The widespread use of ML-based decision making in domains with high societal impact such as recidivism, job hiring and loan credit has raised a lot of concerns regarding potential discrimination. In particular, in certain cases it has been observed that ML algorithms can provide different decisions based on sensitive attributes such as gender or race and therefore can lead to discrimination. Although, several fairness-aware ML approaches have been proposed, their focus has been largely on preserving the overall classification accuracy while improving fairness in predictions for both protected and non-protected groups (defined based on the sensitive attribute(s)). The overall accuracy however is not a good indicator of performance in case of class imbalance, as it is biased towards the majority class. As we will see in our experiments, many of the fairness-related datasets suffer from class imbalance and therefore, tackling fairness requires also tackling the imbalance problem. To this end, we propose AdaFair, a fairness-aware classifier based on AdaBoost that further updates the weights of the instances in each boosting round taking into account a cumulative notion of fairness based upon all current ensemble members, while explicitly tackling class-imbalance by optimizing the number of ensemble members for balanced classification error. Our experiments show that our approach can achieve parity in true positive and true negative rates for both protected and non-protected groups, while it significantly outperforms existing fairness-aware methods up to 25% in terms of balanced error.

The devastating effects of cyber-attacks, highlight the need for novel attack detection and prevention techniques. Over the last years, considerable work has been done in the areas of attack detection as well as in collaborative defense. However, an analysis of the state of the art suggests that many challenges exist in prioritizing alert data and in studying the relation between a recently discovered attack and the probability of it occurring again. In this article, we propose a system that is intended for characterizing network entities and the likelihood that they will behave maliciously in the future. Our system, namely Network Entity Reputation Database System (NERDS), takes into account all the available information regarding a network entity (e. g. IP address) to calculate the probability that it will act maliciously. The latter part is achieved via the utilization of machine learning. Our experimental results show that it is indeed possible to precisely estimate the probability of future attacks from each entity using information about its previous malicious behavior and other characteristics. Ranking the entities by this probability has practical applications in alert prioritization, assembly of highly effective blacklists of a limited length and other use cases.

Most real-world networks are incompletely observed. Algorithms that can accurately predict which links are missing can dramatically speedup the collection of network data and improve the validity of network models. Many algorithms now exist for predicting missing links, given a partially observed network, but it has remained unknown whether a single best predictor exists, how link predictability varies across methods and networks from different domains, and how close to optimality current methods are. We answer these questions by systematically evaluating 203 individual link predictor algorithms, representing three popular families of methods, applied to a large corpus of 548 structurally diverse networks from six scientific domains. We first show that individual algorithms exhibit a broad diversity of prediction errors, such that no one predictor or family is best, or worst, across all realistic inputs. We then exploit this diversity via meta-learning to construct a series of "stacked" models that combine predictors into a single algorithm. Applied to a broad range of synthetic networks, for which we may analytically calculate optimal performance, these stacked models achieve optimal or nearly optimal levels of accuracy. Applied to real-world networks, stacked models are also superior, but their accuracy varies strongly by domain, suggesting that link prediction may be fundamentally easier in social networks than in biological or technological networks. These results indicate that the state-of-the-art for link prediction comes from combining individual algorithms, which achieves nearly optimal predictions. We close with a brief discussion of limitations and opportunities for further improvement of these results.

We address the problem of learning to benchmark the best achievable classifier performance. In this problem the objective is to establish statistically consistent estimates of the Bayes misclassification error rate without having to learn a Bayes-optimal classifier. Our learning to benchmark framework improves on previous work on learning bounds on Bayes misclassification rate since it learns the {\it exact} Bayes error rate instead of a bound on error rate. We propose a benchmark learner based on an ensemble of $\epsilon$-ball estimators and Chebyshev approximation. Under a smoothness assumption on the class densities we show that our estimator achieves an optimal (parametric) mean squared error (MSE) rate of $O(N^{-1})$, where $N$ is the number of samples. Experiments on both simulated and real datasets establish that our proposed benchmark learning algorithm produces estimates of the Bayes error that are more accurate than previous approaches for learning bounds on Bayes error probability.

Emergence of artificial intelligence techniques in biomedical applications urges the researchers to pay more attention on the uncertainty quantification (UQ) in machine-assisted medical decision making. For classification tasks, prior studies on UQ are difficult to compare with each other, due to the lack of a unified quantitative evaluation metric. Considering that well-performing UQ models ought to know when the classification models act incorrectly, we design a new evaluation metric, area under Confidence-Classification Characteristic curves (AUCCC), to quantitatively evaluate the performance of the UQ models. AUCCC is threshold-free, robust to perturbation, and insensitive to the classification performance. We evaluate several UQ methods (e.g., max softmax output) with AUCCC to validate its effectiveness. Furthermore, a simple scheme, named Uncertainty Distillation (UDist), is developed to boost the UQ performance, where a confidence model is distilling the confidence estimated by deep ensembles. The proposed method is easy to implement; it consistently outperforms strong baselines on natural and medical image datasets in our experiments.

In many real-world scenarios where data is high dimensional, test time acquisition of features is a non-trivial task due to costs associated with feature acquisition and evaluating feature value. The need for highly confident models with an extremely frugal acquisition of features can be addressed by allowing a feature selection method to become target aware. We introduce an approach to feature selection that is based on Bayesian learning, allowing us to report target-specific levels of uncertainty, false positive, and false negative rates. In addition, measuring uncertainty lifts the restriction on feature selection being target agnostic, allowing for feature acquisition based on a single target of focus out of many. We show that acquiring features for a specific target is at least as good as common linear feature selection approaches for small non-sparse datasets, and surpasses these when faced with real-world healthcare data that is larger in scale and in sparseness.

Disaster monitoring is challenging due to the lake of infrastructures in monitoring areas. Based on the theory of Game-With-A-Purpose (GWAP), this paper contributes to a novel large-scale crowdsourcing disaster monitoring system. The system analyzes tagged satellite pictures from anonymous players, and then reports aggregated and evaluated monitoring results to its stakeholders. An algorithm based on directed graph centralities is presented to address the core issues of malicious user detection and disaster level calculation. Our method can be easily applied in other human computation systems. In the end, some issues with possible solutions are discussed for our future work.

Fake engagement is one of the significant problems in Online Social Networks (OSNs) which is used to increase the popularity of an account in an inorganic manner. The detection of fake engagement is crucial because it leads to loss of money for businesses, wrong audience targeting in advertising, wrong product predictions systems, and unhealthy social network environment. This study is related with the detection of fake and automated accounts which leads to fake engagement on Instagram. As far as we know, there is no publicly available dataset for fake and automated accounts. For this purpose, two datasets have been generated for the detection of fake and automated accounts. For the detection of these accounts, machine learning algorithms like Naive Bayes, logistic regression, support vector machines and neural networks are applied. Additionally, for the detection of automated accounts, cost sensitive genetic algorithm is applied because of the unnatural bias in the dataset. To deal with the unevenness problem in the fake dataset, Smote-nc algorithm is implemented. For the automated and fake account detection problem, 86% and 96% are obtained, respectively.

Over three decades, the Internet has grown from a small network of computers used by research scientists to communicate and exchange data to a technology that has penetrated almost every aspect of our day-to-day lives. Today, it is hard to imagine a life without online access for business, shopping, and socialising. A technology that has connected humanity at a scale never before possible has also amplified some of our worst qualities. Online hate speech spreads virally across the globe with short- and long-term consequences for individuals and societies. These consequences are often difficult to measure and predict. Online social media websites and mobile apps have inadvertently become the platform for the spread and proliferation of hate speech.

Adversarial examples are slight perturbations that are designed to fool artificial neural networks when fed as an input. In this work the usability of the Fisher information for the detection of such adversarial attacks is studied. We discuss various quantities whose computation scales well with the network size, study their behavior on adversarial examples and show how they can highlight the importance of single input neurons, thereby providing a visual tool for further analyzing (un-)reasonable behavior of a neural network. The potential of our methods is demonstrated by applications to the MNIST, CIFAR10 and Fruits-360 datasets.