Predictive Models As A Service Vs. Training As A Service: Part One

Erik Marcade

Machine learning is on every CXO’s mind at this time. We’ve all heard and tested many use cases for machine learning, spanning various domains. I would like to focus on one point in this blog: the emergence of two distinct categories of machine learning solutions, depending on the type of problem that needs to be solved.

This week and next, I will describe in detail these approaches, which I call “macro modeling/predictive models as a service” and “micro modeling/training as a service.” In addition, I’ll highlight some of the challenges that technology executives need to be aware of when making investment decisions.

Predictive models as a service—macro modeling

When talking about machine learning, some obvious use cases that come to mind are autonomous vehicles and machine learning-powered translation systems. These could be described as general-purpose systems powered by predictive or machine learning techniques. Let’s see how these are generated and consumed.

Challenge 1

First, we’re talking about one system that has been trained on a very large corpus of data in order to get the proper results in many different situations. For example, Toyota says it needs 8.8 billion miles to create a safe autonomous car. The same is true for image recognition, where public image data sets contain more than 100 million images (and the true internal data sets used by Yahoo and others are much bigger than that).

To build a general-purpose predictive model thus requires a gigantic amount of data that you have the right to use for this purpose.

Challenge 2

How many intelligent general-purpose systems do we need? Today, we have many teams focused on autonomous vehicles, image classification, or even translation systems. But how many of these systems do we need on the planet? If a system to drive autonomous vehicles is efficient enough to beat the competition, you can expect that there will be 10 systems or so to equip all the cars on the planet. We expect these systems to work well in cities, in the countryside, during day and night, and so on. Producers will compete on the price and reliability of the sensors.

The same is true for translation systems or even image recognition systems. This is a winner-take-all market. If we push it to the limit, how many true artificial intelligence systems do we need on the earth?

Challenge 3

As always with predictive and machine learning, it’s almost never a “fire-and-forget” activity. Your systems need to be continuously updated as new data comes in or specific rare situations occur, which means that you need to connect them to continuous feeds of data collection for continuous updates and monitoring.

This continuous improvement feature has cost impacts, of course, that will push the need for continuous learning or incremental learning. This in turn will also be used in order to start from general-purpose predictive models to specific models for specific contexts.

Technical challenges

Of course, the fact that these systems can be transported is important. It’s nice to have such systems available as REST APIs on the cloud. This means that they will be available only within connected environments, which will solve many, but not all, of the use cases. Typically, an autonomous car must be able to run even if there is no connection.

On the shared models, through services, speed, and concurrency is very important, as well as exchanged data security and privacy. These are technical challenges that have been solved in the SAP Leonardo machine-learning foundation based on SAP Cloud Platform with Cloud Foundry.

Finally, we’re talking about very large data volumes and very large computing power, which impact on direct operation costs.> Consider, too, the fact that we can expect more improvement on this financial equation in the future (such as the introduction of ASICs) and even on pure electrical consumption, not to mention the amount of data traffic.

Next week, I’ll discuss training as a service – micro-modeling.

See how you can turn insight into action, make better decisions, and transform your business.

Comments

Erik Marcade

About Erik Marcade

Erik Marcade is vice president of Advanced Analytics Products at SAP.

Data Geeks Rewrite The Business Rules Playbook

Robin Meyerhoff

When it comes to digital business, Andrew McAfee knows a thing or two. A principal research scientist at MIT, prolific writer, and management expert, McAfee is a leader in understanding and explaining how digital technologies are changing business, the economy, and society.

At the recent SAP Leonardo Live event in Chicago that focused on digital transformation, McAfee urged his audience to throw out the business playbook they’ve been using for the past 30 years.

“The right way to run a factory in the steam era became a really, really bad way to run it in the era of electrical power,” he said. “Similarly, during a technology transition — and afterwards — the advice you used to follow becomes bad advice.”

McAfee explained that fast, profound shifts are occurring in three key areas: process, company, and industry. And he provided a new playbook to help companies navigate those changes and succeed.

Process: From people to machines

The traditional wisdom about process, which McAfee defines as “getting stuff done,” is to let machines handle the routine work like accounting or record keeping, and have people use their accumulated wisdom to make the judgements calls. This is the playbook of yesterday.

“Profound shifts are occurring in three key areas: process, company, industry”

McAfee explains that in most companies, decisions have typically been based on the highest-paid person’s opinion, or “HiPPOs.” They follow their gut, past experiences, and education, but they are being threatened by what McAfee calls “the Geek” — people who use data to make decisions.

“When the Geek needs to make a tough call, they gather evidence, do the best analysis they can, then they follow the evidence — even if it doesn’t go along with their gut or their experience,” McAfee explains.

“But here is where things get interesting,” he says. “In 136 studies of decision making by HiPPOs versus Geeks, 48 percent of the time HiPPOs added nothing over Geeks’ approach. Furthermore, 46 percent of the time HiPPOs provided an inferior decision. HiPPOs were only clearly better in eight percent of the cases. We need to make HiPPOs an endangered species.”

McAfee believes that with artificial intelligence (AI) and machine learning, “Now we have a new toolkit to help us sift through these crazy amounts data, see patterns, and make very sophisticated, accurate judgements in extremely complicated situations.”

He explained that AI and machine learning technologies have leapfrogged much further ahead today than anyone could have anticipated, and are ready to take over making judgement calls.

“Go is 3,000-year-old Asian strategy game. Computers have been laughably bad at Go. Until last year, when the world’s best Go player became a computer,” said McAfee.

Analyzing the game played by AlphaGo, a Google AI company, experts focused on one particular move — move 37 — that made no sense to human players but ultimately helped the machine win. The lesson learned? AlphaGo doesn’t just play the game better than we do, it plays differently than we do.

McAfee is optimistic: “Together with machines, we’re going to make progress in some very difficult areas. And when we rewrite the business playbook, remember: machines are demonstrating excellent judgement.”

Company: From core to crowd

“For about 25 years we’ve been telling business that to succeed they need to strengthen their core — ‘core competency, core strength, core capabilities,’” said McAfee. “The idea of the core is a small number of things that differentiate you from competitors, realize value for customer, help you succeed in your markets.”

But, he explains, now there are millions of interconnected adults on the internet and if you can activate the energy of the crowd, amazing things can happen.

McAfee provided an example where a Harvard Business School expert on crowd sourcing and innovation Karim Lakhani worked with the National Institute of Health (NIH) and Harvard Medical School to try and improve the ability to sequence human white-blood cell genomes. They got good results.

But when Lakhani opened up an online competition to the crowd as an algorithmic challenge they got amazing results in both accuracy and speed. McAfee says the top results, “showed an improvement that was three orders of magnitude faster, without sacrificing accuracy,” compared to the NIH and Harvard Medical School results.

“We’re seeing companies that don’t focus on growing their core. They embrace the crowd from the start,” said McAfee. “We will see how this plays out. But when we rewrite the business playbook, we need to remind ourselves: the crowd is surprisingly wise.”

Industries: From industry to platform

“I grew up in McKinsey understanding the playbook rule: There is no substitute for knowing an industry inside and out. For the past 30 years, the business playbook has said industry structure determines successful business models,” said McAfee.

But in three very different industries McAfee argues that platform is making the difference when it comes to disruptive innovation.

Take the smart phone industry: The defining moment was when Apple opened up the App Store as a platform for outside developers. For urban transportation, it was Uber and now group fitness is being transformed with ClassPass, a platform that allows people to take classes at gyms by subscribing as members to ClassPass, not the gym.

McAfee explains: “ClassPass says, ‘Don’t join a gym. Sign up with us. You can pick whatever classes you want and get variety.’ To gyms they say, ‘you have some empty spaces. We can fill them. You won’t get the full price but some revenue is better than none.’”

Like with Apple and Uber, the platform for ClassPass brings together products, services, sellers, and consumers.

If platforms work, McAfee believes there are many advantages: You get the network effects of increased demand, companies can control the rules of engagement. With an open platform, you can crowd-source innovation and get additional information, which is used to create better pricing and matching of services.

This blows apart the distinct industry-sector differences people used to assume fueled growth and replaces it with the mandate to find the right platform for your business.

McAfee concludes, “I am pretty confident that the successful businesses of tomorrow are going to have a lot more machines, platforms, and crowds in them than today. I am really confident that following the industrial-age business playbook is a really good recipe for failure.”

This article originally appeared on SAP News Center.

Comments

Robin Meyerhoff

About Robin Meyerhoff

Robin Meyerhoff has been at SAP for over 10 years. She started in product communications and has covered a variety of technologies and topics: analytics, mobile, databases, SAP HANA, cloud, sustainability and corporate social responsibility. She currently works within Global Corporate Affairs as a writer on the Content Team and focused on innovation.

Bringing Ethics To Artificial Intelligence With Crowdsourced Innovation

Jochen Schneider

The ongoing debate over the value of artificial intelligence (AI) swings between two extremes. There’s a utopian vision, where AI results in limitless prosperity and health, the end of poverty and hunger, elimination of crime, human immortality, colonization of the galaxy, and the Omega Point at which people become all-knowing and ever-present. But eventually, a skeptical vision emerges as people anticipate a grim future of human enslavement into mass poverty and endless despair, terminating centuries of scientific, technological, and social progress.

As the benevolent potential of AI brightens, the shadows of unethical possibilities become longer and darker. However, digging deeper into this black-and-white dialogue can help companies blur the lines between the two extremes to deliver significant AI-driven advantages for everyone.

Blending the extremes of AI helps create a future of more human-centered experiences

An excellent way to assess AI is to look through the lens of the end user and customer. Think about it: Would you want to enter a store without a single retail associate? Sure, it’s quick and efficient to have a robot retrieve the item, make algorithm-driven product suggestions, and complete the transaction. But this kind of shopping experience eliminates the personal interaction that often motivates buyers to visit a store in the first place instead of shopping online.  As shoppers peruse the store floor, talk to an associate about product options, and get honest opinions, they often feel more comfortable about the purchasing decisions they’re making. So, human emotion rather than fact becomes the more significant influence.

The same degree of reflection is needed when considering AI for any aspect of the business. While the technology can optimize processes to make work faster and simpler, decision makers often miss the opportunity to create more human-centered experiences, which later increase employee and customer engagement, unify people in a collaborative environment, and build a culture of trust. It’s not enough to adopt technology as a replacement for human tasks; businesses need to think about the customer experience and use existing capabilities and resources to interact more closely with every customer.

Companies can engage in this line of thinking by leveraging a platform for digital innovation. Executives, stakeholders, and technology experts can work together to seamlessly integrate AI capabilities into the business network, now and in the future, to proactively respond to the behaviors and needs of customers, employees, and suppliers. Through open dialogue and multi-perspective thinking, the team can uncover new value by creating improved processes, truly unique shopping experiences, industry-disruptive business models, or entirely new companies.

As knowledge about AI technology grows, this unified approach can help companies become more efficient and engaging without losing the human touch that people demand. The platform can also safeguard the privacy of employee and customer information by providing smart apps and processes that comply with proper business conduct policies and regulations.

In the near future, the government will need to pass legislation to help prevent harmful and immoral outcomes that may become possible with AI. However, companies can’t stand still and hope for the best until then. AI technology and market dynamics will undoubtedly continue to change. And for businesses, this means they must evolve with those shifts – responsibly, ethically, and sustainably – to establish a relationship of trust and loyalty with their potential and existing employees and customers.

If we want to retain humanity’s value in an increasingly automated world, we need to start recognizing and nurturing skills that are uniquely human. Learn about Human Skills for the Digital Future.

Comments

Jochen Schneider

About Jochen Schneider

Jochen is chief digital officer at SAP Innovative Business Solutions in EMEA/ MEE. He dedicates his time to innovation, the next generation’s demand and influence. First and foremost, he processes knowledge on how to make innovation real for customers. Second, he cares about how young professionals can prepare for business and influence the future of work.
He studies modern leadership, entrepreneurship, and exponential growth drivers.
In his present role, he is responsible for devising new go-to-market strategies for innovative custom-tailored solutions. His focus is on solutions that leverage innovation technology, foster new business processes, or business models, consequently growing customer value.

Follow him on LinkedIn: https://www.linkedin.com/in/sjochen/

Human Skills for the Digital Future

Dan Wellers and Kai Goerlich

Technology Evolves.
So Must We.


Technology replacing human effort is as old as the first stone axe, and so is the disruption it creates.
Thanks to deep learning and other advances in AI, machine learning is catching up to the human mind faster than expected.
How do we maintain our value in a world in which AI can perform many high-value tasks?


Uniquely Human Abilities

AI is excellent at automating routine knowledge work and generating new insights from existing data — but humans know what they don’t know.

We’re driven to explore, try new and risky things, and make a difference.
 
 
 
We deduce the existence of information we don’t yet know about.
 
 
 
We imagine radical new business models, products, and opportunities.
 
 
 
We have creativity, imagination, humor, ethics, persistence, and critical thinking.


There’s Nothing Soft About “Soft Skills”

To stay ahead of AI in an increasingly automated world, we need to start cultivating our most human abilities on a societal level. There’s nothing soft about these skills, and we can’t afford to leave them to chance.

We must revamp how and what we teach to nurture the critical skills of passion, curiosity, imagination, creativity, critical thinking, and persistence. In the era of AI, no one will be able to thrive without these abilities, and most people will need help acquiring and improving them.

Anything artificial intelligence does has to fit into a human-centered value system that takes our unique abilities into account. While we help AI get more powerful, we need to get better at being human.


Download the executive brief Human Skills for the Digital Future.


Read the full article The Human Factor in an AI Future.


Comments

Dan Wellers

About Dan Wellers

Dan Wellers is founder and leader of Digital Futures at SAP, a strategic insights and thought leadership discipline that explores how digital technologies drive exponential change in business and society.

Kai Goerlich

About Kai Goerlich

Kai Goerlich is the Chief Futurist at SAP Innovation Center network His specialties include Competitive Intelligence, Market Intelligence, Corporate Foresight, Trends, Futuring and ideation.

Share your thoughts with Kai on Twitter @KaiGoe.heif Futu

Tags:

How Manufacturers Can Kick-Start The Internet Of Things In 2018

Tanja Rueckert

Part 1 of the “Manufacturing Value from IoT” series

IoT is one of the most dynamic and exciting markets I am involved with at SAP. The possibilities are endless, and that is perhaps where the challenges start. I’ll be sharing a series of blogs based on research into knowledge and use of IoT in manufacturing.

Most manufacturing leaders think that the IoT is the next big thing, alongside analytics, machine learning, and artificial intelligence. They see these technologies dramatically impacting their businesses and business in general over the next five years. Researchers see big things ahead as well; they forecast that IoT products and investments will total hundreds of billions – or even trillions – of dollars in coming decades.

They’re all wrong.

The IoT is THE Big Thing right now – if you know where to look.

Nearly a third (31%) of production processes and equipment and non-production processes and equipment (30%) already incorporate smart device/embedded intelligence. Similar percentages of manufacturers have a company strategy implemented or in place to apply IoT technologies to their processes (34%) or to embed IoT technologies into products (32%).

opportunities to leverage IoTSource:Catch Up with IoT Leaders,” SAP, 2017.

The best process opportunities to leverage the IoT include document management (e.g. real-time updates of process information); shipping and warehousing (e.g. tracking incoming and outgoing goods); and assembly and packaging (e.g. production monitoring). More could be done, but figuring out where and how to implement the IoT is an obstacle for many leaders. Some 44 percent of companies have trouble identifying IoT opportunities and benefits for either internal processes or IoT-enabled products.

Why so much difficulty in figuring out where to use the IoT in processes?

  • No two industries use the IoT in the same way. An energy company might leverage asset-management data to reduce costs; an e-commerce manufacturer might focus on metrics for customer fulfillment; a fabricator’s use of IoT technologies may be driven by a need to meet exacting product variances.
  • Even in the same industry, individual firms will apply and profit from the IoT in unique ways. In some plants and processes, management is intent on getting the most out of fully depreciated equipment. Unfortunately, older equipment usually lacks state-of-the-art controls and sensors. The IoT may be in place somewhere within those facilities, but it’s unlikely to touch legacy processes until new machinery arrive. 

Where could your company leverage the IoT today? Think strategically, operationally, and financially to prioritize opportunities:

  • Can senior leadership and plant management use real-time process data to improve daily decision-making and operations planning? Do they have the skills and tools (e.g., business analytics) to leverage IoT data?
  • Which troublesome processes in the plant or front office erode profits? With real-time data pushed out by the IoT, which could be improved?
  • Of the processes that could be improved, which include equipment that can – in the near-term – accommodate embedded intelligence, and then communicate with plant and enterprise networks?

Answer those questions, and you’ve got an instant list of how and where to profit from the IoT – today.

Stay tuned for more information on how IoT is developing and to learn what it takes to be a manufacturing IoT innovator. In the meantime, download the report “Catch Up with IoT Leaders.”

Comments

Tanja Rueckert

About Tanja Rueckert

Tanja Rueckert is President of the Internet of Things and Digital Supply Chain Business Unit at SAP.