Want to Know More About Machine Learning and AI?

Wondering whether you should invest in AI and Machine Learning? That’s a question that the most innovative companies are considering. Why consider it? One good reason is because your competitors have already started. If that doesn’t give you some reason to get motivated, I hope you get started before you are put out of business. To make sure that doesn’t happen, there are  a few things to consider to help you start to explore an investment in machine learning.

It’s the Data, Stupid

Of course, as with any business initiative, you’ll want to create value. And this can be done using machine learning systems. But for those systems to provide value, companies will need to begin by evaluating their organization’s data maturity, but more importantly their readiness to accomplish its data-driven goals. Company’s need to start with an audit of their data warehousing, data scientific research capabilities, data governance and data hygiene. In addition, it’s important to look at the sources, uses, volume, and veracity of all your date, meaning your first-, second-, and third-party data.

Garbage in, Garbage Out

Why is making sure your data so clean? Machine learning is basically taking a computer and making it smart enough to learn from the data it’s fed. We are essentially programming machines to learn. The goal is that after a certain point of time, the computer is able to predict further data. How so? Let’s pretend you want to make your computer predict the weather. So to begin, you might feed the computer weather reports of every hour of every over the past year. What you might end up with is– because the temperature (z) depends on day of the year (x) as well as the time of the day (y), more than two-dimensional curve. In fact, weather is random, so the equation generated by the computer won’t just have 3 variables (x, y, z), it may also have higher powers. So depending on the number of factors in a prediction and the randomness of the outcome, the complexity of the curve can increasingly get more complicated.

So back to the data… And I know you know the story about data: garbage in, garbage out. So hopefully, now you see can why good, clean data is so important to prediction. As the computer is taking the data you feed it to make future predictions, those predictions dependent on the data you are feeding it. So you want the very best data possible. And it takes super computers which are capable of handling large volumes of data, as well as the ability to learn fast and to make fast decisions based on the learning it under goes.

AI and ML Are Not The Same

Often times Artificial Intelligence (AI) and Machine Learning (ML) are used interchangeably. But they are actually different. Artificial Intelligence is the broader concept of machines being able to carry out tasks in a way that we would consider “smart.” Machine Learning is the application of AI based on the idea that we should be able to give machines access to data and let them learn for themselves. Artificial Intelligence devices (devices designed to act intelligently), are often classified into one of two groups: 1) applied and 2) general.

Applied AI is far more common. Applied AI is about systems designed to intelligently trade stocks and shares or drive an autonomous vehicle. Generalized AI is may up of systems or devices that, in theory, can handle any task. And are less common. However, this is where some of the most exciting advancements are happening today.

Deep Learning is A New Area of Machine Learning Research

It was introduced with the objective of moving Machine Learning closer to one of its original goals: that of being Artificial Intelligence. So essentially Deep Learning is a subfield of machine learning concerned with the algorithms inspired by the structure and function of the brain called artificial neural networks. Deep learning has worked it’s way into business language via Artificial Intelligence (AI), Big Data and analytics. Deep learning is an approach to AI which shows great promise when it comes to developing the autonomous, self-teaching systems which are revolutionizing many industries.

The Two Big Ideas: It May Be Possible To Teach Computers to Learn and The Internet is a Source of a Ton of Data

Arthur Samuel, in 1959 is credited as the one who came up with the big idea that it might be possible to teach computers to learn for themselves. That would be in contrast to teaching computers everything they need to know about the world and how to carry out tasks. The second big idea was that the Internet, with huge increase in the amount of digital information being generated, stored and could be used for analysis. So the scientists and engineers realized it would be far more efficient to code computers to think like human beings, and then plug them into the internet to give them access to all of the information in the world.

Neural Networks Are Algorithms

Neural networks are a set of algorithms, modeled loosely after the human brain and designed to recognize patterns. The development of neural networks has been key to teaching computers to think and understand the world in the way we do, in addition to the innate advantages they hold over people such as speed, accuracy and lack of bias. So a Neural Network is a computer system that classifies information in the same way a human brain does. It can be taught to recognize, for example, images, and classify them according to elements they contain. It works on a system of probability – which means that based on data it’s fed, it is able to make statements, decisions or predictions with a degree of certainty. The addition of a feedback loop enables “learning” – by sensing or being told whether its decisions are right or wrong and then can modify the approach it takes in the future.

What Can Machine Learning Applications Do?

Machine Learning applications can read text and work out whether the person who wrote it is making a complaint or offering congratulations. They can also listen to a piece of music, decide whether it is likely to make someone happy or sad, and find other pieces of music to match the mood. They can even compose their own music expressing the same themes, or which they know is likely to be appreciated by the admirers of the original piece.

These are all possibilities offered by systems based around ML and neural networks. The idea is that we should be able to communicate and interact with electronic devices and digital information, as naturally as we would with another human being. And another field of AI – Natural Language Processing (NLP) – has become an exciting area of innovation in recent years, and one which is heavily reliant on machine learning. (And yes, my initials just happen to be NLP, but that doesn’t really mean anything… just a happy coincidence…)

Where is Used?

Take Google for instance. Google is using it in its voice and image recognition algorithms. It is also used by Netflix and Amazon to decide what you want to watch or buy next. And it is also being by researchers at MIT to predict the future.  While Machine Learning is often described as a sub-discipline of AI, we might look at Machine Learning as the state-of-the-art of AI. Why? Perhaps because it is showing the greatest promise to provide tools that industry and society can use to drive change.

More on the practical uses of AI and ML in the future. For now, noodle on that!


VP, Program Executive, Innovation and Transformation Center




Innovation & Disruption In Delivery: Could Your Next Amazon Delivery Be By a Drone?

Ever Had a Package Delivered Late, Not at All Or To the Wrong Address?

One of the most irritating issues with ordering online is whether the package gets delivered and to the right address and on time. I know I’ve experienced this a number of times and what’s interesting is that, as customer’s we don’t always think about the delivery service as the issue, but rather it reflects poorly on the company we by the product from. How to fix this customer experience issue? One option is to deliver packages to consumers’ homes using drones. Could this allow companies to bypass the challenges with that last step of the delivery? It might be for delivery to people’s homes. It might not work at apartments, though, because the drone can’t get into the apartment building. Or can it?

What customers may not know is that the last leg of the delivery is the most expensive and inefficient part of parcel delivery. Customers don’t often think about that the product has to go from a store or warehouse, to the shipper’s delivery center and then from there, be deployed to the customer’s address. It is often not the place you bought the product from that is having the issue. It maybe who their delivery service or services are. It could be the individual who works for the delivery service. I know I personally had package delivered to an address that was similar to mine, but not mine. The individual was new to the delivery route and got mixed up. I had to run after the delivery truck, stop them and tell them they delivered to the wrong building. (I had gotten a text my package was delivered, but it was not on my doorstep or at the post office boxes for my building.) And since this happened more than once, I knew what had happened.

What’s the Solution To Better Customer Experience Delivery?

E-commerce companies, like Amazon, are using drones to speed up the this last part of the delivery process, while cutting costs. The result? Improving the customer experience, customer satisfaction and loyalty. And what’s interesting is even legacy retailers could take advantage of a similar process to grow online sales.

So What’s the Hold Up?

While there are many obstacles to overcome for instance, drone regulations, the development of autonomous flight and traffic control systems for drones, as well as consumer acceptance, there are companies actively trying to figure this all out. For instance, Amazon is working on drone delivery, depending on when and where they have the regulatory support needed to safely delivery packages. They want to use drones to deliver packages to customers around the world in 30 minutes or less. In fact, they have Prime Air development centers in the United States, the United Kingdom, Austria, France and Israel.

Amazon Prime @drnatalie

Photo Source: Amazon

They believe the airspace is safest when small drones are separated from most manned aircraft traffic, and where airspace access is determined by capabilities. To learn more, you can look at Amazon’s airspace proposals here: Best-Served Model for Small Unmanned Aircraft Systems and Revising the Airspace Model for the Safe Integration of Small Unmanned Aircraft Systems.

Disruption to Delivery Logistic Firms

As e-commerce providers like Amazon look for solutions within their own company, many logistics providers are experimenting with drone delivery. These firms also seek to cut costs as well as ward off competition, whether it’s from startups, technology companies or e-commerce companies. In fact, FedEx is betting on automation to Fend off contenders like Uber and Amazon. The shipping giant is investing in autonomous trucks and is interested in delivery robots, drones and an Alexa app. And while there are attempts to get this right, those of us in the innovation space know that #failfast – iterating and pivoting is the key. In my book, it’s ok to fail. You can’t learn what you don’t know, you don’t know unless you try. Trying means you learn something each time. Though the concept of failfast is very popular today, if we look back at Edison, it took him 9,999 times to get the filament for the lightbulb to work on the 10,000th time. What if he gave up? We’d all be in the dark!

How Is Amazon’s Prime Air Trial Drone Deliver Program Progressing?

Amazon have started with a private customer trial, to gather data to continue improve the safety and reliability of their systems and operations. As they gather data, this will bring them closer to realizing this how to use this innovation for all their customers. Does weather affect the delivery? Currently, Amazon is permitted to operate during daylight hours when there are low winds and good visibility. However, they are not using it when it rains, snows or in icy conditions. They feel they need to gather more data to improve the safety and reliability of their systems and operations to expand the offering. They are working with regulators and policymakers in various countries in order to make Prime Air a reality for customers around the world.

Video Source: Amazon

Where Can you Find more Information On the Disruption and Innovation Drone Delivery Can Provide?

In a new report, BI Intelligence examines the benefits drone delivery can provide as an e-commerce fulfillment method. In the report, they look at the different approaches companies are taking to experiment with the new technology and processes involved in this new delivery process. In addition, they look at the key players working in the drone delivery space. And have researched the challenges drone delivery faces in reaching mainstream adoption.

Will Your Industry Be Disrupted? Every Industry Should Be Thinking It Will Be Disrupted!

As I was giving a talk on disruption and innovation, I had many questions from what would be considered very standard legacy firms. What they need to be careful of is being aware of the fact that somewhere, in someone’s basement or garage, someone is probably working on a project that will disruption their industry. It’s customary to do the ostrich: stick you head in the sand. But doing so will only make you a dinosaur, (extinct) if you are not careful.

Disruption and innovation are all around us. Just look at what happened to the taxi industry. Not only did Lyft and Uber transform how customers’ order, receive and pay for rides, but they disrupted an age old industry that had not changed for years. And take GM for instance. They make cars. But they decided to look at cars as a service and invest $500M in Lyft to be part of the cars-as-a-service industry.

Disrupt Yourself or Die

Instead of being one of those industries or companies that waits until an upstart disrupts their revenue model and takes marketshare, why not start innovating within your own company. Too many companies are complacent or don’t have the skills to think outside the box. If you don’t, it may want to seek out a firm that can you help you think through this new and confusing new frontier of design-thinking, innovation and disrupting yourself — as a company and as a person. No one wants to be the company that had the leg up on IBM and caused it’s own demise: i.e, nobody wants their story to go down like Digital Equipment Corporation: DEC.

“Digital Equipment Corporation achieved sales of over $14 billion, reached the Fortune 50, and was second only to IBM as a computer manufacturer. Though responsible for the invention of speech recognition, the minicomputer, and local area networking, DEC ultimately failed as a business and was sold to Compaq Corporation in 1998. The  fascinating modern Greek tragedy in book form by Ed Schein, a high-level consultant to DEC for 40 years, shows how DEC’s unique corporate culture contributed both to its early successes and later to an organizational rigidity that caused its ultimate downfall.” Don’t do a DEC.


Natalie Petouhoff

VP, Program Executive, Innovation and Transformation Center | Salesforce.com