The use of artificial intelligence (AI) and machine learning (ML) has infiltrated many industries and cable is no exception. We are seeing cable network operators increasingly embrace the power of AI and ML to improve the network.

In the Spring of 2020, SCTE launched a working group to create standards for AI and ML for the cable industry. The work of the AI/ML working group represents just one part of the SCTE Explorer Initiative which launched in 2020 as part of SCTE’s Standards program, the only ANSI-accredited platform for developing technical standards supporting cable broadband networks. The Explorer Initiative is made up of seven working groups – including aging in place and telehealth, and telemedicine, smart cities, and more – that represent industries, technologies, or practices that will place significant demands on telecommunications infrastructure.

Tasked with exploring how AI and ML can be leveraged to make the network more efficient, this group consists of more than 50 participants from both inside and outside of cable who have been working diligently to shape industry best practices around data governance, anomaly detection, digital piracy, and automated video and ad analysis. Over the past year, the group has made significant progress toward the creation of standards in several areas that will help the industry advance toward 10G and will support scalability of new technology deployments across the network, including the use of ML for spectral impairment detection, HFC node splits, and video metadata extraction.

Machine Learning and Preventative Network Maintenance

The working group has made significant progress over the past year on creating a standard for how cable operators can apply machine learning for preventative network maintenance (PNM). The initial work has been around the use of ML for spectral impairment detection across the access network to create automated diagnostic reporting and mitigation.

Spectral impairments from a variety of sources contribute to a large percentage of infrastructure issues. We are seeing that machine learning models can be trained with numerous sets of labeled spectral impairment observations, including 15 different types of impairment (i.e. tilt, roll-off, notch, foreign carrier, etc.), to return a complete impairment diagnostic. These models, which learn over time to become more accurate, improve the rapid identification and mitigation of impairments to provide fewer network disruptions. Once complete, a standard for this application will assist cable operators big and small, and improve network service for everyone.

Machine Learning for Network Management

Another area the working group has focused on over the past year is the application of ML for network management, specifically HFC node splits. Commonly used by network operators to provide greater bandwidth and capacity to a specific geographic area, node splits require considerable labor and capital investment. This typically leads operators to prioritize where to invest finite resources, a process that has historically relied on a time intensive manual evaluation effort. Machine learning models are proving effective at evaluating the variables to rapidly provide actionable node split reports. The development of an industry standard for this application promises to speed up network growth, improve network efficiency, and provide customers with reliable high-speed services as the network continues to grow.

Artificial Intelligence and Video Piracy

Another area in which the working group has been busy over the past year is around developing best practices for video metadata extraction.

A standard around the application of artificial intelligence on the network to detect signatures of bad actors will help control video piracy. The work of the group over the past year indicates that the use of AI can significantly reduce video piracy. The successful creation and application of a standard around this use case has the potential to provide remarkable benefits to content creators, streaming servicing, and others. The work of the group around video piracy illustrates the power of having technology experts from both inside and outside of the industry working together to create standards.

Over the next couple of years, we expect to see new methods arise for leveraging artificial intelligence and machine learning to optimize

network operations. And, as the industry advances towards 10G, emerging technologies will continue to place demands on the industry to not only keep pace, but to remain ahead of the rapid acceleration of technology. As this happens, it is critical that these advancements are met with standards that will help to support a new era of connectivity and enable businesses to optimize their products and services to reach customers across the broadband network. Join us to be a part of the standards working groups and help shape the future of connectivity.

The post Making Progress Towards AI/ML Standards for Cable appeared first on Artificial Intelligence.

The advent of artificial intelligence is proving to be of monumental impact in a multitude of fields. The innovative applications have elevated convenience levels from good to better and then to best. The customized services especially, which have proved to be the scoring points of a variety of successful companies owes their growth and profit to artificial intelligence and their personalized algorithms. However, as the saying goes, a coin has two sides, so does any technology or action.

And when it comes t0 AI powered technologies, the crucial question is about privacy. The personalized ads that you see on your device do make you filled with awe and wonder. After all its your choices that are on display even before you have thought about buying them. The possibility of a telepathy is intriguing. However, as we look deep, the concerns and questions rush in and the terrifying logic behind them begins to sow seeds of doubt and fear. We keep on questioning the possibility of AI tools invading the private space.

To be clear, it is an obvious fact that companies have been relying on user data ever since internet came to dominate the tech world. Even after persistent assurance about privacy protection, the tech giants like Amazon, Facebook, Microsoft, Google etc. cannot deny this fact. To prove this point, here is a statement made by Microsoft in 2019,

“We’ve long been clear that we collect voice data to improve voice-enabled services and that this data is sometimes reviewed by vendors.”

This implies that listeners of our personal data includes not just AI, but humans as well. The company also added in the statement that they are working on the privacy policy to add more clarity and that they always ensure that they attain consumer permission before utilizing voice data.

The question is whether we should be willing to have a device that evades every bit of our personal space for the sake of adding personalized ads. The key to this question rests on the fact that the rush one feels after getting their hands on a brand new gadget (even if it violates your privacy), overrides the fear. This minute brain glitch is the working formula of the various devices that depend on personal data.

AI, over the years have become synonymous with change. Some latest additions to this list are Google’s Nest Hub and Amazon’s Echo Show. Both include the similar feature of AI sleep tracking. Impressive on the surface, the devices raise massive privacy questions because of the fact that they include sensors and microphones and cameras to enable microscopic movement detection. Yes, a good night’s sleep is of utmost importance. But the irony strikes with the nights you lose sleep due to the inevitable privacy questions and conflicts that are a part of these devices.

To conclude, AI by definition, is not a threat. It is a harbinger of change that comes with a cost. The good and bad that artificial intelligence is capable of should help customers to draw a line of balance, thereby ensuring that not much is lost in the quest of gaining new things.

The post Artificial Intelligence : Progress or Privacy nightmare? appeared first on Artificial Intelligence.