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: Machine Learning to Help Programmers

Can machine learning help us to write better code?

That sounds like a scary proposition but is not as crazy as it sounds. Many aspects of programming problems such as bug fixing, code snippet suggestion and testing can be modeled as machine learning problems. Many of the same techniques we use in classification and prediction problems can be applied to code optimization scenarios.

Using machine learning models to improve the quality of code and the lifecycle of software programs is an active area of research. Large software companies like Microsoft, Google or Facebook have been leading the charge in this area and these problems become an order of magnitude more critical given their large developer base. Recently, Facebook unveiled Aroma, a new framework that uses machine learning for code recommendations. The ideas in Aroma are still early but they show us the path to a future in which machine learning is a fundamental component of compilers, debuggers and IDEs in order to help developers write better code.

Now let’s take a look at the core developments in AI research and technology this week:

Research

AI researchers from IBM a paper proposing a new distributed deep learning method for speech modeling tasks.

>Read more in this blog post from IBM Research

The DeepMind team published an incredibly comprehensive analysis about the state of unsupervised learning methods.

>Read more in this blog post from DeepMind

AI powerhouse Prowler published a very interesting analysis of the use of incentives to drive optimization in multi-agent reinforcement learning systems.

>Read more in this blog post from the Prowler team

Cool Tech Releases

Facebook unveils Aroma, a framework that uses Aroma to make recommendations in programming code.

>Read more in this blog post from the FAIR team

AI Researchers from Microsoft and Cornell University combined the AirSim drone simulator and reinforcement learning techniques to build autonomous robots that can navigate tunnels.

>Read more in this blog post from Microsoft Research

Facebook open sources Habana backend for the Glow machine learning hardware platform.

>Read more in this blog post from the Facebook engineering team

AI in the Real World

Elon Musk discusses the Tesla Autopilot system in the AI Podcast

>Watch the entire podcast here

Forbes published a review of Amy Webb’s recent book The Big Nine that describes the impact that tech giants can have in the race to dominate AI.

>Read the Forbes article here

The Economist published an incredibly comprehensive study about the battle between Google and DeepMind to dominate the future of AI.

>Read the complete article from The Economist here