The world of digital advertising is no stranger to debates and controversies. Consider YouTube's recent brush after an Adalytics report claimed adult ads are being served in kids’ content. Amidst the chatter, Google launched a counterstudy to shed more light on its technology. 

On the surface, all of this might seem like just another swipe at the tech giants. Given their size, many assume that companies like Google must be cutting corners, denying them the benefit of the doubt. And meanwhile, Google staffers are scratching their heads, asking, “Why are we suddenly the villains?”

But it also brings up a crucial, central question: Can — and more importantly, should — advertising on platforms like YouTube ever be fully automated? If not, where is the proper boundary between machine influence and human oversight?

The immense potential of machine learning is undeniable. We can now sift through vast data sets, identify patterns and target ads with incredible precision and speed. 

But precision doesn’t always translate to brand relevance or safety. Just because a machine can target an adult watching a children’s video doesn't mean it should. Context, after all, is king.

With controversies like the one we’ve seen with YouTube, we start to see some machine-learning pitfalls more clearly. Machines lack nuance. They can’t discern context the way humans can. For instance, while a program can recognize that a video is geared toward children, it’s pretty close to impossible for it to discern whether an adult is watching alongside them, leading to misplaced ads and wasted dollars. 

That said, here are some ways machine learning and process automation can add real value to YouTube advertising:

• Customizing Inclusion Lists: Brands want to ensure their content is running alongside the channels they approve. But with 51 million channels on YouTube, growing at year-over-year rates of over 35%, that's a heavy lift for a human. Enhanced toolsets here could refine advertiser control.
• Optimizing Campaigns: Machine learning can tailor optimization recommendations, enabling brands to hone their campaigns more effectively, so that they resonate better with target audiences.
• Pacing Budgets: It’s crucial to keep your ad spend on track. With the aid of automation, campaigns can better align their spending, ensuring they’re not overspending or underspending the allocated budget.

The value of automation becomes murkier when we move beyond these processes. While automation can manage a bid or predict a user’s next click, it can’t yet grasp the subtleties of human experience or the broader cultural context. We’ve seen machines falter when it comes to recognizing potentially harmful or inappropriate content. 

Remember when ads were routinely placed next to extremist content? Or the gun ads that kept getting past rules that supposedly prohibited them? Those were machine oversights. Targeting ads to kids? Another tech hiccup.

So where does this leave us? As in most areas, the future of YouTube advertising likely lies in a marriage between human and machine. That blend of human strategic expertise, honed through daily interaction with the platform, can provide insights and guidance that transcend basic metrics. Automation can handle the heavy lifting, but you need people to catch the nuances machines might miss. 

In the current landscape, relying too much on YouTube’s tools — or any single platform’s automation — isn’t just risky, it’s potentially reckless. The appetite for automation is growing, but so are the stakes. As the industry moves forward, brands and advertisers will have no choice but to recognize both strengths and limitations of automation — and adapt accordingly.

Pity the poor sales reps in a media and entertainment company. They’re tasked with bringing in more and more revenue for their organizations, yet they compete in an extremely crowded and noisy environment. Worse, the entire sector pitches the same generic data — number of visits or views, gender, time of data, etc. — beefed up with commoditized syndicated segments, such as sitcom fans and news junkies. How can sales-team pitches stand out when they tell the same story as everyone else? It’s simple. They can’t.

What M&E sales teams need to do in order to break through the noise is to tell better stories about audiences. Storytelling wins pitches, as we’ve seen over and over. Most M&E companies try to do just that, which is why they have dedicated ad-research teams tasked with looking for insights in their mounds of data and slicing their audiences into segments as best they can on behalf of advertisers. But I’ve yet to see any research team that isn’t overwhelmed by requests, a situation that’s not likely to change given the advertiser’s appetite for finding the “right” audiences. With these teams overloaded and in reactive mode, sales teams end up missing out on sale opportunities and losing deals because they often can’t get the insights and compelling story they need to close them.

AI-Assisted Selling

This is precisely where AI — specifically unsupervised machine learning — can help. Take a company like CNN, which sees consumers across multiple touch points: website, mobile app, OTT, TV on-demand, etc., all of which generate massive amounts of first-party data. AI can be used to proactively sift through data sets quickly and efficiently, and identify meaningful insights based on connections between users and similarities in their behaviors. Graph technology can automatically identify insights and aggregate these signals so that M&E research and sales teams can find stories in their data at a glance. This means sales teams have more, and better, stories to share, and can become more like a consultant to the advertiser, which is far more strategic and valuable.

AI lets sales teams unlock all of the interesting, nonintuitive stories that are buried inside the company’s massive consumer data. Humans are multidimensional creatures; a viewer may be a mom who is an avid fan of a cooking show, but she may also be a horse fanatic who lives to ride. Another viewer may be a sports fan, but when he’s not watching the game he’s in his basement workshop tinkering away at a DIY project. Put another way, this M&E brand doesn’t just have two obvious audiences, cooking and sports enthusiasts, it has many nuanced audiences all within the existing user base. By unlocking new insights at scale, M&E sales teams can tell new and differentiated stories to advertisers.

Going further, M&E sales teams can leverage that data proactively to create net new opportunities instead of simply responding to existing briefs. For instance, they can create a custom audience and pitch it to a brand that has never advertised with any of their properties before. Or they can cater segments to a brand’s specific campaign goals, such as growing brand awareness among specific audiences.

More interestingly, M&E companies can advise advertisers of the additional affinities shared by an audience, which the brand can then use to drive message development. Let’s say an M&E company discovers that viewers of a science-fiction show also have a strong affinity for steampunk and animal rights. This insight helps the advertiser align campaign and content strategy with the natural interests of the audience.

Win-Win-Win Scenario

By telling better stories with their data, M&E companies break through the noise, close more deals, become more strategic and drive higher-value inventory to the advertisers that want to align with these unique audience segments. Brands can cater their messages to touch on the additional shared affinities within their target audiences, which will make their ads inherently more interesting and relevant. That, in turn, means the audience members themselves enjoy an overall better experience. That’s a win for everyone.

Tim Burke is CEO of marketing strategy platform Affinio.

Comcast said it has acquired Arlington, Virginia-based BlueVector for an undisclosed sum.

The technology company, which was spun off from defense contractor Northrop Grumman two years ago, uses artificial intelligence and machine learning to provide cybersecurity services to companies and government agencies.

BluVector has been owned since 2017 by Philadelphia-based private equity firm LLR Partners.

Comcast said that Eric Malawar, a veteran cybersecurity engineer who once worked for the House Committee on Homeland Security, is now the CEO of BluVector. Malawer’s LinkedIn profile says he joined Comcast in 2017 as head of AI and data strategy.

BluVector’s founding CEO, former IBM executive Kris Lovejoy, will serve as an adviser and consultant to Comcast and its new venture.

Comcast chief information security officer, Noopur Davis, will lead the process of identifying opportunities to leverage the companies’ combined technology and expertise to support new products and initiatives.

Using proprietary technology, BluVector works to detect, analyze and contain a wide range of sophisticated cyber threats, including fileless malware, zero-day malware and ransomware

BluVector doesn't announce the names of its clients, but it said in November that it had one a "multi-million-dollar contract" with a "government agency."

“BluVector is a global leader in leveraging AI and machine learning to defend against advanced cyber threats,” Don Mathis, GM, Growth at Comcast said, in statement. “We’re thrilled that BluVector is part of Comcast and are excited to support its continued growth, even as we explore new opportunities to leverage BluVector technology and expertise. 

In the "you can't spell Pai without AI department," FCC chair Ajit Pai has scheduled a Nov.30 forum on artificial intelligence (AI) and machine learning technologies.

The event, which will be held at the FCC's Washington headquarters and be open to the public, will include hands-on demos of the technology as well as experts to discuss its implications for the communications marketplace and policy.

"Because so much of AI intersects with the Commission’s technological and engineering work," said Pai, "we want to explore what it means for the future of communications. I look forward to bringing in experts to discuss this important issue, so that the FCC and the American public can learn about what’s on the horizon.”

The chairman has already seen, and recognized, the technologies' potential for helping communications accessibility.

Among the FCC's Chairman's Awards for Advancements in Accessibility in 2018 was one for IBM's Content Clarifier, which uses uses "artificial intelligence (AI) algorithms, machine learning models, and natural language processing to simplify, summarize, and augment digital content to increase comprehension for people with cognitive disabilities, the aging population, or those learning English as a second language."

Taking a shot at video-capable smart home hubs from rivals such as Amazon, Google announced that a mix of CE partners are coming out with Smart Display products that work with Google Assistant and are integrated with YouTube and YouTube TV.

Those new products, from partners such as LG Electronics, JBL, and Lenovo, are slated to come out in July, Lilian Rincon, director of product management at Google, announced Tuesday at the Google I/O developer conference.

Those displays will support voice commands used for Google Home with the video components. A YouTube TV customer, for example, will be able to use Google’s new Smart Display to tune into a show and have it stream to the screen, or request a video from YouTube to be sent to the device.

Rincon also demonstrated how a Smart Display could also be used to show recipe information along with a video of step-by-step instructions on how the item is prepared.

Much of Google I/O was used to tout advancements in artificial intelligence, and how Google is applying the technology across multiple products.

Google, for example, has redesigned Gmail with A.I. to include a “smart compose” feature that uses machine learning to suggest phrases as the user writes. Users simply hit the Tab button to use the suggested auto-completing of sentences and phrases. That’s rolling out this month.

Google Assistant is using a technology called Wavenet that creates a more natural sounding voice, with respect to pitch, pace and pauses that can convey meaning, using the raw audio. Google Assistant is also adding six new voices to Google Assistant, including one based on the voice of singer John Legend.

Scott Huffman, VP of Google Assistant engineering, said more than 500 million devices, including about 5,000 connected home devices, are now equipped with Google Assistant.

Among new features coming out in the weeks to come is “Continued Conversation,” which allows some back-and-forth and follow-up requests from the user without having to say, “Hey, Google,” each time. The system is also being made to support multiple actions/requests.

Google also talked up Android P, the new version of its mobile OS. Among the new features is an adaptive battery technology that improves battery life by employing machine learning to understand which apps the user is likely to open in given times of the day. There’s also an auto-brightness feature that accounts not just for current lighting conditions but also the user’s personal preferences.

Google is launching the Android P beta today for the Pixel and seven other “flagship” Android devices. 

Mist, a startup that brings A.I. techniques to wireless networks, has raised a $46 million “C” round led by Kleiner Perkins.

The latest round also included participation from Lightspeed Venture Partners, Norwest Venture Partners, GV (formerly Google Ventures), NTT Docomo Ventures, and Dimension Data.

Mist, which has raised $88 million so far, said it will use the new funds to grow sales and expand its marketing and invest in engineering, with a focus on A.I.

It also aims to develop a channel of managed service providers and work with original equipment manufactures.

Founded in 2014 and based in  Cupertino, Calif., Mist ties data to machine learning and A.I. techniques to underpin “self-learning” wireless networks that optimize the experience. It started to ship its Mist Learning WLAN in 2016, and said it ended 2017 with more than 250 customers.

Among those customers is Verizon, which is using Mist’s technology for its software-defined wireless local area network.

Comcast’s voice remote for its X1 platform has become an increasingly popular feature, and the operator has kept it stocked with a constant flow of updates that, for example, support voice commands for special events such as the Super Bowl or the Winter Olympics.

The technology that underpins that platform is also branching well beyond the TV. Comcast has already started to integrate the X1 voice remote with Xfinity Home, its home security and automation service.

Related: Comcast Adds AI Smarts to Home Security Cameras

Comcast also recently added another voice wrinkle with a Phone Finder feature for Xfinity Mobile. Those customers can activate it by saying, “Xfinity Mobile, find my phone” into the X1 voice remote, or by uttering their 10-digit phone number into the remote.

That’s expected to be just the tip of the voice-enabled spear. Comcast, which has deployed nearly 20 million voice remotes, envisions it will support voice navigation and other voice-based features across a greater number of “domains” that criss-cross the company’s full lineup of services. But a major challenge comes with creating a system that can do that, at scale, without requiring an army of people to manage and update the system manually.

Progress: It’s All In the Telling
Not only does the system need to know and understand an increasingly broader scope of specific and conversational-style commands, it also needs to grasp the intent of the voice command. Is it a search for a TV program or movie, or is the user telling the smart home system to turn off a light or adjust the thermostat?

Comcast is taking on the scale challenge with a machine learning platform developed in-house that works in tandem with an integrated metadata platform. “We have a really good advantage to work with voice, because of our metadata platform,” Comcast Cable executive director of AI product Jeanine Heck said.

That same metadata system has helped to create a foundation for the voice engine.

Early on, Comcast used a more traditional, pattern-based algorithm that relied on manual tuning. But it later realized that machine learning would be required to train that algorithm to maintain a high level of accuracy while keeping pace with requirements as the scope and complexity of the system continued to expand.

Related: AI, Machine Learning to Change the Customer Experience, Comcast’s Watson Says

Machine learning tied to a language model would also become necessary as the voice platform reaches into more domains while adapting to a broader range of conversation-style commands. “We saw that the machine could learn how to accurately find intent in a better way than it did prior to that,” Heck said, adding that Comcast’s dependence on deep learning to process natural language has only become more pronounced over time. “Our machines can learn how to adapt to those [new domains].”

Added Jonathan Palmatier, Comcast Cable’s vice president of product management, voice control: “You’d need an army of people that are trying to capture every possible way that you can construct a phrase, and [that’s where] it starts to get untenable.”

In addition to expanding voice support to multiple Xfinity services, Comcast has also been working on how voice commands can support customer support functions.

Comcast’s work around A.I. has extended to the cameras that are attached to its Xfinity Home platform to help pinpoint when those cameras should capture movement.

In the wake of a 24/7 recording feature with motion-detection introduced last year, Comcast  recently added A.I. and machine learning techniques that train the system to “identify the most relevant image in a motion event,” Chris Ganster, senior director, product at Comcast Cable, explained in this blog post.

RELATED: Comcast Brings 24/7 Video Recording to Xfinity Home

Those additions extend and enhance the movement detection capabilities of the earlier iteration of the system.

“While the system was smart enough to recognize movement, it was less good at capturing exactly when that movement was taking place (this resulted in a lot of a pictures of the tip of a dog’s tail or the edge of a car bumper in the far corner of an image),” he noted.

Comcast, he added, took some of the AI and machine learning techniques used to refine the X1 Voice Remote and applied them to technology called “computer vision.”

From there, engineers from Comcast’s Applied AI Team in Washington, D.C., worked with those involved with Xfinity Home to implement a “homegrown algorithm” that focuses on movement, centers it, and “delivers a thumbnail that is instantly recognizable.”

“So now instead of squinting to see the tip of your dog’s tail, you’ll see a clear image of your dog, or the delivery truck, or whatever else triggers your motion detector,” Ganster noted.

As a result, Comcast said it has already seen a 20% rise in customers using the Xfinity Home app to view recorded moments.

Ganster said Comcast intends to add more AI-focused features to its Xfinity Home cameras that, for example, will use optical zoom to deliver better, more useful images.

Comcast has more than 1 million Xfinity Home subs.

Denver -- The cable industry has no lack of network data. The challenge, according to panelists who spoke at a Cable-Tec Expo workshop on Wednesday (Oct. 18) in the Customer Journey track, is to use that information about the home and plant in the right way.

The common goal is better customer experience. “My job is to capture as much data around the customer as possible,” Comcast vice president of personalization Martin Marcinczyk said, “and to leverage that data to provide context for the workforce, products and customer service.”

To illustrate his point, Marcinczyk shared a video of Comcast going beyond normal operations in support of customers recently upended by disastrous weather. But he also acknowledged the deluge of information that Comcast itself faces: more than 250 pieces of data around the delivery platform multiplied millions of times. “Our customers interact with us a billion times a day,” he said. “And we don’t have one system, we have hundreds.”

Read More: Complete coverage of Cable-Tec Expo 2017

Comcast appears to be making headway in leveraging these massive amounts of data. “We have so much information,” said Marcinczyk, “we can begin to predict when we’ll have successful events -- or not so successful events.”

The vision of proactive maintenance is compelling, but progress is not universal. As a reality check, Gary Cunha, senior director, product management at ARRIS, shared input received earlier this year from one tier 2 operator: “We’re still reactive; when a spike in calls hits the call center, we really have no idea what caused it.”

Step one is a change in mindset. Cunha recommended thinking of the call center not as the first, but the last point of contact, like the goalkeeper on a soccer team. ‘Think of operations as a soccer team,” he said. “There are about seven opportunities to prevent the goal.”

Other winning combinations include optimized technical workforces, proactive home and network management and subscriber self-care. Cunha shared results from two cases. One involved analyzing the root cause of open work orders that were generating a flood of calls. The solution led to a reduction in repeat truck rolls and fault-based service calls, and a $1.6 million savings in one quarter. In the second case, assisting a large European operator burdened with a surge in WiFi-related service calls, ARRIS set up a program for telemetry analysis and evaluation of channel utilization, access points and client received signal strength indication (RSSI).

Customer self-care can be a big win, too. Cunha said the top two reasons for WiFi-related service calls, accounting for 30 percent in this category, are simply requests for SSIDs and passwords. He called that situation “a prime opportunity to exploit customer readiness for technology.”

The industry’s R&D arm has also been analyzing access network data. CableLabs Principal Architect Karthik Sundaresan updated workshop attendees on the organization’s efforts “to learn from that data and take concrete actions or steps to make an experience better.”

Like Cunha, Sundaresan said the right mental framework was essential. “Before I create a machine to do something, first I have to visualize” he said. “Can I understand it?”

Sharing graphical representations of DOCSIS 3.1 modulation error ration (MER) data from Comcast, Sundaresan discussed individual cable modem performances, some less stable than others. About one, he noted: “It’s nice for about two weeks, then something happens and MER drops by about 3dB.”

Plotted by dB level, frequencies and time, these three-dimensional graphs - looking in one case like the Himalayas and in another like the Grand Canyon - revealed much about individual modems. But that kind of analysis only goes so far. “You can see where this is leading,” he said. “Machine learning is the end goal.”

To that end, Sundaresan talked about discovering common characteristics across pools of modems and using algorithms such as a sliding median and threshold comparison for initial anomaly detection. Training a predictive model also involves data labels, such as (in the case of modems) wide, sharp, roll-off and tilt.

Sundaresan described the trained model as a “convolutional neural network,” or a way to understand what each of several layers is learning and then to remember that learning for future prediction. He said the CableLabs predictive engine was built open source, in Python, and with Keras as the neural network API for deep learning systems such as TensorFlow, CNTK or Theano.

Denver -- Comcast Cable president and CEO Dave Watson tagged artificial intelligence and machine learning as a next big area of focus, especially to mine device and network data in ways that improve customer care.

Read More: Complete coverage of Cable-Tec Expo 2017

“We just hosted an AI demo day (in Philadelphia), it’s incredible … if there’s ever an industry that’s primed and could benefit from taking data and making it available to agents and customers, it’s ours. I think it’s really going to change the customer experience, profoundly.”

Watson spoke here during Wednesday’s (Oct. 18) luncheon panel at SCTE•ISBE Cable-Tec Expo, with Arris CEO Bruce McClelland and Multichannel News technology editor Jeff Baumgartner. Major themes: Gigabit services, WiFi resilience, and mobile 5G.

McClelland, who subscribes to Comcast’s Gigabit service in the Atlanta area, described it as “really responsive and pretty addictive.”

As for what services will warrant that much raw throughput: “The last thing I’m worried about is whether we’re going to consumer all the bandwidth -- I think we are,” in part because of advancements in virtual reality, augmented reality, 4K/HDR video, and the Internet of Things.

Watson also emphasized Comcast’s wireless intentions, both mobile and WiFi.

“By 2020, there’s going to be something like 50 devices hanging off of the WiFi, so it better be good,” he said. To address in-home coverage, he referenced Comcast’s investment in Palo Alto, Calif.-based Plume, which makes WiFi “pod”-styled extenders that will essentially create a mesh network in the home.

RELATED: Comcast Launches ‘xFi,’ Invests in Plume

Arris is similarly focused on mobile 5G small cells, for higher speeds, better connectivity and improved latency.

“It’s not unlike the HFC network, as we continue to split nodes and reduce service group sizes … there’s an opportunity to be very disruptive in the mobile space.”

Watson noted that it took about a decade for 4G wireless to fully launch.

“I’m not sure it’s as dire as an either/or,” in terms of 5G’s potential as both a mobile backhaul opportunity, and a last mile replacement. “It can be complimentary and additive, but it’s going to take a while.”

Change is the new constant in the media and entertainment industry, as the fundamental nature of competition in the industry has shifted. No longer do television shows merely compete with television shows, theatrical releases with theatrical releases, video games with video games, etc. With more content now available than any of us can consume in a lifetime, all content and content offers are competing with each other in real time for a precious slice of each consumer’s time.

In this era of hyper competition and digital connectedness, a new business model is emerging, one powered by data, with the consumer firmly at its center. Data has become M&E’s secret weapon for sustaining and improving consumer engagement and satisfaction. It’s data that can reconcile a world of ubiquitous distribution with a consumer’s finite amount of time to consume content. And it’s just getting started: By 2020, data scientists estimate that for every human on the planet, there will be 1.7 MB of data being produced. Every second.

Leveraging that data is essential as M&E companies compete for consumers’ time, advocacy and money across multiple platforms. Those that know consumers best will win.

Enter cognitive computing, which has the ability to examine the choices consumers are making, and under what circumstances, in order to understand and predict what consumers will want next. Although some pundits have framed artificial intelligence and machine learning as a person-versus-machine comparison, at IBM we view the most productive implementation of AI as a symbiosis of person and machine.

Computers are good at rapidly processing enormous amounts of data and looking for hidden patterns and valuable attributes. The human brain is highly adept at other skills, such as generalization and abstraction. Together these make up Augmented Intelligence and the possibility of better decision making in all functions of business, from production and distribution to sales and marketing.

Two core domains are driving the application of machine learning. One is “audience” (customer or consumer) insight -- reaching a new level of personalization by understanding their affinities, traits, likes, dislikes, and how they respond to media. The other is “content” insight -- the enrichment of metadata and understanding what’s in the content to exploit it in new ways, in new formats and across new channels of distribution. Combine these, and M&E companies can apply cognitive insights to improve KPIs, including ad sales, content ultimates, productivity and efficiency, and margin growth.

Consider the ongoing Major League Baseball season. Using cognitive technology, media companies can now take baseball-related video content and better understand it in terms of sentiment -- including a closer look at actions that may indicate crowd excitement, or tuning in to crowd noise and the tone of the sportscaster’s commentary to gauge key moments -- and surface it to fans for a more personalized viewing experience.

Cognitive computing is a natural enhancement to existing business practices that will make day-to-day operations more effective while helping media companies create a better consumer/user experience. AI is a positive disruption, one that demonstrates the art of the possible. Most importantly, it’s an opportunity to thrive in an age of constant change -- the firms that are better equipped to extract insights from data are the likely winners.

Steve Canepa is general manager, global telecommunications, media & entertainment industry, at IBM. Image by John Lund/Getty Images.

Implementing artificial intelligence systems can be technically challenging and expensive, but it doesn’t have to be.

So says DimensionalMechanics, a startup based in Bellevue, Wash., that claims to have a developed a platform that can put A.I. within reach of a wide range of companies, with an initial  focus on those in the media and entertainment industry.

The goal is to lower that technology and economic bar in a way that makes A.I. “more accessible to organizations” without requiring them to have a technical background in areas such as deep learning and machine learning, company CEO and co-founder Rajeev Dutt, said, noting that many are also looking for A.I. solutions that are not just affordable but customizable as well.

To help achieve some of those goals, DimensionalMechanics has introduced NeoPulse AI Studio, a set of applications based on the company’s underlying framework that, it says, can help businesses and other organizations rapidly create and design customized A.I. solutions. That product complements the company’s pre-built AI models in areas such as image and video analysis and recommendations systems.

The company, which has raised $6.7 million and intends to raise a “B” round this fall, is also getting a boost into the media and entertainment world through a strategic alliance with GrayMeta, a company that specializes in automated metadata collection, curation and search.

GrayMeta, which counts ABC, AMC, CBS, Deluxe, DirecTV, Disney, HBO, NBCUniversal and Showtime among its clients, is also the first to offer NeoPulse AI to the media and entertainment sector, DimensionalMechanics said.

Dutt said the media, entertainment and advertising industries are among the biggest producers and consumers of data, providing a “proving ground for a lot of machine learning technologies.”

Some use-case examples include a photo-ranking system that was “trained” using 2 million images to determine which ones might make an ad or news article more likely to grab attention or drive and maximize traffic. The technology is also being used to help editors analyze and write headlines that can improve click rates.

On the video side, the company also provides A.I. solutions to drive recommendations.

“There’s a fairly broad range of applications,” Dutt said.

DimensionalMechanics has carved out a set of business models, including cloud software for independent developers, on-premises solutions that can simulate the cloud-based system while keeping a company’s data close to the vest, as well as a way for partners to resell and monetize their A.I. models through the NeoPulse AI Store.

Founded in 2015, DimensionalMechanics currently has 11 employees.

Philadelphia -- If customer experience is indeed the most important product, then it’s probably time to move machine learning up on the priority list. 

That was the gist of Monday’s kickoff workshop on machine learning and network operations, where executives detailed how the techniques can be applied to network optimization and customer care. 

“Right now, we throw away more data than we use,” said Tom Cloonan, CTO for Arris's Cloud & Network Solutions division and the moderator of the session. “We can do better, in terms of the decisions we make about how to modify plant, split a node or not, replace an amplifier, and ultimately, know whether customers are happy or not.”

Jason Schnitzer, founder of Applied Broadband, described his work to optimize DOCSIS 3.1-based networks using the three metrics of optimization: Modulation Error Ratio (MER), Codeword Error Ratio (CER), and downstream receive power -- no small feat, given that DOCSIS 3.1 essentially obviates traditional 6 MHz channel widths, instead spreading 7,680 OFDM (Orthogonal Frequency Division Multiplexing) subcarriers, between 24 MHz and 192 MHz of spectrum. 

“If you’re collecting data from tens of millions of devices, doing 7,680 samples, four or so times a day, you quickly get into very large-scale data,” Schnitzer said. Ultimately, DOCSIS 3.1 will enable operators approach the Shannon Limit by enabling multiple modulation profiles across the entire population of cable modems. “So, in a sense, optimization is necessary, for 3.1 to be successful.” 

Chris Menier, VP of products and marketing for Guavus, said machine learning is growing out of the industry’s early steps into “big data” -- which began in departmental silos. “Network operations, field operations, billing, care, they all had their own tools, reports, and dashboards,” he said. 

Next came data warehousing and federation, then automation use cases, he said. “Now, it’s about enriching and correlating the data -- where did it happen on the network? To what type of device? What was it playing when it failed?”

By detecting anomalies, and correlating them with additional data, operators can identify and address problems before they impact customers. And if care agents are armed with enriched data, in a timely manner, “you can turn someone into a promoter, from a detractor, in NPS [Net Promoter Score] terms,” Menier said.

By applying machine intelligence, anomalies can be detected, correlated and even automatically repaired. Does it work? “It sure does,” Menier said, to the tune of $70 million in savings for an unspecified operator.

Watchwith said it has acquired Arris’s Media  Analysis Framework (MAF), a cloud-based machine learning and automated metadata-generation platform that adds smarts to Watchwith’s native digital advertising platform for premium video.

Arris’s MAF technology is designed to understand the content of video automatically and its context, which plays a big role in deciding when and where native ads (i.e. ads that are placed inside the show itself) are presented in the video that a consumer is watching,  Zane Vella, Watchwith’s CEO, said.

That metadata, he said, brings automation to the cue points in the video where it makes the most sense to place the native ads – which can be interactive overlays or on-screens polls.

And that “machine learning” element? “The more information we feed into it, the smarter it gets,” Vella explained, noting that the analysis is based on the underlying video itself as well as the audio, including natural language processing.

As for context, this placement system for native advertising uses a variety of “recipes” for different types of content, whether it’s a news cast, a sporting event, or a serialized drama.

He said this capability, which is already being used and tested on shows such as USA Network’s Mr. Robot and for campaigns for LG and Toyota, is not only important for digital video delivered on traditional platforms, such as set-top boxes, but also on over-the-top streaming devices and smartphones. “The opportunity and need to monetize those assets in new ways is growing,” he said.

And, for these systems to achieve scale, the decision engine that makes these decisions must shift away from manual processes and instead become automated, which, in turn, reduces operations costs. Watchwith says the new automated process has the potential to eliminate thousands of man-hours.

Watchwith also believes that achieving that level of scale is also vital because programmers must find ways to build and deploy native digital video ad platforms that can compete with those from the likes of Facebook, YouTube and Snapchat.

“At a time when networks are reducing linear ad loads, the need for native digital ad products like Watchwith is higher than ever before,” Vella said.

Watchwith’s partners include Fox, NBCUniversal and Viacom.

Watchwith and Arris aren’t strangers. Arris is an investor in Watchwith, and the two companies have been working together on integration and product development for years. Vella said it was decided that Watchwith’s acquisition of MAF represented the best go-to-market strategy for the technology.

Watchwith and Arris aren’t disclosing the financial terms of the deal, but “several” engineers from Arris, including MAF leader Faisal Ishtiaq, PhD, that developed the technology are joining Watchwith and opening up a new R&D office in Chicago.

The deal comes just the day after Arris agreed to sell its Whole Home Solution assets to Ottawa-based Espial.

Get more #INTX2016 news.

Boston -- “Machine learning” will pave the way to delivering better experiences including “how quickly we can put apps on the X1 platform,” Comcast EVP and chief technology officer Tony Werner explained here in an introductory overview at the Imagine Park opening session on Monday. 

He envisioned “a ton of very cool apps,” comparing the ability of DOCSIS 3.1 to replace “a backhoe or forklift” in implementing new services for cable operators.

“I love Full Duplex,” Werner enthused about a symmetrical multi-gigabit project underway at CableLabs and being demoed here by Nokia, before he scooted off stage to support other Comcast executives at the opening of their nearby booth on the INTX show floor. 

As fellow panelist and visionary Andy Lippman, senior scientist and Associate Director of the Massachusetts Institute of Technology Media Lab, showed his “Ultimate Media” project, Werner quickly noted that, “You don’t have all the online sources loving it.”  Werner cited the prioritization for certain news sources within the MIT algorithm.

For his part, Lippman envisioned a shifting media landscape in which TV networks assume the role that Hollywood studios have taken in the motion picture industry -- they become venture capital sources for creative producers.  He called it a re-invention of the broadcast industry. 

Lippman also emphasized the speed of change, citing the rapid move away from schedule programming to on-demand programming.