In this paper, we propose a deep learning approach for smartphone user identification based on analyzing motion signals recorded by the accelerometer and the gyroscope, during a single tap gesture performed by the user on the screen. We transform the discrete 3-axis signals from the motion sensors into a gray-scale image representation which is provided as input to a convolutional neural network (CNN) that is pre-trained for multi-class user classification. In the pre-training stage, we benefit from different users and multiple samples per user. After pre-training, we use our CNN as feature extractor, generating an embedding associated to each single tap on the screen. The resulting embeddings are used to train a Support Vector Machines (SVM) model in a few-shot user identification setting, i.e. requiring only 20 taps on the screen during the registration phase. We compare our identification system based on CNN features with two baseline systems, one that employs handcrafted features and another that employs recurrent neural network (RNN) features. All systems are based on the same classifier, namely SVM. To pre-train the CNN and the RNN models for multi-class user classification, we use a different set of users than the set used for few-shot user identification, ensuring a realistic scenario. The empirical results demonstrate that our CNN model yields a top accuracy of 89.75% in multi-class user classification and a top accuracy of 96.72% in few-shot user identification. In conclusion, we believe that our system is ready for practical use, having a better generalization capacity than both baselines.

Drug - drug interactions ( DDIs) account for over 30% of all adverse drug reactions ( ADRs) ca ses and often occur when co - medicate more than two drugs. More alarmingly, it stays a significant ADR - mediated morbidity every year [1], and this ramps up withdrawn - risks of a drug from the market and thu s pulls a strong disincentive to drug development [2] . Though it is ideal for detecting all negative DDIs during clinical trials, DDIs - induced - ADRs cases are often reported at clinical uses and post - marketing surveillance, which pose a severe threat to public health. A study concerning the relationship between DDIs and the mortality rate of elderly hospitalized patients concludes that over 62.77% of patients present at least one DDI, and this may amount strictl y to the death of these patients [3] . Besides, DDIs also expand the length of stay and cost of hospitalization [4] .

--In recent years, deep learning methods have outperformed other methods in image recognition. This has fostered imagination of potential application of deep learning technology including safety relevant applications like the interpretation of medical images or autonomous driving. The passage from assistance of a human decision maker to ever more automated systems however increases the need to properly handle the failure modes of deep learning modules. In this contribution, we review a set of techniques for the self-monitoring of machine-learning algorithms based on uncertainty quantification. In particular, we apply this to the task of semantic segmentation, where the machine learning algorithm decomposes an image according to semantic categories. We discuss false positive and false negative error modes at instance-level and review techniques for the detection of such errors that have been recently proposed by the authors. We also give an outlook on future research directions. The stunning success of deep learning technology, convolu-tional neural networks (CNN) in particular [1]-[3], has led to a rush towards technology development for new applications that ten years ago would have been considered unrealistic.

Artificial intelligence(AI) systems and humans communicate more and more with each other. AI systems are optimized for objectives such as error rate in communication or effort, eg. computation. In contrast, inputs created by humans are often treated as a given. We investigate how humans providing information to an AI can adjust to reduce miscommunication and improve efficiency while having to change their behavior as little as possible. These objectives result in trade-offs that we investigate using handwritten digits. To create examples that serve as demonstrations for humans to improve, we develop a model based on a conditional convolutional autoencoder (CCAE). Our quantitative and qualitative evaluation shows that in many occasions the generated proposals lead to lower error rates, require less effort to create and differ only modestly from the original samples.

Recently, it was found that the maximization of MIAS can be used as an intrinsic motivation for artificial agents. In literature, the term empowerment is used to represent the maximum of MIAS at a certain state. While empowerment has been shown to solve a broad range of reinforcement learning problems, its calculation in arbitrary dynamics is a challenging problem because it relies on the estimation of mutual information. Existing approaches, which rely on sampling, are limited to low dimensional spaces, because high-confidence distribution-free lower bounds for mutual information require exponential number of samples. In this work, we develop a novel approach for the estimation of empowerment in unknown dynamics from visual observation only, without the need to sample for MIAS. The core idea is to represent the relation between action sequences and future states using a stochastic dynamic model in latent space with a specific form. This allows us to efficiently compute empowerment with the "Water-Filling" algorithm from information theory. We construct this embedding with deep neural networks trained on a sophisticated objective function. Our experimental results show that the designed embedding preserves information-theoretic properties of the original dynamics.

Most of the data-driven approaches applied to bearing fault diagnosis up to date are established in the supervised learning paradigm, which usually requires a large set of labeled data collected a priori. In practical applications, however, obtaining accurate labels based on real-time bearing conditions can be far more challenging than simply collecting a huge amount of unlabeled data using various sensors. In this paper, we thus propose a semi-supervised learning approach for bearing anomaly detection using variational autoencoder (VAE) based deep generative models, which allows for effective utilization of dataset when only a small subset of data have labels. Finally, a series of experiments is performed using both the Case Western Reserve University (CWRU) bearing dataset and the University of Cincinnati's Center for Intelligent Maintenance Systems (IMS) dataset. The experimental results demonstrate that the proposed semi-supervised learning scheme greatly outperforms two mainstream semi-supervised learning approaches and a baseline supervised convolutional neural network approach, with the overall accuracy improvement ranging between 3% to 30% using different proportions of labeled samples.

We use Asynchronous Advantage Actor Critic (A3C) for implementing an AI agent in the controllers that optimize flow of traffic across a single intersection and then extend it to multiple intersections by considering a multi-agent setting. We explore three different methodologies to address the multi-agent problem - (1) use of asynchronous property of A3C to control multiple intersections using a single agent (2) utilise self/competitive play among independent agents across multiple intersections and (3) ingest a global reward function among agents to introduce cooperative behavior between intersections. We observe that (1) & (2) leads to a reduction in traffic congestion. Additionally the use of (3) with (1) & (2) led to a further reduction in congestion.

Once users have shared their data online, it is generally difficult for them to revoke access and ask for the data to be deleted. Machine learning (ML) exacerbates this problem because any model trained with said data may have memorized it, putting users at risk of a successful privacy attack exposing their information. Yet, having models unlearn is notoriously difficult. After a data point is removed from a training set, one often resorts to entirely retraining downstream models from scratch. We introduce SISA training, a framework that decreases the number of model parameters affected by an unlearning request and caches intermediate outputs of the training algorithm to limit the number of model updates that need to be computed to have these parameters unlearn. This framework reduces the computational overhead associated with unlearning, even in the worst-case setting where unlearning requests are made uniformly across the training set. In some cases, we may have a prior on the distribution of unlearning requests that will be issued by users. We may take this prior into account to partition and order data accordingly and further decrease overhead from unlearning. Our evaluation spans two datasets from different application domains, with corresponding motivations for unlearning. Under no distributional assumptions, we observe that SISA training improves unlearning for the Purchase dataset by 3.13x, and 1.658x for the SVHN dataset, over retraining from scratch. We also validate how knowledge of the unlearning distribution provides further improvements in retraining time by simulating a scenario where we model unlearning requests that come from users of a commercial product that is available in countries with varying sensitivity to privacy. Our work contributes to practical data governance in machine learning.

University of California, Davis Abstract--Despite the plethora of studies about security vulnerabilities and defenses of deep learning models, security aspects of deep learning methodologies, such as transfer learning, have been rarely studied. In this article, we highlight the security challenges and research opportunities of these methodologies, focusing on vulnerabilities and attacks unique to them. W ith the widespread adaptation of deep neural networks (DNN), their security challenges have received significant attention from both academia and industry, especially for mission critical applications, such as road sign detection for autonomous vehicles, face recognition in authentication systems, and fraud detection in financial systems. There are three major types of attacks on deep learning models, namely adversarial attacks, data poisoning, and exploratory attacks. Particularly, adversarial attacks, which aim to carefully craft inputs that cause the model to misclassify, has been extensively studied and many defence mechanisms have been proposed to alleviate them. These attacks are of paramount importance because they are effective, moderately simple to launch, and often transferable from one model to another. In literature, there are several survey and review papers on deep learning security and defence mechanisms. In this article, we focus on security of a much less explored area of machine learning - machine learning methodologies. Machine learning methodologies have been widely used to mitigate the restrictions and assumptions of a typical machine learning process. A typical DNN training process assumes large labeled dataset(s), access to high computational resources, non-private and centralized data, standard training and hyper-parameter tuning, and fixed task distribution over time. However, these assumptions are often difficult to realize in practice. Notwithstanding the proliferation of these machine learning methodologies, their security aspects have not been comprehensively analyzed, if ever studied. In this article, we focus on potential attacks, security vulnerabilities, and future directions specific to each learning methodology.

With the rise of machine learning and data science, applied everywhere and changing every industry, it's no wonder that experts in machine learning are handsomely paid and much looked after. If you've already read a couple of data science and machine learning books, it's time to focus on deep learning: Neural Networks, Keras, Tensorflow, Scikit-learn, etc. Introduction to Machine Learning with Python is a smooth introduction into machine learning and deep learning. It doesn't assume any knowledge about coding and Python in particular and it introduces fundamental concepts and applications of machine learning, discussing various methods through examples. That's the best book I've ever seen for an entry level Deep Learning Engineer. If you've already completed a couple of machine learning projects, you know something about Keras or Tensorflow, you've used scikit-learn then I have two recommendations for you.