The Comcast brand is synonymous with home entertainment and telecommunications. While Comcast is well known for its products, it has evolved into a technology company focused on leading new development initiatives to enhance its products, solutions, and customer experiences. Technology is used as a business enabler across the organization. For Comcast's leadership team, improving and optimizing customer service has become one of the focal points of its technology development strategy. "As technology trends develop and new products become available, we're able to broaden the range of what we can do for our customers, which drives new ideas at the business level," says Jason Michener, director of software development and engineering at Comcast.

Korean telecommunications company KT announced plans to invest 300 billion won ($257 million) over the next four years to become an artificial intelligence (AI) company, Korean press reported. The Korean telco also said it aims to hire nearly 1,000 specialists in the AI field with the aim of creating new value propositions in line with the deployment of 5G networks in the country. KT rolled out its AI-based service, called Giga Genie, in January of 2017. This AI service was initially offered in the form of a television set-top box. The company has been recently expanding the application of the AI-based service to speakers, apartments, hotels and cars.

At its NXTWORK conference in Las Vegas, Juniper Networks announced on Monday new capabilities for its Mist and Contrail solutions that were designed to help its enterprise customers. For Mist, Juniper announced what it claimed was the first "AI Driven Self Diving Network" for enterprises that use Mist's AI engine. It also uses Mist's microservices cloud to streamline IT operations to help simplify troubleshooting across both wired and wireless networks. Using Mist's integrated AI-engine, which is called Marvis, the Mist platform identifies the root cause of issues across various IT domains, such as WLAN, LAN, WAN and security, and automatically resolves them when possible. If the issue is outside the domain of the access network, Marvis will provide a set of recommended actions to help IT managers resolve their issues.

T-Mobile USA CEO John Legere has surfaced as the possible new CEO at WeWork, where he would follow the troubled tenure of co-founder Adam Neumann. The Wall Street Journal reported that Legere is in discussions with the office-sharing startup, which was bailed out recently by SoftBank Group. The topic of succession at T-Mobile came up during a sit-down Legere had with USA TODAY's Ed Baig last week, fresh off the announcements that T-Mobile would be flipping the switch on its 5G network on Dec. 6, Legere was joined in the conversation by T-Mobile president and COO Mike Sievert to discuss the remaining obstacles to T-Mobile's pending merger with Sprint and to make the case that the merger will result in more, not less competition, and more jobs. SoftBank already holds a major stake in Sprint. The Journal article stated that there's no guarantee Legere would take on the WeWork challenge.

One of the things about newly hyped technologies is that they are often described in magical terms and Artificial Intelligence (AI) is no different. Reading the technology press and vendor announcements make it seem as if AI can cure all your business' problems and allow you to leapfrog the competition, succeed in new markets in days and delight your customers while reducing costs by 80%. What is particularly fascinating about AI is that the hype is not new, in fact it is decades old. The first Neural Network, a foundation of today's AI research, was actually created in the late 1950's. This led to a large leap in expectations, which were never realized.

Conventional multiuser detection techniques either require a large number of antennas at the receiver for a desired performance, or they are too complex for practical implementation. Moreover, many of these techniques, such as successive interference cancellation (SIC), suffer from errors in parameter estimation (user channels, covariance matrix, noise variance, etc.) that is performed before detection of user data symbols. As an alternative to conventional methods, this paper proposes and demonstrates a low-complexity practical Machine Learning (ML) based receiver that achieves similar (and at times better) performance to the SIC receiver. The proposed receiver does not require parameter estimation; instead it uses supervised learning to detect the user modulation symbols directly. We perform comparisons with minimum mean square error (MMSE) and SIC receivers in terms of symbol error rate (SER) and complexity.

In this paper, we consider a platform of flying mobile edge computing (F-MEC), where unmanned aerial vehicles (UA Vs) serve as equipment providing computation resource, and they enable task offload-ing from user equipment (UE). We aim to minimize energy consumption of all the UEs via optimizing the user association, resource allocation and the trajectory of UA Vs. To this end, we first propose a Convex optimizAtion based Trajectory control algorithm (CA T), which solves the problem in an iterative way by using block coordinate descent (BCD) method. Then, to make the real-time decision while taking into account the dynamics of the environment (i.e., UA V may take off from different locations), we propose a deep Reinforcement leArning based Trajectory control algorithm (RA T). In RA T, we apply the Prioritized Experience Replay (PER) to improve the convergence of the training procedure. Different from the convex optimization based algorithm which may be susceptible to the initial points and requires iterations, RA T can be adapted to any taking off points of the UA Vs and can obtain the solution more rapidly than CA T once training process has been completed. Simulation results show that the proposed CA T and RA T achieve the similar performance and both outperform traditional algorithms. Liang, Kezhi and Nauman are with the Department of Computer and Informantion Science, Northumbria University, Newcastle upon Tyne, UK, NE1 8ST. Cunhua and Arumugam are with School of Electronic Engineering and Computer Science, Queen Mary University of London, E1 4NS, U.K. Wei is with National Mobile Communications Research Lab, Southeast University, China. I NTRODUCTION With the popularity of computationally-intensive tasks, e.g., smart navigation and augmented reality, people are expecting to enjoy more convenient life than ever before. However, current smart devices and user equipments (UEs), due to small size and limited resource, e.g., computation and battery, may not be able to provide satisfactory Quality of Service (QoS) and Quality of Experience (QoE) in executing those highly demanding tasks. Mobile edge computing (MEC) has been proposed by moving the computation resource to the network edge and it has been proved to greatly enhance UE's ability in executing computation-hungry tasks [1].

A primary challenge in single-cell RNA sequencing (scRNA-seq) studies comes from the massive amount of data and the excess noise level. To address this challenge, we introduce a hierarchical autoencoder that reliably extracts representative information of each cell. In an extensive analysis, we demonstrate that the approach vastly outperforms state-of-the-art techniques in many research sub-fields of scRNA-seq analysis, including cell segregation through unsupervised learning, visualization of transcriptome landscape, cell classification, and pseudo-time inference.

At Microsoft we are working on many Artificial Intelligence (AI) conversational platforms, transforming the way companies engage their customers. When building them, one of the most important success factors is providing these AI with the right personality: right for your company… and right for your customers. Please remember, in case your AI conversational platform had a strong "personality", never call him or her "bot". Otherwise, you may have quite a "personal" response…:-) It's important to know the difference between conversational AI and conventional chatbots (or just "bots"). Chatbots serve up canned responses to anticipated requests and statements.

Classical stochastic models, such as the Erd os-R enyi, Barabasi-Albert, and the stochastic block model generate graphs based on a predefined set of parameters, such as the probability of edge formation within and between communities [1]. In contrast, modern approaches to graph generation based on deep learning, including NetGAN [2], GraphGAN [3], and GraphRNN [4], are flexible enough to learn multiple different properties of an input graph simultaneously. The graphs generated by these architectures may be used for downstream learning tasks such as data augmentation [5], recommendation [6], and link prediction [7]. Many real-world networks consist of entities with complex mutual interrelations. Such networks cannot be modeled effectively as graphs with simple pairwise relations, despite the fact that pairwise relations provide a wealth of information for learning. Studying higher-order relationships in a graph is fundamental for our understanding of the network behavior and function. Higher-order relationships are usually termed hyperedges (collections of more than two nodes) [8, 9] or network motifs (recurrent node connectivity patterns that are statistically significant compared to some ground truth random graph model) [10]. These higher-order structures are the actual building blocks of complex networks, as they capture fundamental functional properties.