleneck". You will learn three more models in this story after Part 1: K-Nearest Neighbors (KNN), Naive Bayes, and Decision Tree. KNN is a non-generalizing machine learning model since it simply "remembers" all of its train data.

Research on intelligent tutoring systems has been exploring data-driven methods to deliver effective adaptive assistance. While much work has been done to provide adaptive assistance when students seek help, they may not seek help optimally. This had led to the growing interest in proactive adaptive assistance, where the tutor provides unsolicited assistance upon predictions of struggle or unproductivity. Determining when and whether to provide personalized support is a well-known challenge called the assistance dilemma. Addressing this dilemma is particularly challenging in open-ended domains, where there can be several ways to solve problems. Researchers have explored methods to determine when to proactively help students, but few of these methods have taken prior hint usage into account. In this paper, we present a novel data-driven approach to incorporate students' hint usage in predicting their need for help. We explore its impact in an intelligent tutor that deals with the open-ended and well-structured domain of logic proofs. We present a controlled study to investigate the impact of an adaptive hint policy based on predictions of HelpNeed that incorporate students' hint usage. We show empirical evidence to support that such a policy can save students a significant amount of time in training, and lead to improved posttest results, when compared to a control without proactive interventions. We also show that incorporating students' hint usage significantly improves the adaptive hint policy's efficacy in predicting students' HelpNeed, thereby reducing training unproductivity, reducing possible help avoidance, and increasing possible help appropriateness (a higher chance of receiving help when it was likely to be needed). We conclude with suggestions on the domains that can benefit from this approach as well as the requirements for adoption.

Networks and temporal point processes serve as fundamental building blocks for modeling complex dynamic relational data in various domains. We propose the latent space Hawkes (LSH) model, a novel generative model for continuous-time networks of relational events, using a latent space representation for nodes. We model relational events between nodes using mutually exciting Hawkes processes with baseline intensities dependent upon the distances between the nodes in the latent space and sender and receiver specific effects. We demonstrate that our proposed LSH model can replicate many features observed in real temporal networks including reciprocity and transitivity, while also achieving superior prediction accuracy and providing more interpretable fits than existing models.

As the number of smart devices that surround us increases, so do the opportunities to create smart socially-aware systems. In this context, mobile devices can be used to collect data about students and to better understand how their day-to-day routines can influence their academic performance. Moreover, the Covid-19 pandemic led to new challenges and difficulties, also for students, with considerable impact on their lifestyle. In this paper we present a dataset collected using a smartphone application (ISABELA), which include passive data (e.g., activity and location) as well as self-reported data from questionnaires. We present several tests with different machine learning models, in order to classify students' performance. These tests were carried out using different time windows, showing that weekly time windows lead to better prediction and classification results than monthly time windows. Furthermore, it is shown that the created models can predict student performance even with data collected from different contexts, namely before and during the Covid-19 pandemic. SVMs, XGBoost and AdaBoost-SAMME with Random Forest were found to be the best algorithms, showing an accuracy greater than 78%. Additionally, we propose a pipeline that uses a decision level median voting algorithm to further improve the models' performance, by using historic data from the students to further improve the prediction. Using this pipeline, it is possible to further increase the performance of the models, with some of them obtaining an accuracy greater than 90%.

In this paper we share findings from our effort to build practical machine translation (MT) systems capable of translating across over one thousand languages. We describe results in three research domains: (i) Building clean, web-mined datasets for 1500+ languages by leveraging semi-supervised pre-training for language identification and developing data-driven filtering techniques; (ii) Developing practical MT models for under-served languages by leveraging massively multilingual models trained with supervised parallel data for over 100 high-resource languages and monolingual datasets for an additional 1000+ languages; and (iii) Studying the limitations of evaluation metrics for these languages and conducting qualitative analysis of the outputs from our MT models, highlighting several frequent error modes of these types of models. We hope that our work provides useful insights to practitioners working towards building MT systems for currently understudied languages, and highlights research directions that can complement the weaknesses of massively multilingual models in data-sparse settings.

Machine prediction algorithms (e.g., binary classifiers) often are adopted on the basis of claimed performance using classic metrics such as sensitivity and predictive value. However, classifier performance depends heavily upon the context (workflow) in which the classifier operates. Classic metrics do not reflect the realized utility of a predictor unless certain implicit assumptions are met, and these assumptions cannot be met in many common clinical scenarios. This often results in suboptimal implementations and in disappointment when expected outcomes are not achieved. One common failure mode for classic metrics arises when multiple predictions can be made for the same event, particularly when redundant true positive predictions produce little additional value. This describes many clinical alerting systems. We explain why classic metrics cannot correctly represent predictor performance in such contexts, and introduce an improved performance assessment technique using utility functions to score predictions based on their utility in a specific workflow context. The resulting utility metrics (u-metrics) explicitly account for the effects of temporal relationships on prediction utility. Compared to traditional measures, u-metrics more accurately reflect the real world costs and benefits of a predictor operating in a live clinical context. The improvement can be significant. We also describe a formal approach to snoozing, a mitigation strategy in which some predictions are suppressed to improve predictor performance by reducing false positives while retaining event capture. Snoozing is especially useful for predictors that generate interruptive alarms. U-metrics correctly measure and predict the performance benefits of snoozing, whereas traditional metrics do not.

As engineered systems grow in complexity, there is an increasing need for automatic methods that can detect, diagnose, and even correct transient anomalies that inevitably arise and can be difficult or impossible to diagnose and fix manually. Among the most sensitive and complex systems of our civilization are the detectors that search for incredibly small variations in distance caused by gravitational waves -- phenomena originally predicted by Albert Einstein to emerge and propagate through the universe as the result of collisions between black holes and other massive objects in deep space. The extreme complexity and precision of such detectors causes them to be subject to transient noise issues that can significantly limit their sensitivity and effectiveness. In this work, we present a demonstration of a method that can detect and characterize emergent transient anomalies of such massively complex systems. We illustrate the performance, precision, and adaptability of the automated solution via one of the prevalent issues limiting gravitational-wave discoveries: noise artifacts of terrestrial origin that contaminate gravitational wave observatories' highly sensitive measurements and can obscure or even mimic the faint astrophysical signals for which they are listening. Specifically, we demonstrate how a highly interpretable convolutional classifier can automatically learn to detect transient anomalies from auxiliary detector data without needing to observe the anomalies themselves. We also illustrate several other useful features of the model, including how it performs automatic variable selection to reduce tens of thousands of auxiliary data channels to only a few relevant ones; how it identifies behavioral signatures predictive of anomalies in those channels; and how it can be used to investigate individual anomalies and the channels associated with them.