Make vs. Buy: Applications Exploiting Sentiment Analysis Services

In reviewing key end-of-year reports, articles and research summaries, a few stand out. Most notably, let’s look at Machine Learning as a Service: Part 1 (Sentiment analysis: 10 applications and 4 services) from Towards Data Science.

Two sections of that June 2018 report stand out:

1. What can I do with sentiment analysis?

This section sets the stage for coming up with potential enterprise use cases. It lists ten good examples from published academic research.

Actions: Ideate

  • Check the references in footnotes 4 through 13
  • Identify all the text streams your enterprise already captures somehow (and other related text streams you might also exploit, such as Twitter commentary that somehow relates to your business)
  • Given the text streams, which of the ten use cases might apply to your industry and business?
    • Don’t limit yourself to single use cases
      • How might you combine use cases to impact your business?
      • Consider both opportunity and threat scenarios
      • What happens to your business if a competitor emerges that is exploiting these services (or applications that depend on use of such services)?
      • Do not make technology implementation assumptions at this stage. (All four vendors listed in this research provide sentiment analysis services from “the cloud.” But there are non-cloud implementations if you need them. Park this issue as you conceptualize potential uses and experiment with the cloud based technology.)

2. What are some good sentiment analysis services?

This section is a straightforward snapshot of sentiment analysis services. It describes and evaluates Amazon, IBM, Google and Microsoft mid-2018 services. (Details will continue to evolve as the service providers enhance their offerings.)

Action: Experiment with the technology.

  • Assess the feasibility of the business scenarios you’ve envisioned above.
  • This is not the same thing as building the “solution” yourselves.

Most enterprises should not build:

  • The underlying Natural Language Processing technologies
  • Their own sentiment analysis services.
  • The production application that relies in whole or part on sentiment analysis services.

Most enterprises should:

  • Experiment to get a grasp of the feasibility of achieving their business objectives
  • Seek out specialist firms with demonstrated subject matter expertise in their industry
  • Follow the section on Five Easy Criterial To Seek described in our blog post on 5 Easy Criteria To Get Quick Returns on AI Investments

Disclosure: I am the author of this article and it expresses my own opinions. This is not a sponsored post. I have no vested interest in any of the entities mentioned in this post. Neither does the Analyst Syndicate.

Drone Dirigibles… No Really: Dirigibles!

If you’re hanging on to images of the Hindenburg, it’s time to let those go. …

No matter how hard I would try to talk about drone dirigibles with clients, I could literally hear their eyes roll back in their heads. I guess compared to quadcopters racing around dropping packages on people’s heads, a safe, dependable solution that has been called the “Swiss army knife” of air vehicles just doesn’t sound sexy.

True, dirigibles aren’t designed to win the sprints.  They are the tortoise in the fable, however.  

If you’re hanging on to images of the Hindenburg, it’s time to let those go. They haven’t been flying hydrogen bags for a while now, and when was the last time you heard of a dirigible crash?  Most dirigibles have safety features that make a fast, hard landing such as a crash nearly impossible. 

Networking

High Altitude(HA) drone airships can be used to replace satellites for communications. Signals don’t have to penetrate the atmosphere or travel as far, and that means less lag time. If something goes wrong with the drone airship it can land for repairs or be retrofitted with new equipment, which would be a costly endeavor for satellites if it could be done at all. Facebook and Google both have invested in HA fixed wing drones outfitted with solar power that are capable of staying in the air for long periods of time, but the problem with a fixed wing remains the same closer to the ground as it does up high: it moves fairly quickly compared to things that hover. Google also had been working on Hot Air Balloons to provide networking to remote areas of the planet.

Big Movers

Currently, if you need to manufacture a product you deal with a lot of shipping. The raw goods have to be shipped to either a refinery or a foundry or to wherever the manufacturing plant is. Then they leave that process and travel farther into the supply chain. Product moves from the interior of a country via infrastructure to the ports, and you guessed it: more shipping.  One of the pluses of dirigibles is they can carry a lot of weight over long distances for very little cost. Again, they won’t win any speed races, but they will show up with larger payloads and they don’t require much by way of infrastructure to get the job done. In a country where torrential rainfall can wash out roads and bridges, for example, dirigibles can fly above the damage.  Got a problem with pirates raiding freight ships? Dirigibles fly high over them. Same for getting humanitarian relief to areas experiencing natural disasters, or to remote places that have traditionally been difficult to provide support, like the Antarctic.

Um… About That Border Wall and Surveillance

The Cato Institute in one of its reports said border security using drones was not feasible. That report was based on using Predator drones, which is misleading for a number of reasons and is not the only option. The report also did not engage any level of imagination to use existing technologies, or to modify, outfit, or change configurations of the Predator for applications in a peacetime role. In any case, an airship’s flexibility lends itself perfectly to this sort of mission. It can be outfitted with any number of sensors, including multispectral image sensors, microphones, sonar, LIDAR, radar, and many others. Not only can they spot something or someone coming to the border from miles away, but track it, report its location, and keep tracking it even if it goes over the border. Whether it’s a low flying smuggler or a desert owl, it can be monitored. With some AI, image and facial recognition software could be employed, which means we’d know who or what was at the border, and even if a person made it across the border, with facial recognition software, it would be hard to stay in the country for long without ending up on a local police or federal monitor at some point.

The other great thing about dirigibles is that they can be tethered–you can have them in remote locations, with a ground station below and all the power they need. If the power goes out, or you want them to be more self-sufficient, you can use thin-film solar for the outer skin of the craft.

China developed the Yuanmeng, which is the largest airship at 75 meters in length, and from a military standpoint it can be used to monitor ground and/or sea activity. It can also be equipped with radar capable of monitoring all air traffic; there have even been suggestions that it could have quantum radar capable of detecting stealth planes and other vehicles.

These dirigibles can be used in other stationary surveillance situations, like stadiums, large outdoor events, prisons, high traffic areas, search and rescue, patrolling long stretches of property, or pipelines. They can monitor rivers, lakes, damns, and ports. Small low altitude versions are relatively cheap, do not require much power (if any solar is used), and are easy to maintain or upgrade. And small hobbyist dirigibles are capable of achieving high altitudes, so if you were monitoring your acreage on a farm, you could do that 24/7 with the right  payload and  tethering with a power source.

Travel?

The Chinese are working on a civilian version of the Yuanmeng for air travel. I imagine this would be cheaper than traditional air travel. It would take longer and would be an alternative to buses for a country that has notorious infrastructure problems during the rainy season. It would likely be similar to a train carrying both passengers and commercial freight. 

A Platform

Because dirigibles are capable of slow circling and staying over an area, they would make excellent platforms for other devices to land on, recharge, and continue their missions.  There are some plans for using a dirigible as a platform for swarming drones. You may have seen the video of swarms being launch from a larger jet plane. That makes them essentially a single mission deployment, whereas a dirigible platform would make for multiple swarming missions. This could translate to a civilian applications, like mail delivery or surveying.

Downsides and Takeaways

Speed is the one major downside. Dirigibles are not going to be winning a Famous Company’s Pizza Delivery contract so the customer doesn’t wait longer than 30 minutes. However, they are a stable platform for moving freight and can carry more than their rotor and fixed wing cousins, so they might be a solution for the Last Mile, especially for items weighing over a pound. They also make fairly large targets for the anti-drone crowd. Obviously, there are not as many manufacturers of drone dirigibles and the sizes vary from small (about a meter) to very large (tens of meters). Dirigibles may become a more flexible and cheaper security solution for both home and commercial use.

If speed is not a factor, airships might be the right solution for data gathering in your business. Or a multifaceted approach might be your answer–traditional drones for just the speed work, dirigibles for constant surveillance and heavy lifting missions. Much to consider if you are thinking about adding drones to your operation.

The Longest Last Mile

Imagine a rain of hot soup, a hail of pills, and a meteor shower of packages…

The Logistics of Delivery Drones

Don’t believe the hype…

I used to get a question from clients on a fairly regular basis—the question went along the lines of: When do Delivery Drones begin being a thing? My answer would disappoint most, but really you have to kind of do the math on this one, before you realize just how much has to be done before we get there.

Besides the obvious problems of battery life versus distance along with weight and how much a single drone delivery costs and how many flights before ROI. There’s problems about where and how to land, like should we all have designated landing areas in our yards or rooftops. If it’s a building rooftop, who does the sorting and delivering or is everyone’s stuff just left in a pile? Assuming it’s not a giant online retailer, but say the local pharmacy, do you drop the drugs without a human presence, or do you have to hover until a human with a pin code arrives? And where and how does the drone recharge? I’ve seen and heard so many possible systems!

But even beyond all that logistical dilemma lies a much larger one…

The real logistical issues are going to be with the UAS Traffic Management (UTM) systems. It seems like such an easy thing until you think about a city like Manhattan. You’ll have drone messengers, food delivery, delivery services, laundry services, diaper services, retail delivery, groceries, and just about every other thing you can imagine.  Millions of drones. All in a perfect and non-stop ballet of precision and avoidance!

…and that’s not considering the thousands of pilotless air taxis!

Or an entire system of priorities, like emergency equipment, police and fire drones, unmanned ambulances, etc., etc.

Sunny with a chance of hot soup rain and possible pharmaceutical hail…

The photo above was a common one at the turn of the 19th century, when there might be only two cars in an entire county and yet somehow they’d find each other.

Imagine a rain of hot soup, a hail of pills, and a meteor shower of packages…

How do you go about creating a system that will not only manage the flights of millions of payload-heavy drones but also monitors the manned flights above and intersecting that air space as well as any unidentified objects and what to do about them? A system like this will no doubt require AI, and all of the AI trappings/baggage that will go along with it. But more importantly, how does the system track and manipulate millions of devices down to the centimeter.

The UTM is going to become an incredibly intricate part of any city’s government. It will need to have the ability to include special missions, new parameters, and will no doubt be a huge budget line item for maintenance and repair. It’s going to need auditors and supervisors and investigators. It will need to interact with other city systems, private business systems, as well as state and federal systems. If you’re getting the impression that this is going to cost millions if not billions of dollars, you’d be right.

Currently, there are UTM tests under way at the FAA test areas. NASA has been testing UTM at its Ames facilities, and other private companies are also testing UTM.

How does a system like this work? The details are still sketchy. But here’s what you can probably expect. Companies and individuals will register with the UTM (either an area or city system) and will get a hardware package to add to their drone.This will make sure each drone has the minimum requirements to operate in the system and will include a communications chip like 5G, an altimeter, a GNSS chip (similar to a GPS chip but much more accurate), a chip that stores flight data and a unique ID.

And this is just from the city side… you still would have a whole other system on the drone owner side that would include, package data such as weight and mass, energy management,craft maintenance, mileage, flight logs and any additional pertinent data.

5G Most Disruptive Technology Change Ever

Always look at infrastructure changes to make easy predictions about the future. You could get very rich.

A decade ago I attended meetings around the world where the topic was “how can we, as a country, join the Internet revolution?”   Brazil and Columbia stick in my mind. Don’t even get me started on Australia and their wasteful endeavor to create a National Broadband Network(NBN). I never had the floor but I wanted to stand up and shout “deregulation!”  That is what sparked the internet revolution in the United States. In 1993, here in Michigan, it cost 8 cents a minute for telephone calls that went outside your immediate area code. You could be a mile away from your ISP’s nearest POP (Point Of Presence) and see outrageous phone bills that ratcheted up quickly at $4.80 an hour.  At RustNet we sold internet access for $19.50/month. If we wanted to get customers in a different area code we had to put stacks of dial-up modems in an office in that area code. Then we backhauled the traffic to our main office and sent the packets out to the internet through our upstream provider in Chicago. (Anyone remember Net99?).

The big break up of ATT had occurred in 1982 and the regional telephone companies (Baby Bells) started to compete for your business after the 1996 telecom deregulation. Per minute charges went away just in time to fuel the rapid growth of internet subscribers. By that time the telcos offered their own backhaul so you did not need to maintain huge stacks of modems in every POP. You just paid for a T1 to the telephone company’s Central Office (CO) and they delivered the calls to you.

In 1995 I published a business plan for How to Start an ISP. It gave me great visibility into the wave of deregulation that was sweeping the world. As each country figured out that per minute charges were holding them back they would deregulate, encourage competition, and I would see sales of the plan going to that country. South Africa and Mozambique used my plan as a starting point. The internet took off.  By 2005 you could tell which countries still had per minute charges. They had Internet Cafes because people could not afford to dial-in.

Of course 4G spelled the end to all that. Now you can get internet on your phone and, if you can tether your phone to your computer, you use that for internet access. I can get 95 Mbs over Verizon 4G.

Well 5G is going to explode many things. And it is coming fast. Ericsson predicts there will be one billion 5G subscribers in six short years.

What is different about 5G?  It is very, very, fast. Huawei has tested 5G connections at 70 gigabits per second. Gigabits. At that speed even immersive experiences like SecondLife will work. No wonder people are excited.

But what could this do for security?

5G introduces new networking paradigms. It is going to have dramatic effects on the Internet of Things (IoT) as very small, low power radios will be able to connect. That will pose an opportunity for data theft and continue the weekly news cycle of privacy violations that we have come to know and love.

But think about what these speeds will do to your typical enterprise (and SMB) networks. Why would anyone use the pokey internet connection at work when they get 5G at home and on their smart devices?  Businesses have already moved the critical tools they need to the cloud, (email to Office365, document sharing to Microsoft hosted Sharepoint or Google Docs, or DropBox, HR systems, Salesforce, etc.) They don’t need your network at all. And if you force them in through a VPN they are going to be tunneling through your pokey network to get access to those mission critical services.

Ever see the scene in Gettysburg where General Buford rants about how clearly he can see what will happen in the morning? 

The hardwired connection is dead for office use. Sure, every firewall vendor will add 5G radios to their UTM devices for remote offices and HQ, just as they have added 4G. But going through a gateway means dealing with the slow wifi in the office.  It will be faster for users to jump on the 5G network themselves. So they will.

Goodbye cable triple play. We won’t need twisted pair, CAT5, or fiber to the home anymore. All home devices, including your TV, will connect directly to the internet via 5G.

New, very fast growing, businesses will start up to address these problems.

Here is what happens next.

Stage 1. A startup that is probably already out there will introduce a policy overlay to the carrier networks. An enterprise will just enroll all employ devices and manage what they can do over the network. It will be like a virtual UTM. They will encrypt traffic, filter content, and apply firewall rules. Managed Service Providers will do that policy work for SMBs.

Stage 2. The carriers will recognize that they have created a monster as every enterprise starts cancelling their lease line subscription. Seeing the opportunity they will start to develop their own service offerings for security.

Stage 3. One carrier, late to the game, will acquire the fastest growing 5G security management platform from Stage 1.

Stage 4. All the other carriers will cut off that 5G management platform for their own networks and make their own acquisitions.

Stage 5. All carriers will bundle security into their offerings. Network security will finally be part of the internet.

 This whole time frame will play out by 2030.

Thank you technology.

Originally published at Forbes.com December 6, 2017