Recent work (Ilyas et al, 2019) suggests that adversarial examples are features not bugs. If adversarial perturbations are indeed useful but non-robust features, then what is their origin? In order to answer these questions, we systematically examine the learning dynamics of adversarial perturbations both in the pixel and frequency domains. We find that: (1) adversarial examples are not present at initialization but instead emerge very early in training, typically within the first epochs, as verified by a novel breakpoint-based analysis; (2) the low-amplitude high-frequency nature of common adversarial perturbations in natural images is critically dependent on an implicit bias towards sparsity in the frequency domain; and (3) the origin of this bias is the locality and translation invariance of convolutional filters, along with (4) the existence of useful frequency-domain features in natural images. We provide a simple theoretical explanation for these observations, providing a clear and minimalist target for theorists in future work. Looking forward, our findings suggest that analyzing the learning dynamics of perturbations can provide useful insights for understanding the origin of adversarial sensitivities and developing robust solutions.

We propose a new embedding method, named Quantile-Quantile Embedding (QQE), for distribution transformation, manifold embedding, and image embedding with the ability to choose the embedding distribution. QQE, which uses the concept of quantile-quantile plot from visual statistical tests, can transform the distribution of data to any theoretical desired distribution or empirical reference sample. Moreover, QQE gives the user a choice of embedding distribution in embedding manifold of data into the low dimensional embedding space. It can also be used for modifying the embedding distribution of different dimensionality reduction methods, either basic or deep ones, for better representation or visualization of data. We propose QQE in both unsupervised and supervised manners. QQE can also transform the distribution to either the exact reference distribution or shape of the reference distribution; and one of its many applications is better discrimination of classes. Our experiments on different synthetic and image datasets show the effectiveness of the proposed embedding method.

Matrix square roots and their inverses arise frequently in machine learning, e.g., when sampling from high-dimensional Gaussians $\mathcal{N}(\mathbf 0, \mathbf K)$ or whitening a vector $\mathbf b$ against covariance matrix $\mathbf K$. While existing methods typically require $O(N^3)$ computation, we introduce a highly-efficient quadratic-time algorithm for computing $\mathbf K^{1/2} \mathbf b$, $\mathbf K^{-1/2} \mathbf b$, and their derivatives through matrix-vector multiplication (MVMs). Our method combines Krylov subspace methods with a rational approximation and typically achieves $4$ decimal places of accuracy with fewer than $100$ MVMs. Moreover, the backward pass requires little additional computation. We demonstrate our method's applicability on matrices as large as $50,\!000 \times 50,\!000$ - well beyond traditional methods - with little approximation error. Applying this increased scalability to variational Gaussian processes, Bayesian optimization, and Gibbs sampling results in more powerful models with higher accuracy.

Explainable AI provides insight into the "why" for model predictions, offering potential for users to better understand and trust a model, and to recognize and correct AI predictions that are incorrect. Prior research on human and explainable AI interactions has focused on measures such as interpretability, trust, and usability of the explanation. Whether explainable AI can improve actual human decision-making and the ability to identify the problems with the underlying model are open questions. Using real datasets, we compare and evaluate objective human decision accuracy without AI (control), with an AI prediction (no explanation), and AI prediction with explanation. We find providing any kind of AI prediction tends to improve user decision accuracy, but no conclusive evidence that explainable AI has a meaningful impact. Moreover, we observed the strongest predictor for human decision accuracy was AI accuracy and that users were somewhat able to detect when the AI was correct versus incorrect, but this was not significantly affected by including an explanation. Our results indicate that, at least in some situations, the "why" information provided in explainable AI may not enhance user decision-making, and further research may be needed to understand how to integrate explainable AI into real systems.

This work proposes a structured approach to baselining a model, identifying attack vectors, and securing the machine learning models after deployment. This method for securing each model post deployment is called the BAD (Build, Attack, and Defend) Architecture. Two implementations of the BAD architecture are evaluated to quantify the adversarial life cycle for a black box Sentiment Analysis system. As a challenging diagnostic, the Jigsaw Toxic Bias dataset is selected as the baseline in our performance tool. Each implementation of the architecture will build a baseline performance report, attack a common weakness, and defend the incoming attack. As an important note: each attack surface demonstrated in this work is detectable and preventable. The goal is to demonstrate a viable methodology for securing a machine learning model in a production setting.

An adversarial attack paradigm explores various scenarios for the vulnerability of deep learning models: minor changes of the input can force a model failure. Most of the state of the art frameworks focus on adversarial attacks for images and other structured model inputs, but not for categorical sequences models. Successful attacks on classifiers of categorical sequences are challenging because the model input is tokens from finite sets, so a classifier score is non-differentiable with respect to inputs, and gradient-based attacks are not applicable. Common approaches deal with this problem working at a token level, while the discrete optimization problem at hand requires a lot of resources to solve. We instead use a fine-tuning of a language model for adversarial attacks as a generator of adversarial examples. To optimize the model, we define a differentiable loss function that depends on a surrogate classifier score and on a deep learning model that evaluates approximate edit distance. So, we control both the adversability of a generated sequence and its similarity to the initial sequence. As a result, we obtain semantically better samples. Moreover, they are resistant to adversarial training and adversarial detectors. Our model works for diverse datasets on bank transactions, electronic health records, and NLP datasets.

Recent work has highlighted the vulnerability of many deep machine learning models to adversarial examples. It attracts increasing attention to adversarial attacks, which can be used to evaluate the security and robustness of models before they are deployed. However, to our best knowledge, there is no specific research on the adversarial attacks for multi-view deep models. This paper proposes two multi-view attack strategies, two-stage attack (TSA) and end-to-end attack (ETEA). With the mild assumption that the single-view model on which the target multi-view model is based is known, we first propose the TSA strategy. The main idea of TSA is to attack the multi-view model with adversarial examples generated by attacking the associated single-view model, by which state-of-the-art single-view attack methods are directly extended to the multi-view scenario. Then we further propose the ETEA strategy when the multi-view model is provided publicly. The ETEA is applied to accomplish direct attacks on the target multi-view model, where we develop three effective multi-view attack methods. Finally, based on the fact that adversarial examples generalize well among different models, this paper takes the adversarial attack on the multi-view convolutional neural network as an example to validate that the effectiveness of the proposed multi-view attacks. Extensive experimental results demonstrate that our multi-view attack strategies are capable of attacking the multi-view deep models, and we additionally find that multi-view models are more robust than single-view models.

Being able to model and forecast international migration as precisely as possible is crucial for policymaking. Recently Google Trends data in addition to other economic and demographic data have been shown to improve the forecasting quality of a gravity linear model for the one-year ahead forecasting. In this work, we replace the linear model with a long short-term memory (LSTM) approach and compare it with two existing approaches: the linear gravity model and an artificial neural network (ANN) model. Our LSTM approach combined with Google Trends data outperforms both these models on various metrics in the task of forecasting the one-year ahead incoming international migration to 35 Organization for Economic Co-operation and Development (OECD) countries: for example the root mean square error (RMSE) and the mean average error (MAE) have been divided by 5 and 4 on the test set. This positive result demonstrates that machine learning techniques constitute a serious alternative over traditional approaches for studying migration mechanisms.

The oil and gas industry is turning to AI to help cut operating costs, predict equipment failure, and increase oil and gas output. A faulty well pump at an unmanned platform in the North Sea disrupted production in early 2019 for Aker BP, a Norwegian oil company, according to an account in the Wall Street Journal. The company installed an AI program that monitors data from sensors on the pump, flagging glitches before they can cause a shutdown, stated Lars Atle Andersen, VP of operations for the firm. Now he flies in engineers to fix such problems ahead of time and prevent a shutdown, he stated. Aker BP employed a solution from SparkCognition of Austin, Texas.

Job Number: R0083391 Artificial Intelligence/Machine Learning Data Scientist The Challenge Are you excited at the prospect of unlocking the secrets held by a data set? Are you fascinated by the possibilities presented by the IoT or recent advances in machine learning and artificial intelligence? In an increasingly connected world, massive amounts of structured and unstructured data open up new opportunities. As a data scientist, you can turn these complex data sets into useful information to solve global challenges. Across private and public sectors -- from fraud detection to cancer research to national intelligence -- you know the answers are in the data.