Emergent // Future: Open Sourced AI, AWS AI Tools, FPGA’s, and more
This week we look into the efforts to open source artificial intelligence by Google DeepMind and OpenAI, how Amazon AWS is re:Inventing its platform for AI with new services and FPGA’s for EC2.
Plus our favorite articles from the past week about AI and some of the projects making us smarter.
Open Sourcing AI ⚡️
You might have heard: Google DeepMind is open sourcing its AI training platform and making the code available on GitHub for researchers and developers to experiment with.
By open sourcing the 3D game-like platform, DeepMind hopes the research community can make use of it and help shape the development of the platform going forward.
But did you know… The move comes after OpenAI open sourced Universe, their platform that features a thousand different environments for training AI to play games and use apps.
Universe lets AI agent “use a computer like a human does: by looking at screen pixels and operating a virtual keyboard and mouse.”
Plus: In keeping with our mission to democratize access to AI, we’ve open sourced the AWS AMI for training your own style transfer models.
re:Inventing AWS ☁️
Amazon AI launched with three tools for developers at re:Invent last week: image recognition, text-to-speech, and a chatbot builder.
The platform is Amazon’s answer to the Google Cloud ML platform, which provides machine learning services with pre-trained models.
Quick overview of the three tools:
- Rekognition — an image recognition service for identifying objects and scenes in photos.
- Polly — a text-to-speech service with 47 different voices in 24 different languages.
- Lex — speech recognition and natural language understanding service for building conversational chatbots, like Alexa.
Amazon also debuted their FPGA instances for EC2, called F1. These reprogrammable chips make it easy to switch context within an application without having to switch the hardware.
ALSO: Amazon and Techstars are partnering on a 13-week conversational AI startup accelerator. They’re offering companies $20K in initial investment and $100K convertible note option upon completion of the program.
What We’re Reading 📚
- “Humanity’s greatest fear is about being irrelevant.”Anthropologist Genevieve Bell explains why being scared about AI has more to do with our fear of each other than killer robots. (The Guardian)
- Why Deep Learning Matters and What’s Next for Artificial Intelligence. Deep learning is ushering in an era that will fundamentally alter the way we live, work, and communicate like the industrial revolution. (Algorithmia)
- Google, Facebook, and Microsoft Are Remaking Themselves Around AI. Artificial intelligence is not only reshaping the technology these tech giants use but how they organize and operate their businesses. (Wired)
- Marc Andreessen on flying and autonomous cars, IoT, AR, VR, and what tech will disrupt next. Whether the subject is job creation, the future of autonomous vehicles, or the remarkable progress in the quest to build flying cars, Marc Andreessen is as optimistic as he’s ever been. (The Verge)
- Alexa, Tell Me Where You’re Going Next. Amazon’s VP of Alexa talks about machine learning, chatbots, and whether industry is strip-mining AI talent from academia. (Backchannel)
Things Making Us Smarter 🛠
- Connecting Torch to StarCraft
- Fast Face-swap Using Convolutional Neural Networks
- High-Resolution Image Inpainting using Multi-Scale Neural Patch Synthesis
- Building a machine learning workstation for under $5000
- Lessons Learned from Deploying Deep Learning at Scale
Emergent Future is a weekly, hand-curated dispatch exploring technology through the lens of artificial intelligence, data science, and the shape of things to come. Subscribe here.
Originally published at blog.algorithmia.com on December 6, 2016.