In the News
Google infuses AI everywhere
Google announced several products that rely heavily on AI at its recent I/O event:
- Google Home – a direct competitor to Amazon Echo
- Google Assistant – a direct competitor to Siri
- Allo and Duo – direct competitors to Whatsapp or Skype with built-in chatbots
AI will create 'useless class' of human, predicts bestselling historian
Smarter artificial intelligence is one of 21st century’s most dire threats, writes Yuval Noah Harari in follow-up to Sapiens
More news
- Are IBM's claims about Watson and AI fraudulent? More
- Google developed its own chips (ASICS) to run deep learning algorithms faster than with GPUs. They claim their advance is equivalent to a 7-year advance on the Moore power law. This is huge, and they are keeping their designs private. More
- Amazon is building autonomous drones. More
- AI is now on the White House's agenda. More
Featured
Moderate Images with Deep-Learning
Moderate and filter images with a simple API. Determine if images contain nudity, how many people or faces they contain, if they are natural photographs or not etc.
Learning
Engineering Intelligence Through Data Visualization at Uber
Interesting piece by the Uber engineering team on the way they empower people internally through their data visualization team.
Facebook's AI backbone: Flow
Building AI is hard—so Facebook is building AI that builds AI. Here are some details they shared.
Residual neural networks are an exciting area of deep learning research
The power of residual neural networks have become apparent with Microsoft's first place in the last Imagenet competition. This article highlights several recent papers that show the potential of such networks
Neural Networks are impressively good at compression
By learning "smart" representations of data, images or sounds, neural networks could be an interesting approach to compression.
Software tools & code
OpenAI Gym Beta
OpenAI Gym is a toolkit for developing and comparing reinforcement learning (RL) algorithms. It consists of a growing suite of environments (from simulated robots to Atari games), and a site for comparing and reproducing results.
Google Parsey McParseface
Google released SyntaxNet, an open-source neural network framework for TensorFlow that provides a foundation for Natural Language Understanding systems. The release comes with Parsey McParseface, an English parser trained by Google.
Some thoughts
If Google’s right about AI, that’s a problem for Apple
Today, Apple’s being led properly day-to-day and doing very well overall. But if the landscape shifts to prioritize big-data AI services, Apple will find itself in a similar position as BlackBerry did almost a decade ago
About
This newsletter is a weekly collection of AI news and resources. If you find it worthwhile, please forward to your friends and colleagues, or share on your favorite network!
Share on Twitter · Share on Linkedin · Share on Google+
Suggestions or comments are more than welcome, just reply to this email.
Thanks!