The Intuition Machine Letter —2nd Edition
Welcome to the bi-weekly letter covering Deep Learning Patterns, Methodology and Strategy. We’ve come up away to organize the topics to appeal to the broadest of audiences. The more general topics always are at the top, while more specialized ones are towards the bottom. We hope that this newsletter we appeal to all those interested in Deep Learning developments.
For this issue, we continue to look at various ethical viewpoints regarding the development of AI and its impact on our future. We also look at how the development of AI parallels development in psychology and the different approaches one can use in creating AI. We also explore new tools and courses that may help one learn about AI development, and dive deeper into the latest researches and developments on AI.
Ethics
Six Months Later, France has Formulated their Deep Learning Strategy
Six months ago, I wrote that “The West is Unaware of the Deep Learning Sputnik moment”. It turns out mathematician Cedric Villani began a 6 month journey to learn all that needs to be learned about Deep Learning.
Artificial Personhood is the Root Cause Why A.I. is Dangerous to Society
I have now come to the realization that human compatible AI is a problem that is inextricably intertwined with human civilization. It cannot be solved because present human civilization isn’t structured in a manner that is aligned with the needs of humanity. You cannot achieve human beneficial A.I. without drastically remaking human civilization.
AI’s Symbol Grounding Problem and the Future of Work
Here’s the kicker though, there are plenty of work that people are simultaneously something people love to do, are good at and the world needs.
Growing a Machine Learning Practice Inside a Large Financial Firm
I recently had the pleasure of speaking with Zachary Hanif, Director in the Center for Machine Learning (C4ML) at Capital One.
How artificial intelligence will transform how we gesture
Smartphones are new tools for body rehabilitation, sports training and motor development.
Is it Time to Panic about American Ignorance of Deep Learning?
Eric Schmidt in a recent Verge report has a few remarks about the current state of AI policy in the United States:
Tim Berners-Lee: we must regulate tech firms to prevent ‘weaponised’ web | Technology
The inventor of the world wide web warns over concentration of power among a few companies ‘controlling which ideas are shared’
A history of machine translation from the Cold War to deep learning
I open Google Translate twice as often as Facebook, and the instant translation of the price tags is not a cyberpunk for me anymore. That’s what we call reality.
This Is What Happens When Bitcoin Miners Take Over Your Town — POLITICO Magazine
Eastern Washington had cheap power and tons of space. Then the suitcases of cash started arriving.
Malta Blockchain Regulation Proposal: Legal Personality for DAOs and Smart Contracts — ChainStrategies
Technology Arrangements On March 2, 2018, the Blockchain Malta Association hosted an event entitled “Reflections on the DLT Regulation.” A number of lawyers and industry experts were invited to debate the consultation document issued by the Government of Malta and wherein the country’s intention to regulate the sector was presented. The consultation document proposed three…
Stephen Fry describing our future with artificial intelligence and robots — YouTube
In a beautiful monologue of 6 minutes Stephen Fry takes us all the way back to ancient Greek Mythology (Prometheus, Pandora) and then via the world wide web …
Opinion | The Tyranny of Convenience — The New York Times
All the personal tasks in our lives are being made easier. But at what cost?
IOHK | Charles Hoskinson at LSE, Cardano’s goals for Africa. — YouTube
Cardano’s goals for Africa and the developing world. Charles Hoskinson addresses packed crowd at LSE. The London School of Economics has a long history of li…
In Search of the Distinctively Human | The Philosophy of Blade Runner 2049
AI and Psychology
How knowledge about different cultures is shaking the foundations of psychology
Psychological phenomena have long been thought of as universal. But it turns out scientists may have been blinded by their own culture.
media.ccc.de — Computational Meta-Psychology
Computational theories of the mind seem to be ideally suited to explain rationality. But how can computations be subverted by meaning, em…
AI Design and Development
The Machine Learning Reproducibility Crisis « Pete Warden’s blog
Gosper Glider Gun I was recently chatting to a friend whose startup’s machine learning models were so disorganized it was causing serious problems as his team tried to build on each other’s work and share it with clients. Even the original author sometimes couldn’t train the same model and get similar results! He was hoping…
Clues for Which I Search and Choose — arg min blog
An outsider tour of reinforcement learning, Part 9. Simple methods solve apparently complex RL benchmarks
Building Literate Machines with Dr. Adam Trischler — Microsoft Research
Try Microsoft Edge
A fast and secure browser that’s designed for Windows 10
Smerity.com: Backing off towards simplicity — why baselines need more love
When we lose accurate baselines, we lose our ability to accurately measure our progress over time.
Microsoft: FPGA Wins Versus Google TPUs For AI
Microsoft recently disclosed Project Brainwave, which uses pools of FPGA’s for real-time machine-learning inference, marking the first time the company has shared architecture and performance details of their FPGA-based chips. I take an in depth look at the implications for Intel and Xilinx.
AI in storytelling: Machines as cocreators | McKinsey & Company
Computers don’t cry during sad stories, but they can tell when we will.
Game Theory | Worlds Revealed: Geography & Maps at The Library Of Congress
Worlds Revealed: Geography and Maps at The Library Of Congress. A blog from the Library of Congress.
Tools and Courses
IDEA — nonverbal algorithm assembly instructions
IDEA is a series of nonverbal algorithm assembly instructions, created by Sándor P. Fekete, Sebastian Morr, and Sebastian Stiller.
MIT 6.S099: Artificial General Intelligence
An introduction to deep learning through the applied task of building a self-driving car. Taught by Lex Fridman.
Style Transfer as a Service· Rodrigo Castro
Picmatix.com is an online tool that creates digital art by transferring styles to user pictures. This blog post outlines the serverless architecture that powers the website.
Mini World of Bits benchmark
Each environment is an HTML page that is 210 pixels high, 160px wide (i.e. identical to ATARI ALE simulator dimensions). The top 50 pixels (with a yellow background) contain the task query — a description of what the agent should do in the environment.
Invited talk:Deep Reinforcement Learning with Subgoals (David Silver) on Vimeo
David Silver
Capsule net with GRU | Kaggle
Research
Frontiers | Illusory Motion Reproduced by Deep Neural Networks Trained for Prediction | Psychology
The cerebral cortex predicts visual motion to adapt human behaviour to surrounding objects moving in real time. Although the underlying mechanisms are still unknown, predictive coding is one of the leading theories. Predictive coding assumes that the brain’s internal models (which are acquired through learning) predict the visual world at all times and that errors between the prediction and the actual sensory input further refine the internal models.
Understanding Deep Learning through Neuron Deletion | DeepMind
Deep neural networks are composed of many individual neurons, which combine in complex and counterintuitive ways to solve challenging tasks, ranging from machine translation to Go. This complexity grants neural networks their power but also earns them their reputation as confusing and opaque black boxes. Understanding how deep neural networks function is critical for explaining their decisions and enabling us to build more powerful systems.
Transfer Your Font Style with GANs — The Berkeley Artificial Intelligence Research Blog
The BAIR Blog
Research Blog: Semantic Image Segmentation with DeepLab in Tensorflow
- DeepLab-v3+ is not used to power Pixel 2’s portrait mode or real time video segmentation. These are mentioned in the post as examples of features this type of technology can enable.↩
Research Blog: Using Evolutionary AutoML to Discover Neural Network Architectures
What is wrong with VAEs?
Latent Variable ModelsSuppose you would like to model the world in terms of the probability distribution over its possible states with .The world may be com…
Journal Article
[1802.08864v1] One Big Net For Everything
Book
Artificial Intuition : The Improbable Deep Learning Revolution
I challenge you to find a field as interesting and exciting as Deep Learning. This book is a spin-off from my previous book “The Deep Learning AI Playbook”. The Playbook was meant for a professional audience. This is targeted to a much wider audience. There are two kinds of audiences, those looking to explore and those looking to optimize. There are two ways to learn, learning by exploration and learning by exploitation. This book is about exploration into the emerging field of Deep Learning.
The Deep Learning AI Playbook
“Whatever you are studying right now if you are not getting up to speed on deep learning, neural networks, etc., you lose,” says Mark Cuban.
“We are going through the process where software will automate software, automation will automate automation.”