Last Week in AI
Every week, my team at Invector Labs publishes a newsletter to track the most recent developments in AI research and technology. You can find this week’s issue below. You can sign up for it below. Please do so, our guys worked really hard on this:
From the Editor
Is decentralized artificial intelligence(AI) possible? Currently, AI is totally based on centralized model controlled by incumbents with access to large volumes of data and expensive data science talent. This is a classic rich get richer model which seems unsustainable in the long run. The emergence of blockchain technologies and advances in cryptographic techniques such as homomorphic encryption has triggered a new generation of platforms that attempt to decentralized AI solutions that move the control from a central authority to a network of participants.
The decentralized AI space has been exploding with new offerings such as Ocean Protocol, OpenMined or SingularityNet based on blockchain technologies. Until now, most of those offers have been constrained to pilots and small sample applications. That picture changed last week when SingularityNet released the first version of their decentralized AI platform on the Ethereum blockchain. While SingularityNet has achieved notoriety for powering AI systems like Sophia the Robot, the general availability of their platform represents a complete different label. Now developers can write new models that run on SingularityNet’s platform without relying on a centralized authority.
Now let’s take a look at the core developments in AI research and technology this week:
AI researchers from IBM published a paper proposing a new method for retrieving images based on natural language.
Google AI researchers demonstrate how to use reinforcement learning for robot navigation.
Microsoft Research published a method to high performance AI algorithms for the classification of medical images.
Cool Tech Releases
Decentralized platform SingularityNet reach general availability on the Ethereum mainnet.
Uber engineers detail the data pipeline architecture used to power many of the machine learning workflows at Uber.
AI in the Real World
Encounter AI built Mai, a digital assistant to help the visually impaired to order fast food.
The New York times debates the challenges and realities of ethics in AI
China is aggressively competing with The United States for supremacy in the AI market.