Comcast is tapping into artificial intelligence (AI) and machine learning (ML) to gain valuable insights across its networks, and to provide a better customer experience. It has also come up with an internal program to make sense of all of the buckets of data that its ML and AI systems are gathering. Previously, Comcast's Tony Werner and Matt Zelesko spoke to FierceTelecom about how company's virtualization efforts were based on the deployment of its next-gen X1 video platform. Comcast's use of AI and ML were also founded on the back of X1. With X1, Comcast was able to establish complete telemetry and visibility into the platform, establish incremental roll outs and roll back to previous versions if there's a problem.

Artificial intelligence (AI) is evolving the landscape of telecommunications and those communication service providers (CSPs) at the forefront of the change stand the most to gain. Machine learning is an exciting ability of AI, which begins with simple commands overseen by humans, like a child learning to ride a bike. When the parent is comfortable, once the algorithms are working correctly, the training wheels are taken off and AI can be involved in operations and processes. One of the major benefits of utilizing AI and machine learning capabilities is that it allows telco employees to focus on more strategic and higher priority initiatives, while mundane and tedious tasks are automated through AI. Telcos also know the value of maintaining customer satisfaction.

The edge is a rich source of new data, whether it's from subscribers connected to a telco's latest 5G network, or radio signals detected in a battlefield environment. Faced with the huge data volumes and an increasing demand for low-latency analysis and action, the processing needs to occur local to the source – at the edge itself. Furthermore, security, compliance and corruption risks may dictate whether the data is permitted to be moved away from the edge at all, and if allowed, in what curated form. A data center or cloud used in isolation cannot meet all these emerging needs. But most customers don't want to reinvent the wheel when analyzing data at the edge, and instead prefer to reuse proven technologies like AI, while adopting architectures that have already been honed inside a data center or cloud.

NVIDIA and Ericsson today announced they are collaborating on technologies that can allow telco operators to build high-performing, efficient and completely virtualized 5G radio access networks (RAN). These virtualized networks can enable faster and more flexible introduction of new AI and IoT services. The collaboration, announced by NVIDIA founder and CEO Jensen Huang at a keynote ahead of the start of MWC Los Angeles, brings together Ericsson's expertise in RAN technology with NVIDIA's leadership in GPU-powered accelerated computing platforms, as well as AI and supercomputing. Telcos are exploring alternative technologies and RAN architectures amid growing interest for virtualization, while securing the best possible user experience. A major industry challenge is how to virtualize the complete RAN solution in a cost-, size- and energy-efficient way, comparable with traditionally built RAN networks.

The automatic classification of applications and services is an invaluable feature for new generation mobile networks. Here, we propose and validate algorithms to perform this task, at runtime, from the raw physical channel of an operative mobile network, without having to decode and/or decrypt the transmitted flows. Towards this, we decode Downlink Control Information (DCI) messages carried within the LTE Physical Downlink Control CHannel (PDCCH). DCI messages are sent by the radio cell in clear text and, in this paper, are utilized to classify the applications and services executed at the connected mobile terminals. Two datasets are collected through a large measurement campaign: one labeled, used to train the classification algorithms, and one unlabeled, collected from four radio cells in the metropolitan area of Barcelona, in Spain. Among other approaches, our Convolutional Neural Network (CNN) classifier provides the highest classification accuracy of 99%. The CNN classifier is then augmented with the capability of rejecting sessions whose patterns do not conform to those learned during the training phase, and is subsequently utilized to attain a fine grained decomposition of the traffic for the four monitored radio cells, in an online and unsupervised fashion.

Over the past decade, data center services have shifted from predominantly web-centric to cloud-centric. Today, they are shifting once again, this time from the era of cloud computing and into the intelligent era. With massive amounts of data generated during digitalization, filtering out and automatically reorganizing useful information -- then mining the valuable data using Artificial Intelligence (AI) -- is arguably the key to the intelligent era. According to the Huawei Global Industry Vision (GIV), 97% of large enterprises will use AI by 2025. Indeed, more and more enterprises regard AI as the primary strategy for digital transformation.

Today, any typical modern-day smartphone is able to scan faces, documents, QR codes, capture super-resolution photos, recognise gestures, voice and perform multiple other tasks besides answering calls and texts. These handheld devices are the epitome of software and hardware engineering; and to do these tasks, they require state-of-the-art image recognition and NLP models running in the background. Image and language models are at the heart of many machine learning applications today and training these models is a computational nightmare with increasing data. Google has been using TensorFlow Lite for taking pictures on its flagship model Pixel. For Portrait mode on Pixel 3, Tensorflow Lite GPU inference accelerates the foreground-background segmentation model by over 4x and the new depth estimation model by over 10x vs CPU inference with floating-point precision.

Massive multiple-input multiple-output (MIMO) systems require downlink channel state information (CSI) at the base station (BS) to better utilize the available spatial diversity and multiplexing gains. Convolutional neural network (CNN)-based CSI feedback compression schemes has received a lot of attention recently due to significant improvements in compression efficiency; however, they still require reliable feedback links to convey the compressed CSI information to the BS. Instead, we propose here a CNN-based analog feedback scheme, called AnalogDeepCMC, which directly maps the downlink CSI to uplink channel input. Corresponding noisy channel outputs are used by another CNN to reconstruct the DL channel estimate. Not only the proposed outperforms existing digital CSI feedback schemes in terms of the achievable downlink rate, but also simplifies the operation as it does not require explicit quantization, coding and modulation, and provides a low-latency alternative particularly in rapidly changing MIMO channels, where the CSI needs to be estimated and fed back periodically.

Mobile operators are integrating AI-enabled applications and edge-cloud computing with their 5G networks to accelerate enterprise digital transformation and unleash new business opportunities. According to the Computing Technology Industry Association (CompTIA), IoT, artificial intelligence and 5G networks are among the emerging solutions offering the greatest business and financial opportunities in the digital age. These technologies have the power to transform the business landscape and how enterprises operate on a daily basis, and they could transform the productivity and growth speed of small or large corporations. These emerging technologies must process and transmit massive quantities of data in real time, something only possible with 5G and AI. The increased speed and capacity of 5G networks enhanced by artificial intelligence will enable the development and spread of technologies such as autonomous vehicles, smart cities, and virtual reality.

Financial firms are working to identify potential fraud by analyzing how customers hold their phones, how fast they type and other information about mobile interactions--and the strategy is yielding results. Using artificial-intelligence tools to crunch behavioral data is often a more secure way to verify customers than traditional means such as passcodes, experts say. The Federal Bureau of Investigation warned companies last month that cybercriminals can circumvent typical multifactor-authentication techniques. One way is by calling a telecommunications company, posing as a customer and getting a service agent to switch that person's phone number to the criminal's device. The fraudster can then have the individual's bank send a one-time passcode to the phone and gain access to the target's bank account.