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
What does it take to run machine learning at scale? Despite the advancements in machine and deep learning, most of the toolset and infrastructure to run intelligent programs at scale remains relatively basic. Furthermore, as an industry, I don’t believe we have enough experience to yet understand the best practice of running large scale machine learning workflows.
The challenges of building state-of-the-art machine learning infrastructures is rapidly becoming one of the main roadblocks for the mainstream adoption of these technologies. Fortunately, companies such as Netflix, Facebook, Microsoft, Google or Uber that are at the forefront of large-scale machine learning implementations have been actively open sourcing some of the frameworks and architecture guidance they follow to achieve scale and efficiency in machine learning solutions. Just this week, we saw new releases in this area from engineering teams at Facebook and Netflix. The challenge remains to incorporate those architecture best practices and frameworks into mainstream machine learning platforms.
Now let’s take a look at the core developments in AI research and technology this week:
Facebook published a new paper using a technique called Zero-Shot Learning for identifying images associated with large volumes of text.
OpenAI published a new paper that proposes an energy-based method for identifying concepts in 2D and 3D environments.
AI researchers from IBM just proposed a new method to detect phonetic similarities in Chinese natural language phrases.
Cool Tech Releases
AI Powerhouse OpenAI released Spinning up in Deep RL, an educational resource to master reinforcement learning.
Facebook open sourced FBGEMM to run server-side machine learning models efficiently at scale.
Netflix recently shared some insights about the infrastructure the media giant uses to run machine learning workflows at scale across dozens of teams.
AI in the Real World
AI News Anchor: This week China’s news agency Xinhua the first AI version of some of their popular news anchors.
AI for Mental Health: IBM Research has been working on a serious of AI models to improve language pattern recognition to help clinicians diagnose and treat mental health disorder.
General AI in the US Military: The US Air Force discussed some of their work laying out the foundation for general AI.