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: Can AI be Decentralized?
Decentralized artificial intelligence(AI) is one of the emerging trends in the AI ecosystem. The traditional centralized model of AI solutions have steadily contributed to increase the gap between large companies and startups in the space. In the current ecosystem, its very difficult for a new company to compete with the likes of Google, Microsoft or Facebook in terms of AI talent, investments or access to large datasets. Similarly, its unclear how developing nations will be able to compete with top economies for the adoption of AI technologies.
Decentralized AI architecture have the potential to democratize access to knowledge and talent. While the decentralized AI space has been booming with innovation, the practical implementations and adoption remain very limited. This week, we published a detailed analysis about the path of adoption of decentralized AI platforms.
Now let’s take a look at the core developments in AI research and technology this week:
AI Research
Facebook published a summary of reinforcement learning method used to teach robots to learn.
>Read more in this blog post from the Facebook AI Research team
Google AI researchers published a paper about how to use deep learning for generating accurate 3D representations of moving objects in an scene.
>Read more in this blog post from the Google AI team
Baidu AI researchers published a paper proposing a technique for parallelizing text to speech models.
>Read more in this blog post from the Baidu AI Research team
Cool AI Tech Releases
Facebook open sourced Pythia, a PyTorch-based framework for multitasking vision and language tasks.
>Read more in this blog post from the Facebook engineering team
The PyTorch team also discussed torchvision 0.3, a new library for building computer vision models.
>Read more in this blog post from the PyTorch team
TensorWatch is a cool new library for visualizing deep learning models.
AI in the Real World
The US Army and North Carolina State University have been working on a new framework for continual learning and improving the memory of neural networks.
>Read more in this blog post from the US Army
Google has been working with several medical center developing AI that can help to better detect lung cancer.
>Read more in this article from the New York Times
Samsung developed deepfake AI, a system that can generate small video clips from a single photo. The Mona Lisa example is a big freaky.
