Scenarioplanning - social futures

AI periodicering

The Four Waves of A.I. - Kai-Fu Lee 2018
... What changed everything a decade or so ago was an approach called “deep learning” ...
The first stage is “Internet A.I.” Powered by data flowing through the web ...
The second wave is “business A.I.”  ... 
The third wave  ... call it “perception A.I.” ... 
The fourth wave is ... “autonomous A.I.” ... 



Get Ready for the Insurtech Revolution, Act 2
The second act, which Schreiber calls “Deep Seeing,” will revolutionize core insurance products and pricing. He said the second act will have a profound impact as it deploys artificial intelligence (AI) and machine learning along with predictive modeling and nontraditional data sets.
“The next insurance leaders will use bots, not brokers, and AI, not actuaries,” he proclaimed to the audience at the ITC.
... offer a product such as a will or, perhaps in the future, a genetic test, for free as a way of getting to the insurance sale. He also sees non-insurance firms entering.
Daniel Glaser, CEO of insurance broker and risk consulting giant Marsh McLennan, also has a view on the next phase of insurtech. But his view includes brokers. He swatted down the Lemonade leader’s notion that bots will replace brokers. “That’s just nonsense,” he declared. “I’m on the other side of that bet.” ... He predicts companies will at first go about creating proprietary AI and machine learning applications but that those will not be sustainable any more than proprietary customer relations management systems were. ... 2018

Artificial intelligence: Transforming the insurance industry ... ... 

Deep Learning frameworks

Deep Learning frameworks  
Google vs Apple vs Amazon/Facebook/Microsoft/ONNX  

Battle of the Deep Learning frameworks — Part I: 2017 
... At the moment, TensorFlow by Google seems to be the most used deep learning framework. ... 

Open Neural Network Exchange (ONNX) 
Amazon Web Services, Facebook and Microsoft

Facebook  and Microsoft announced ONNX
The Exchange makes it easier for machine learning developers to convert models between PyTorch and Caffe2.
... Apple’s CoreML for example helps developers convert a very limited number of models. At this point CoreML doesn’t even support TensorFlow. ... 

Amazon joins Facebook and Microsoft in support of open-source AI platform ... 

Business and AI

Lasse Rouhiainen  
Keynote speaker on emerging technologies, social media and video marketing. ...
Artificial Intelligence: 101 Things You Must Know Today About Our Future ... 

Artificial Intelligence and Machine Learning for Business: A No-Nonsense Guide to Data Driven Technologies
Steven Finlay. Amazon Media EU 2017.   
... The focus is very much on practical application, and how to work with technical specialists (data scientists) to maximize the benefits of these technologies. ... 

Microsoft AI

Microsoft Research Labs 
Microsoft Creates New AI Lab to Take on Google's DeepMind ...
Eric Horvitz, director of Microsoft Research Labs 

Videos about AI

The incredible inventions of intuitive AI 
Maurice Conti - TEDTalk 2017
AI: From passive AI to Generative AI ... 

AI on Track to Achieving Superintelligence?  
World Economic Forum: Center for the Fourth Industrial Revolution. 
Professor Rita Singh shares how AI may shape our understanding of the past. ... ... 

Next Generation Robots  
Boston Dynamics, Asimo, Da Vinci, SoFi ... 

Risk society - AI

Future of Life Institute  
Works to mitigate existential risk from advanced artificial intelligence (AI).
Its founders include MIT cosmologist Max Tegmark.

Tegmark, Max 2017: Life 3.0: Being Human in the Age of Artificial Intelligence 
Chapter 3: The near future. ... 
... what can be done to maximize the chances of a positive outcome ...
Tegmark’s solutions to inevitable mass unemployment are a stretch ... ... 

MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) ... 

DataRobot, Inc.

Become an AI-Driven Enterprise with Automated Machine Learning
DataRobot enables users to build and deploy machine learning models

AI subjects

AI Chipsets  
...  Enter a suite of new processors found on an
SoC—“system on a chip.” Huawei, Apple, Alphabet,
IBM, NVIDIA, Intel and Qualcomm are all working
new systems architecture and SoCs,

Cognitive Computing
Cognitive computing systems use natural language processing and artificial intelligence in order to understand our intentions. 

Nine big companies dominate AI 
Alphabet, Amazon, Microsoft, IBM, Facebook and Apple in the US, along with Chinese behemoths Tencent, Baidu and Alibaba.
On the investment side, Intel Capital, Google
Ventures, GE Ventures, Samsung Ventures, Tencent
and In-Q-Tel lead.

Ambient Interfaces - Zero-UI
E.g.: Amazon Echo,  Microsoft Kinect, ... ... ... andy-goodman

Future Today Institute


Nick Bostrom’s SuperIntelligence

Review v/Andy Hines
A healthy fear of superintelligence 

AI developmental pathway:

1st, task intelligence,
in which AI outperforms people in a single task. We are currently here.
2nd, general intelligence,
in which AI is as intelligent as people in a wide range of tasks.
This may be decades away ... or closer.
3rd, superintelligence,
in which AI is orders of magnitude more intelligent than people across-the-board

Decades away, right? Probably, but Bostrom posits the interesting scenario that the takeoff from general to superintelligence could happen very fast, on the order of days or weeks. That would mean essentially no time to prepare a world in which humans are no longer the smartest species on the planet.

Strawberry Fields scenario
Bostrom mentions a treacherous turn in which AI realizes at some  points that humans are trying to limit it, and begins to conceal its behavior. And he talks about “perverse instantiation,” which basically includes the Strawberry Fields scenario that Musk has popularized.

Andy Hines

What Elon Musk Really Thinks of Artificial Intelligence -- and Why You Should Care  
... 2. Musk's fears of A.I. created tension with his friend, Google co-founder Larry Page. Musk targeted Google as the company most likely to let its A.I. get out of control. ... Many argue that Musk is creating a good-vs-evil storyline. By portraying his own companies as being on the good side, he'll have an easier time attracting talent for cheap. ...

... Alphabet, could have perfectly good intentions but still “produce something evil by accident”—including, possibly, “a fleet of artificial intelligence-enhanced robots capable of destroying mankind.”
Strawberry fields scenario  
Even robots with a seemingly benign task could indifferently harm us. “Let’s say you create a self-improving A.I. to pick strawberries,” Musk said, “and it gets better and better at picking strawberries and picks more and more and it is self-improving, so all it really wants to do is pick strawberries. So then it would have all the world be strawberry fields. Strawberry fields forever.” No room for human beings. ... 

Artificial Intelligence

Explainable Artificial Intelligence (XAI) - Mr. David Gunning 
New machine-learning systems will have the ability to explain their rationale, characterize their strengths and weaknesses, and convey an understanding of how they will behave in the future.

AI Needs New Clichés - Molly Wright Steenson
The old ones aren’t helping, and the new ones are old. ... 

AI in manufacturing

This $2 Billion AI Startup Aims to Teach Factory Robots to Think  
Deep learning in manufacturing ... ... 

We apply machine learning and deep learning to robotics and machine tools
Preferred Networks released open source deep learning framework Chainer v4

AI development

DevOps tooling - data science toolkits  

AutoML project is an approach made in order to automate the design of machine learning models. James Kobielus 2018. ... 

Featuretools - Python library for automated feature engineering
An open source framework.

Big data's biggest secret: Hyperparameter tuning  
... Open source projects such as the Fregata library are tackling the problem of tuning and executing regression models at massive scale ... 

Synthetic (aka artificial) training data, will become the lifeblood of most AI projects. Solution providers will roll out sophisticated tools for creation of synthetic training data and the labels and annotations needed to use it for supervised learning.

The surge in robotics projects and autonomous edge analytics will spur solution providers to add strong reinforcement learning to their AI training suites.
AI solution providers will add collaborative learning to their neural-net training tools.

It will also be useful in for optimizing distributed AI architectures such as generative adversarial networks (GANs) in the IoT,

Transfer learning - knowledge transfer - inductive transfer - meta learning
Machine learning, once implemented, tends to be specific to the data and requirements of the task at hand. Transfer learning is the act of abstracting and reusing those smarts ... 

James Kobielus 2018 ... 
James Kobielus

Decision intelligence engineering

Why businesses fail at machine learning  
Decision intelligence engineering  
...  the applied side is a very different discipline from the algorithms research side ...  you need an interdisciplinary team. ... ...

Open Source Machine Learning Libraries  
Choosing an Open Source Machine Learning Library: TensorFlow, Theano, Torch, scikit-learn, Caffe ... 

Top 16 Open Source Deep Learning Libraries and Platforms  
By all measures, TensorFlow is the undisputed leader.
Keras, Caffe, Microsoft Cognitive Toolkit, and PyTorch completing the top five. ... 

Google Cloud

Machine learning research
Stanford Vision and learning Lab 
Computer vision and human vision

The blockchain-enabled intelligent IoT economy 
The convergence of AI, blockchain, and IoT ... 

Facebook AI