February 15, 2018

What You Need To Know About AI

Lately, when clients come to me as a consultant, Artificial Intelligence (AI) usually comes up in our conversation. And when I’m asked, “How do I use it?” that tends to actually mean “What is it?” Let’s reach a basic understanding of AI so when you’re ready to explore what AI can offer, your discussion can be as productive as it can be.

What Is Artificial Intelligence?

As the name implies, AI is the intelligent behavior of machines. Most companies could utilize AI to interpret complex data. Here’s how it works (in a very simple way): The AI model asks a set of data a question and returns an answer. To accomplish this task, an AI model needs to understand the data it’s interpreting. So, for AI to deliver accurate, useful information, an AI model needs to be trained on the data it’s given. We’ll get into that soon, but first let’s talk about that data.

Learning From Data

For Artificial Intelligence to work, it needs to learn from specific kinds of data. With organized, not random, data, an AI platform can learn what it needs to. Let’s say you want to train an AI model to identify dogs in images. Organized data would consist of animal images, including dogs, to help the AI discern what is and what is not a dog. Random data (in this case, images of tables, lawnmowers, anything not reasonably close to our concept of a dog) doesn’t help the AI distinguish dogs from other animals. Know what’s in your data and you should be able to avoid any randomness.

At this point, I should clarify that AI isn’t actually telling you what something is — AI tells you what something probably is not. This is determined through a prediction percentage -- you’re not teaching an AI model to know an animal is a dog but rather training it to tell you it’s a certain percentage confident it recognized a dog. If you’ve been feeding your AI model with images of only dogs and cats, then introduce an image of a squirrel, your AI will be less certain of what it sees. But, once you teach the AI model what a squirrel looks like, the AI can discern with more certainty.

How To Train Your AI

Machine learning, the ability for computers to learn without being programmed, can take place a couple different ways. One way is supervised learning, where an AI model infers a function from labeled training data. Those images of animals I mentioned earlier would be labeled “dog,” “cat,” "hippopotamus,” etc. to help the AI learn. The other machine learning method is unsupervised learning. This is where machine learning draws inferences from data sets consisting of input data without labeled responses — you provide a bunch of pictures of animals with no descriptions and let the AI figure out what is and what is not a dog. Remember to, with either approach, provide organized data that your AI will learn from.

Going With Google

Google provides a lot of data sets and pre-trained AI models for purchase. But, they’re all about object recognition and will only do a good job of recognizing things that existed when the model was trained. So, a Google AI model may have learned what a bike and pogo stick are at some point, but a newer invention, like a Segway, could confuse it.

Beyond Cats And Dogs

I’ve led or been on some teams that have worked on training data for tasks other than image recognition, like classifying audio samples or, during an innovation session, recognizing what restaurant is delivering food by identifying food service workers and recognizing what was being delivered. The latter is an excellent example of what can happen when you train very specific models. In this application, the AI model learned from uniforms, not beverage cups, cars or even people. The data set provided in this case included shapes of what delivery drivers could be carrying and the clothes they wore.

The accuracy of your AI model is really about the bucket of data you give it. You need a lot of pictures, and they need to be of a similar quality. The more similar they are, the less training you likely need.

What Could Go Wrong?

Okay, our hypothetical AI is up and running. What is at stake if it’s wrong? Within your own organization, determine the impact of a false positive or false negative. AI that determines what is and isn’t email spam could afford to let some spam through but could create problems if it marks something important as spam. Imagine the consequences of AI providing results of x-rays with a false negative cancer diagnosis.

Now, if you feel more comfortable with how AI works, you can begin the challenging task of figuring out where it fits in your organization.

For information about Artificial Intelligence and Google Assistant, download our Google Assistant POV by filling out the form below.

Originally published on Forbes on Jan 8, 2018.

June 13, 2017

Engineering Jedi: Why I Couldn’t Resist Echo Show

For those who have been living under a rock these past couple weeks, Amazon announced a new member to their successful family of voice assistants: the Echo Show. I already own an Echo Dot and Amazon Echo. And while technically interesting, I didn’t see a compelling use case for the Amazon Look, (I know I look fabulous in my Target T-Shirts—I don’t need Alexa to confirm).

So why then did I plunk down two hundred and change to pre-order the Echo Show? I can think of three things right off the bat.

One screen to rule them all

One of the things that surprised me when I first started using my Amazon Echo was how many “smart” devices I already had in my home. Thermostat. Television. Lights. I also discovered after just a couple of weeks devices that didn’t work in the Amazon Echo ecosystem felt clunky. While at first turning on the kitchen light via voice seemed a novelty, in short order I found myself wanting to control everything from that centralized interface.

The challenge has been some of the more complex devices, my ecobee thermostat for example, can be tricky to get just right without some visual indicators. It’s true there is a phone app, and it’s true I could just walk to each thermostat and use the controls on the device. But the idea of being able to go to one screen on my kitchen counter that is always on for everything has some appeal. Whether I’m looking to adjust temperature in a different area of the house, know when the sprinklers are next scheduled to come on, or see what zone of my security system the cat just tripped, I envision the Echo Show could be my one-stop-shop.

A personal coach

Current incarnations of Echo have been a welcome addition to my kitchen. There are lots of cooking skills available for Alexa, (which is awesome because I have none of my own). She’s great at reading me recipes. But there are times when I just need some visual guides. (Did she say cinnamon or cumin? #worstfrenchtoastever). Of course, you can always ask Alexa to repeat something, or, stop and wash off your hands and then double check on your smartphone. However, a plethora of voice controlled video tutorials sitting on the counter next to my toaster is not without merit.

No more arguments over song lyrics

I listen to lots of music on Echo. The speaker, (even on the Dot when you consider the price point), is noteworthy. Interrupting to ask what the current song being shuffled is, or who the artist is sometimes kills my jam. Also, it stops me from singing along, which is something I do frequently much to the dismay of my son who is home from college for the summer. On occasion, he has even suggested I might not have got the words just right. (Apparently, “saving his life from this warm sausage tea” may not be an actual lyric in “Bohemian Rhapsody.” I’ll have to get back to you on that one.) My point is I think there is something pretty cool about always being able to glance over at the lyrics of whatever you’re listening to.

Of course, the Echo Show is not without its drawbacks. One of the things I most appreciate about my other Echo household members is that the interactions are omni-directional. Often I’m not even in the same room as one of my devices. I regularly shout from the top of my stairs (much to the bewilderment of my dog) asking Alexa for the temperature outside or what is on my calendar for the day. Requiring that I face the screen and knowing the limitations of my aging eyes to perform certain functions may take some getting used to.

Still, I’m betting Echo Show will prove as or more useful than her siblings. And if sales of previous Amazon Echo devices are any clue, I’m not alone. Working at a forward-thinking technology company like Bottle Rocket, we normally get access to developer kits and pre-release software. So, if you have an idea that seems particularly well-suited for the Echo Show you should reach out. One of our iOT strategists can help you hit the ground running.

May 8, 2017

Engineering Jedi: Alexa Lingo

It’s an exciting time for voice. Amazon’s Alexa has come into her own these last couple of years. Some analysts estimate as many as 8.2 million devices have been sold since late 2014. I personally find myself talking to Alexa multiple times a day, every day. It’s truly a remarkable feat of technology.

The engineer in me is fascinated by Alexa. And, being at Bottle Rocket where I work on the frontline of all things technology, I recently decided I wanted to write my own Alexa app, uhm I mean skill (which you’ll learn about later). Bottle Rocket promotes a learning culture, so I quickly tapped into other engineers and strategists here who were already tinkering (and in some cases, more than tinkering) with voice and lots of impressive things in the “personal digital assistant” space.

Much to my surprise, I found that even as a veteran engineer, I had some trouble following the conversation. While Alexa hasn’t even officially turned 3 yet, a whole vernacular has popped up around her that can be a little overwhelming.

So, before I rolled up my sleeves and started coding my first Alexa skill, I put together this handy little glossary of Alexa lingo.

Alexa Development Terminology

Wakeword

Except in the case of the Echo Tap, which has a physical button, Echo has multiple microphones that are always listening. Think of the device as being in standby mode. It is not fully activated and comprehending until you call out the wakeword. By default, this wakeword is “Alexa.” There are currently four other wakewords you can set on the device.

Skills

Skills are essentially apps for Alexa. The list of available skills for Alexa is growing every day. If you haven’t done so before, spend a few minutes browsing some of the most popular.

Invocation

The invocation is the word or words used to identify a particular skill. I’ve heard it described as synonymous to an app name, but I think a better analogy is the app icon since you may choose to call your skill “Greatest Alexa Skill” but might settle on an invocation word that’s less of a mouthful, like “G.A.S.”

Intent

This one doesn’t directly relate to the spoken script with Alexa, but rather intent is the “what” in what are you trying to accomplish by speaking to Alexa in the first place.

Utterance

Utterances represent the variances of spoken language and all the nuance that implies. Think of all the different ways someone might ask about the weather. What’s the weather? What’s my weather? What is my weather? What is the weather like? That list can get very long very quickly. Getting utterances right can be tough, but Amazon’s guidelines are helpful.

Slot

Slot is another word for what programmers and mathematicians call variables. If you think back to algebra, x in the equation 50+x=75 would be the variable. In Alexa’s vernacular x is the slot.

Developing for Alexa

Now that you know the terms in play, you can begin to see how they fit together.

Wakeword
Invocation
Utterance
Slot
Intent

Alexa, ask Southwest about my flight info.
<Respond with information about an upcoming flight>

Alexa, ask Coke Freestyle for today’s top mix.
<Respond with information about the most popular Freestyle mix for today’s date>

Alexa, tell NPR to remind me when Way With Words starts.
<Set a reminder for when the program “Way With Words” is scheduled to next air>

Eureka! Now you’re speaking Alexa!

Of course, a lot more goes into building a great voice experience than just understanding the terminology. Publishing an Alexa skill is a blending of engineering, strategy, and quality assurance. Amazon’s submission process requires knowledge of policy guidelines, cloud-based security, and a combination of functional and experiential testing. Lucky for you (and me), my colleagues here at Bottle Rocket have a head start.

I encourage you to schedule a demonstration of Bottle Rocket’s voice expertise. Even if you aren’t quite sure how an Alexa skill fits into your overall digital strategy, seeing some of the exciting work going on here will get the wheels turning

© 2020 Bottle Rocket. All Rights Reserved.