9. We Need To Talk About Facebook
US$528.45b Market Cap
US$482.92 GDP
http://statisticstimes.com/economy/countries-by-projected-gdp.php
US$28b (2016) revenue
US$27.174 GDP
41.49m pop
20,658 employees
37.95m pop1.2b users
10. We Need To Talk About Google
US$725.34b Market Cap
US$93b (2016) revenue (3 x Facebook)
1.2b users
11. Google Page Ranking
A B C D
0.25 0.25 0.25 0.25
Probability distribution for the likelihood that a person randomly clicking on links will arrive at any particular page.
A
B
C
D
0.75
𝑃𝑅(𝐴) = 𝑃𝑅(𝐵) + 𝑃𝑅(𝐶) + 𝑃𝑅(𝐷)
15. Label Features Vectors
Line
1
1 0 1 1
Line
2
0 1 0 1
Line
3
0 0 0 1
Line
4
1 1 1 1
Sum of vectors 4
If vectors = 4, then image = face
• “Labels” and
“features”
• ’Award points’
• Set a rule
Machine Pattern Recognition
17. 86 x 109
(billion)
neurons
You have the
potential for
more thoughts
than there are
atoms in the
Universe - of
which there
are ~1080
x 24
potential
connections
= 213!
http://www.madsci.org/posts/archives/1999-01/915377761.Ns.r.html
Neural Network
19. Applied Neural NetworkInput Layer
Algorithms 1 (Hidden Layer)
Algorithm 2 (Output Layer)
“A”
“E”
“I”
“O”
“U”
Error Correction (‘Back Propagation’ loop)
Difference
between the
required output
and the machine
generated output
Short vowel sounds
“I”
After many
iterations
in training
Data Algorithm Error Correction Prediction
“I”
“E”
24. A.I. – The Next 10 Years
A.I. skills crunch worsens
Rapid automation of A.I. processes
Explainable AI (XAI) Enters the Mainstream
Augmented Intelligence – enhance rather than replace people
Startups will overtake Amazon, Google, IBM and Microsoft (Gartner)
China will take the lead in A.I. (and may have done so already)
Voice will become the main A.I. UX
Biology-based design paradigm (e.g. the ‘Enterprise Immune System’)
Now
Future
26. Skills For A.I.
Use A.I. as a
savvy consumer
ProduceWithA.I.ConsumeA.I.
Build A.I.
Use A.I. to
augment work
Have A.I.
“done to you”
A.I. not included in the curriculum
Self-learning
A.I. included in curriculum
Training Self-learning
PhD-level
problem solving
Manage A.I. solutions
HE and/or
Specialist
Experience
31. THURSDAY April 12TH 2018
A.I. FOR ABSOLUTE BEGINNERS
Become aware of what A.I. is
and what it can do for you.
Understand the difference
between A.I., Machine
Learning, and Big Data.
Learn what questions you need
to be asking and what skills are
needed to get in front of the A.I.
wave.
A.I. Demystified‘A.I. Demystified’ is designed to quickly give you the tools you need to understand
what A.I. is, how it works, and how you can put it to use.
http://bit.ly/2ofquAA
Bristol
Lets start unpicking A.I. by building a metaphor for it.
Imagine that you are driving a very smart combine harvester.
As it cuts through the crop it applies A.I. – or Machine Learning to be more precise – capabilities.
The first thing it does is classifies the crop in terms of whether its is wheat or chaff. The wheat is separated into a tank and the chaff is blown out of the back of the combine harvester.
Next, it looks for anomalies – are there other species of plants in the crop?
Then it analyses the size of the grains, and clusters them into groups – something that could be useful when combined with geographic and soil analysis data.
Finally, it can forecast ahead. It can use regression analysis to predict the grain sizes and the overall yield.
Driving the Combine Harvester is a diesel engine which in turn drives an electrical generator, which powers its sensors, computing and communication capabilities.
We can think of A.I. as the Combine Harvester and the crop operating in the field, and Machine Learning as the engine that drives the ’intelligence’.
The point of A.I. is to make predictions.
In the case of a combine harvester the outcomes could be a fully automated machine, real-time information to the markets, and information to form the basis of future crop planning.
Our use of AI is vast.
AI is marching into our lives, affecting how we live, work and entertain ourselves. From voice-powered personal assistants like Siri and Alexa, to more iunderlying and fundamental technologies such as behavioral algorithms, suggestive searches and autonomously-powered self-driving vehicles boasting powerful predictive capabilities, there are many examples of artificial intellgence in use today.
A true artificially-intelligent system is one that can learn on its own. True A.I. can improve on past iterations, getting smarter and more aware, allowing it to enhance its capabilities and its knowledge.
Some respected and prominent scientists and inventors worry about the ‘Singularity’ - the idea that the invention of artificial superintelligence will abruptly trigger runaway technological growth, resulting in unfathomable changes to human civilization. This, in effect, is when AI will write itself independently of human input.
Thankfully we are a long way from this eventuality, but to prevent it from happening we need people to be aware of how AI works and how to control it.
The goal of this course is to do just that – give you the tools and understanding necessary for you to take control of AI.
At secondary level, the range of Maths that can be covered in AI projects just explodes.
A.I. can provide motivating reasons for tackling difficult concepts by grounding them in a practical context.
Take a look at these zeros and ones.
It is a binary representation of something, but what?
We first sort the data out into rows
We then call row (line) data ‘labels’ and the zeros and ones ‘features’
Next, we add a ‘vector’ column to ’award points’ for rows that match the shape you are looking for
Then we set an algorithm for classifying the whole image based on the sum of the vectors
17
18
In this greatly simplified model we can visualise error correction in phone-based voice recognition.
Error correction will be used extensively in the learning process prior to the software being put on the phone, to train the software to correctly interpret different languages, accents and dialects.
When you get the phone, you’ll load the appropriate algorithms for your particular language and locale, but the software will still have some learning to do in order to fully adjust to your unique voice.
The more you talk into your voice recognition system, the more error correction it will go through until eventually it will work as smoothly as possible.
So, most important question that we must ask is –
Are we in danger of having A.I. done to us by elites – big business, government, and academia?
Take the case of Cambridge Analytica an A.I. company linked to Brexit, and Trump campaigns, and more recently the elections in Kenya
A.I. also raises questions about:
Privacy
Taxation
Objective truth
Accountability
What are the implications for the increasing convergence between biology and technology?
https://www.theguardian.com/technology/2017/may/07/the-great-british-brexit-robbery-hijacked-democracy#img-5
“Predictions are difficult to make, especially when they are about the future” but…
Many believe that China, due to the amount of data available through its huge population, is already taking a lead in AI.
Rapid Automation of A.I. data processes - E.g. AutoML, Driverless AI
In the West, states and companies will pull in opposite directions in terms of open algorithms and data privacy. No such concerns in China, many Asian states, and Russia.
EU GDPR requires firms to ‘show their workings’
Companies want a ‘black box approach’ to drive down costs of AI. Western democracies want openness and transparency.
Shorter term predictions –
By 2019, more than 10% of IT hires in customer service will mostly write scripts for bot interactions.
Through 2020, organizations using cognitive ergonomics and system design in new artificial intelligence projects will achieve long-term success four times more often than others.
By 2020, 20% of companies will dedicate workers to monitor and guide neural networks.
By 2019, startups will overtake Amazon, Google, IBM and Microsoft in driving the artificial intelligence economy with disruptive business solutions.
By 2019, artificial intelligence platform services will cannibalize revenues for 30% of market-leading companies.
Gartners
Longer-term implications for system design:
Biology will increasingly guide system design (Bio inspired computing)
E.g. https://www.darktrace.com Cyber security modelled on the immune system
Other examples - Autodesk Dreamcatcher (AI driven CAD), Neural Networks, Genetic Algorithms, phototropic solar arrays, smart swarms.
With self-organising Neural Networks, and Goal Driven Design, new development paradigms are likely to emerge
Key Question - what in nature must do what x system does to survive?
Links
http://www.cognilytica.com
https://www.theverge.com/2017/8/3/16007736/china-us-ai-artificial-intelligence
https://www.uscc.gov/sites/default/files/Research/DGI_China%27s%20Industrial%20and%20Military%20Robotics%20Development.pdf
https://www.gartner.com/imagesrv/media-products/pdf/rage_frameworks/rage-frameworks-1-34JHQ0K.pdf
http://www.zdnet.com/article/machine-learning-may-supercharge-enterprise-architecture/
https://www.gartner.com/binaries/content/assets/events/keywords/symposium/esc28/esc28_digitalbusiness.pdf
‘AI Demystified’ – is a one-day workshop - is designed to quickly give you the tools you need to understand how AI works and how you can put it to use.
With no prerequisites – apart from curiosity and basic Excel skills – professionals from all walks of life will get the tools they need to make sense of AI. Throughout the workshop, the focus will be on the practical applications of AI.
AI can seem inaccessible and complex at first, but this workshop is designed to quickly give you the tools you need to understand how AI works and how you can use it.