'Phrenology' has an old-fashioned ring to it. It sounds like it belongs in a history book, filed somewhere between bloodletting and velocipedes. We'd like to think that judging people's worth based on the size and shape of their skull is a practice that's well behind us. However, phrenology is once again rearing its lumpy head. In recent years, machine-learning algorithms have promised governments and private companies the power to glean all sorts of information from people's appearance.

Datasets are rarely a realistic approximation of the target population. Say, prevalence is misrepresented, image quality is above clinical standards, etc. This mismatch is known as sampling bias. Sampling biases are a major hindrance for machine learning models. They cause significant gaps between model performance in the lab and in the real world. Our work is a solution to prevalence bias. Prevalence bias is the discrepancy between the prevalence of a pathology and its sampling rate in the training dataset, introduced upon collecting data or due to the practioner rebalancing the training batches. This paper lays the theoretical and computational framework for training models, and for prediction, in the presence of prevalence bias. Concretely a bias-corrected loss function, as well as bias-corrected predictive rules, are derived under the principles of Bayesian risk minimization. The loss exhibits a direct connection to the information gain. It offers a principled alternative to heuristic training losses and complements test-time procedures based on selecting an operating point from summary curves. It integrates seamlessly in the current paradigm of (deep) learning using stochastic backpropagation and naturally with Bayesian models.

When solving machine learning problems, simply training a model based on a problem-specific training machine learning algorithm does not guarantee either that the resulting model fully captures the underlying concept hidden in the training data or that the optimum parameter values were chosen for model training. Failing to test a model's performance means an underperforming model could be deployed on the production system, resulting in incorrect predictions. Choosing one model from the many available options based on intuition alone is risky. By generating different metrics, the efficacy of the model can be assessed. Use of these metrics reveals how well the model fits the data on which it was trained.

Dealing with imbalanced data is a prevalent problem while performing classification on the datasets. Many times, this problem contributes to bias while making decisions or implementing policies. Thus, it is vital to understand the factors which causes imbalance in the data (or class imbalance). Such hidden biases and imbalances can lead to data tyranny, and a major challenge to a data democracy. In this chapter, two essential statistical elements are resolved: the degree of class imbalance and the complexity of the concept, solving such issues helps in building the foundations of a data democracy. Further, statistical measures which are appropriate in these scenarios are discussed and implemented on a real-life dataset (car insurance claims). In the end, popular data-level methods such as Random Oversampling, Random Undersampling, SMOTE, Tomek Link, and others are implemented in Python, and their performance is compared. Keywords - Imbalanced Data, Degree of Class Imbalance, Complexity of the Concept, Statistical Assessment Metrics, Undersampling and Oversampling 1. Motivation & Introduction In the real-world, data are collected from various sources like social networks, websites, logs, and databases. Whilst dealing with data from different sources, it is very crucial to check the quality of the data [1]. Data with questionable quality can introduce different types of biases in various stages of the data science lifecycle. These biases sometime can affect the association between variables, and in many cases could represent the opposite of the actual behavior [2].

Cybersecurity is of the utmost concern for financial institutions (FIs) of all types, ranging from community credit unions to multibillion-dollar international banking conglomerates to everyday consumers. More than 2 million fraud reports were issued to the Federal Trade Commission in 2020, reaching a total loss of more than $3 billion. One survey found that 47 percent of businesses around the world have reported being victimized by digital crime within the past two years, with losses totaling $42 billion. Fraudsters are also growing more advanced in their tactics, leveraging sophisticated technologies like artificial intelligence (AI) and machine learning (ML) to deploy millions of attacks simultaneously. The overwhelming volume of attacks has put organizations on the back foot, scrambling to find countermeasures to the account takeovers (ATOs), phishing attacks and other schemes they face by the thousands every day.

Machine learning (ML) prediction APIs are increasingly widely used. An ML API can change over time due to model updates or retraining. This presents a key challenge in the usage of the API because it is often not clear to the user if and how the ML model has changed. Model shifts can affect downstream application performance and also create oversight issues (e.g. if consistency is desired). In this paper, we initiate a systematic investigation of ML API shifts. We first quantify the performance shifts from 2020 to 2021 of popular ML APIs from Google, Microsoft, Amazon, and others on a variety of datasets. We identified significant model shifts in 12 out of 36 cases we investigated. Interestingly, we found several datasets where the API's predictions became significantly worse over time. This motivated us to formulate the API shift assessment problem at a more fine-grained level as estimating how the API model's confusion matrix changes over time when the data distribution is constant. Monitoring confusion matrix shifts using standard random sampling can require a large number of samples, which is expensive as each API call costs a fee. We propose a principled adaptive sampling algorithm, MASA, to efficiently estimate confusion matrix shifts. MASA can accurately estimate the confusion matrix shifts in commercial ML APIs using up to 90% fewer samples compared to random sampling. This work establishes ML API shifts as an important problem to study and provides a cost-effective approach to monitor such shifts.

We have built a novel system for the surveillance of drinking water reservoirs using underwater sensor networks. We implement an innovative AI-based approach to detect, classify and localize underwater events. In this paper, we describe the technology and cognitive AI architecture of the system based on one of the sensor networks, the hydrophone network. We discuss the challenges of installing and using the hydrophone network in a water reservoir where traffic, visitors, and variable water conditions create a complex, varying environment. Our AI solution uses an autoencoder for unsupervised learning of latent encodings for classification and anomaly detection, and time delay estimates for sound localization. Finally, we present the results of experiments carried out in a laboratory pool and the water reservoir and discuss the system's potential.

We consider a popular family of constrained optimization problems arising in machine learning that involve optimizing a non-decomposable evaluation metric with a certain thresholded form, while constraining another metric of interest. Examples of such problems include optimizing the false negative rate at a fixed false positive rate, optimizing precision at a fixed recall, optimizing the area under the precision-recall or ROC curves, etc. Our key idea is to formulate a rate-constrained optimization that expresses the threshold parameter as a function of the model parameters via the Implicit Function theorem. We show how the resulting optimization problem can be solved using standard gradient based methods. Experiments on benchmark datasets demonstrate the effectiveness of our proposed method over existing state-of-the art approaches for these problems. The code for the proposed method is available at https://github.com/google-research/google-research/tree/master/implicit_constrained_optimization .

Countless research works of deep neural networks (DNNs) in the task of credit card fraud detection have focused on improving the accuracy of point predictions and mitigating unwanted biases by building different network architectures or learning models. Quantifying uncertainty accompanied by point estimation is essential because it mitigates model unfairness and permits practitioners to develop trustworthy systems which abstain from suboptimal decisions due to low confidence. Explicitly, assessing uncertainties associated with DNNs predictions is critical in real-world card fraud detection settings for characteristic reasons, including (a) fraudsters constantly change their strategies, and accordingly, DNNs encounter observations that are not generated by the same process as the training distribution, (b) owing to the time-consuming process, very few transactions are timely checked by professional experts to update DNNs. Therefore, this study proposes three uncertainty quantification (UQ) techniques named Monte Carlo dropout, ensemble, and ensemble Monte Carlo dropout for card fraud detection applied on transaction data. Moreover, to evaluate the predictive uncertainty estimates, UQ confusion matrix and several performance metrics are utilized. Through experimental results, we show that the ensemble is more effective in capturing uncertainty corresponding to generated predictions. Additionally, we demonstrate that the proposed UQ methods provide extra insight to the point predictions, leading to elevate the fraud prevention process.

Tables in scientific papers contain a wealth of valuable knowledge for the scientific enterprise. To help the many of us who frequently consult this type of knowledge, we present Tab2Know, a new end-to-end system to build a Knowledge Base (KB) from tables in scientific papers. Tab2Know addresses the challenge of automatically interpreting the tables in papers and of disambiguating the entities that they contain. To solve these problems, we propose a pipeline that employs both statistical-based classifiers and logic-based reasoning. First, our pipeline applies weakly supervised classifiers to recognize the type of tables and columns, with the help of a data labeling system and an ontology specifically designed for our purpose. Then, logic-based reasoning is used to link equivalent entities (via sameAs links) in different tables. An empirical evaluation of our approach using a corpus of papers in the Computer Science domain has returned satisfactory performance. This suggests that ours is a promising step to create a large-scale KB of scientific knowledge.