This press release includes information that constitutes forward-looking statements made pursuant to the safe harbor provision of the Private Securities Litigation Reform Act of 1995, including statements about Amdocs' growth and business results in future quarters. Although we believe the expectations reflected in such forward-looking statements are based upon reasonable assumptions, we can give no assurance that our expectations will be obtained or that any deviations will not be material. Such statements involve risks and uncertainties that may cause future results to differ from those anticipated. These risks include, but are not limited to, the effects of general economic conditions, Amdocs' ability to grow in the business markets that it serves, Amdocs' ability to successfully integrate acquired businesses, adverse effects of market competition, rapid technological shifts that may render the Company's products and services obsolete, potential loss of a major customer, our ability to develop long-term relationships with our customers, and risks associated with operating businesses in the international market. Amdocs may elect to update these forward-looking statements at some point in the future; however, the Company specifically disclaims any obligation to do so.

ST. LOUIS September 12, 2017 Amdocs (NASDAQ: DOX), a leading provider of software and services to communications and media companies, is launching Smartbot, an artificial intelligence (AI) and machine learning-based bot that enables digital service providers (DSPs) to provide customer care, sales and marketing engagements to even the most discerning millennial. Amdocs Smartbot with Microsoft Cognitive Services, specifically Microsoft Language Understanding Intelligent Service and Text Analytics API, provides leading DSPs with the ability to transform how they deliver highly personalized, self-service interactions with customers that are simple, quick and helpful. As an Amdocs commissioned global study1 that surveyed over 7,000 consumers and enterprises today revealed, bots lack the intelligence and emotional awareness to address complex consumer issues and must evolve from rules-based bots to AI smartbots. Amdocs' new Smartbot addresses this gap as it is integrated with industry domain knowledge of DSP best practices and business processes. Also, with Amdocs Smartbot's deep integration with the service provider's core information systems such as CRM, order management and product catalog, DSPs can now see a 360-degree view of the customer and understand the context of the interaction, enabling them to grow care-to-commerce revenue by making more relevant predictive care and promotional offers to customers.

Answer Set Programming (ASP) is a well-known paradigm of declarative programming with roots in logic programming and non-monotonic reasoning. Similar to other closely related problem-solving technologies, such as SAT/SMT, QBF, Planning and Scheduling, advancements in ASP solving are assessed in competition events. In this paper, we report about the design and results of the Sixth ASP Competition, which was jointly organized by the University of Calabria (Italy), Aalto University (Finland), and the University of Genoa (Italy), in affiliation with the 13th International Conference on Logic Programming and Non-Monotonic Reasoning. This edition maintained some of the design decisions introduced in 2014, e.g., the conception of sub-tracks, the scoring scheme, and the adherence to a fixed modeling language in order to push the adoption of the ASP-Core-2 standard. On the other hand, it featured also some novelties, like a benchmark selection stage classifying instances according to their empirical hardness, and a "Marathon" track where the top-performing systems are given more time for solving hard benchmarks.

We consider the problem of sequentially making decisions that are rewarded by "successes" and "failures" which can be predicted through an unknown relationship that depends on a partially controllable vector of attributes for each instance. The learner takes an active role in selecting samples from the instance pool. The goal is to maximize the probability of success in either offline (training) or online (testing) phases. Our problem is motivated by real-world applications where observations are time-consuming and/or expensive. We develop a knowledge gradient policy using an online Bayesian linear classifier to guide the experiment by maximizing the expected value of information of labeling each alternative. We provide a finite-time analysis of the estimated error and show that the maximum likelihood estimator based produced by the KG policy is consistent and asymptotically normal. We also show that the knowledge gradient policy is asymptotically optimal in an offline setting. This work further extends the knowledge gradient to the setting of contextual bandits. We report the results of a series of experiments that demonstrate its efficiency.

Random walks are at the heart of many existing deep learning algorithms for graph data. However, such algorithms have many limitations that arise from the use of random walks, e.g., the features resulting from these methods are unable to transfer to new nodes and graphs as they are tied to node identity. In this work, we introduce the notion of attributed random walks which serves as a basis for generalizing existing methods such as DeepWalk, node2vec, and many others that leverage random walks. Our proposed framework enables these methods to be more widely applicable for both transductive and inductive learning as well as for use on graphs with attributes (if available). This is achieved by learning functions that generalize to new nodes and graphs. We show that our proposed framework is effective with an average AUC improvement of 16.1% while requiring on average 853 times less space than existing methods on a variety of graphs from several domains.

In conventional chemisorption model, the d-band center theory (augmented sometimes with the upper edge of d-band for imporved accuarcy) plays a central role in predicting adsorption energies and catalytic activity as a function of d-band center of the solid surfaces, but it requires density functional calculations that can be quite costly for large scale screening purposes of materials. In this work, we propose to use the d-band width of the muffin-tin orbital theory (to account for local coordination environment) plus electronegativity (to account for adsorbate renormalization) as a simple set of alternative descriptors for chemisorption, which do not demand the ab initio calculations. This pair of descriptors are then combined with machine learning methods, namely, artificial neural network (ANN) and kernel ridge regression (KRR), to allow large scale materials screenings. We show, for a toy set of 263 alloy systems, that the CO adsorption energy can be predicted with a remarkably small mean absolute deviation error of 0.05 eV, a significantly improved result as compared to 0.13 eV obtained with descriptors including costly d-band center calculations in literature. We achieved this high accuracy by utilizing an active learning algorithm, without which the accuracy was 0.18 eV otherwise. As a practical application of this machine, we identified Cu3Y@Cu as a highly active and cost-effective electrochemical CO2 reduction catalyst to produce CO with the overpotential 0.37 V lower than Au catalyst.

The use of Artificial Neural Networks (ANNs) towards developing Artificial Intelligence (AI) has undergone a renaissance in the past decade. Out of the many emergent techniques for training ANNs that are collectively referred to as'Deep Learning', Deep Reinforcement Learning (DRL) is proving to be a particularly general and powerful method, with applications ranging from video games [1] to autonomous driving [2]. While most applications of reinforcement learning have traditionally used reinforcement signals derived from performance measures that are explicit to the task - e.g. the score in a game or grammatical errors in a translation, when considering AI systems that are required to have a significant interaction with humans - e.g. the autonomous vehicle - it is critical to consider how the human's preference for objects, events, or actions can be incorporated into the behavioral reinforcement for the AI, particularly in ways that are minimally obtrusive [3], [4]. Such behavioral adaptations occur naturally during social interactions and form the bedrock of social mechanisms that build trust and rapport between strangers [5], [6]. In this paper, we present a novel approach that uses decoded human neurophysiological and ocular time-series data as an implicit reinforcement signal for an AI agent that is driving a virtual automobile.

Optimization algorithms for training deep models not only affects the convergence rate and stability of the training process, but are also highly related to the generalization performance of the models. While adaptive algorithms, such as Adam and RMSprop, have shown better optimization performance than stochastic gradient descent (SGD) in many scenarios, they often lead to worse generalization performance than SGD, when used for training deep neural networks (DNNs). In this work, we identify two problems of Adam that may degrade the generalization performance. As a solution, we propose the normalized direction-preserving Adam (ND-Adam) algorithm, which combines the best of both worlds, i.e., the good optimization performance of Adam, and the good generalization performance of SGD. In addition, we further improve the generalization performance in classification tasks, by using batch-normalized softmax. This study suggests the need for more precise control over the training process of DNNs.

To the best of our knowledge, this paper presents the first large-scale study that tests whether network categories (e.g., social networks vs. web graphs) are distinguishable from one another (using both categories of real-world networks and synthetic graphs). A classification accuracy of $94.2\%$ was achieved using a random forest classifier with both real and synthetic networks. This work makes two important findings. First, real-world networks from various domains have distinct structural properties that allow us to predict with high accuracy the category of an arbitrary network. Second, classifying synthetic networks is trivial as our models can easily distinguish between synthetic graphs and the real-world networks they are supposed to model.

Multi-scanner Antivirus systems provide insightful information on the nature of a suspect application; however there is often a lack of consensus and consistency between different Anti-Virus engines. In this article, we analyze more than 250 thousand malware signatures generated by 61 different Anti-Virus engines after analyzing 82 thousand different Android malware applications. We identify 41 different malware classes grouped into three major categories, namely Adware, Harmful Threats and Unknown or Generic signatures. We further investigate the relationships between such 41 classes using community detection algorithms from graph theory to identify similarities between them; and we finally propose a Structure Equation Model to identify which Anti-Virus engines are more powerful at detecting each macro-category. As an application, we show how such models can help in identifying whether Unknown malware applications are more likely to be of Harmful or Adware type.