The document discusses how machine learning has the potential to deliver new value in three ways: by automating manual human tasks, augmenting human capability, and making the impossible possible. It provides examples of companies using technology to automate insurance claims processing, create scent sensors by combining biology and technology, and develop a next generation interface that silently captures commands. The document also examines emerging digital assistant platforms and how technology is transforming various aspects of daily life such as commuting, food, and exercise.
6. Create entirely new jobs, not
possible before.
Support decision making, and
enable humans to accomplish a
broader range of tasks.
Automate manual human
tasks
Augment human
capability
Make the impossible
possible
Intelligent machines can focus
on repetitive tasks, at wide
scales.
01 / 02 / 03 /
Machine learning is powerful. It has the
opportunity to deliver new value to our world in
three ways.
Opportunities
7. Nexar records and
shares driving incidents
with insurance providers
for faster claims.
01/ Automate Manual Tasks
20. Scale of Trust
“Amazon has a Net Promoter Score of 69,
which is more than 30% above its industry
average.”
980 B 1 T
21. The Most Important Skill:
How to Learn (and Relearn)
“The illiterate of the 21st Century
will not be those who cannot
read and write, but those who
cannot learn, unlearn, and
relearn.”
— Alvin Toffler
22.
23. The Automation Jobless
“The number of jobs lost to more efficient
machines is only part of the problem…
automation may prevent the economy
from creating enough new jobs…
automation is beginning to move in and
eliminate service and office jobs too.”
24. The Automation Jobless
“The number of jobs lost to more efficient
machines is only part of the problem…
automation may prevent the economy
from creating enough new jobs…
automation is beginning to move in and
eliminate service and office jobs too.”
February 24, 1961
Notes de l'éditeur
Improvements in computing power have largely kept pace with Moore’s Law. After 4 decades of exponential increases, the world is now doubling an immense amount of processing power in every two-year period, which is leading to astonishing leaps forward in technological capabilities.
As technology becomes cheaper, world demand is being met at lower price points and fueling an explosion of devices with evermore connections. Sophisticated artificial intelligence devices are now mass-market and better known as personal assistants by the names of Amazon Alexa, Apple Siri, and Google Assistant.
The combinatorial effects of these technologies – mobile, cloud, artificial intelligence, sensors and analytics among others – are accelerating progress exponentially (see Figure1). Once we overcome physical and chemical limitations that are inhibiting exponential gains in mass-market technologies such as battery storage and wireless charging, it is likely that the pace of change will accelerate even faster.
Key Message
It’s a common misconception to think about Artificial Intelligence and relate it to human-like robots. While that might be one outcome, really what we mean to say is Machine Learning.
Talk Track
Thinking of AI as human-intelligent robots is common misconception. Artificial Intelligence (AI) is the ability for any system or machine to replicate human capabilities. Today, when we talk about Artificial Intelligence, we really mean Machine Learning.
So what is Machine Learning?
https://medium.com/netflix-techblog/artwork-personalization-c589f074ad76
https://research.netflix.com/research-area/recommendations
https://blogs.nvidia.com/blog/2018/06/01/how-netflix-uses-ai/
https://store.steampowered.com/labs/
Netflix, something that started as a simple predictive suggestion AI has now fundamentally changed - sometimes the best suggestion algorithm isn’t the goal. Netflix has a 100 person team working on that 90 sec window… the time they’ve got to get you interested. They run 300k product “tests” as their AI has moved from predictive to prescriptive - trying to train you how to become a better Netflix user. We all see something different because we’re different. Netflix is constantly iterating to keep their $14 billion on top.
[Note // Each header (e.g. 01/ Automate) is a link to its section. Click ahead if the conversation wants to focus on a single benefit]
Key Message
Machine Learning provides three near-in opportunities for all of us. It can:
Automate manual human tasks
Augment human capability
Making the impossible
Talk Track
We’re seeing three distinct, human benefits in how Machine Learning can support us today and in the future.
Automate manual human tasks
Machines can begin to take over repetitive tasks, or tasks that aren’t currently scalable by human effort alone.
Augment human capability
Having assessed large data sets, and having immediate recall, we can leverage Machine Learning to support decision making, and identify specific business cases to consider or prioritize.
Making the impossible, possible
Consider the implications flight travel had on the tourism industry. Whole economies grew from our ability to travel globally. Consider the new business models tomorrow’s interfaces (e.g. virtual reality), enabled by Machine Learning, will open up for us.
Next, we’ll highlight current Machine Learning solutions and key examples around each of these benefits.
Talk Track
Nexar uses Machine Learning to be your second set of eyes on the road. With an understanding of vehicle movement, road conditions and driver habits— Nexar detects critical driving moments that occur in-transit, records them and provides detailed incident reports should an accident occur. These reports can be shared directly with an auto-insurer, including a pre-trimmed video of the incident— significantly speeding up the claims and/or litigation process.
Talk Track
Koniku (https://koniku.com) mixes biology, silicone and machine learning— combining them into sensors that can detect particular types of smells. Keeping in mind these are things humans are physically incapable of (some dogs can though). Scent sensors can be used during airport security to make a transactional-less process. Each scent sensor can seek a variety of factors, from threats to Destination Custom’s requirements.
// Note: Emphasize how it replacing/scaling dogs in security
Talk Track
MIT project AlterEgo captures “voice” before a user speaks it, and is able to send that command to digital assistants. Could this be how people communicate with machines in the future, or even with each other?
She wakes you up early! (prescriptive) She's learned you sleep in, got IBM Watson Weather info, pulled in current traffic, etc.
She informs you it is time to eat your next digestible sensor.
She wakes you up early! (prescriptive) She's learned you sleep in, got IBM Watson Weather info, pulled in current traffic, etc.
She informs you it is time to eat your next digestible sensor.
Guided Grocery Experience
Digital health assistant, personalized food advertisement, “you’re low on energy, grab the fruit”, partnership with food/store providers with UHC.
Girl Scouts Attack!
Connected Doctor
Doctor reviews your health behind a remote sensing dashboard of the future and encourages you as you head back to the office for the day.
VR Shared Experiences
You finish the day off in a new virtual reality gym game, burning off some cookies!
It used to be Uber which scared companies, now it’s Amazon.
https://www.retently.com/blog/companies-high-nps/