Every week, Invector Labs publishes a newsletter that covers 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: PyTorch 1.5
PyTorch has established itself as one of the top two frameworks for deep learning developments. Initially incubated by Facebook, PyTorch rapidly developed a reputation from being an incredibly flexible framework for rapid experimentation and prototyping gaining thousands of fans within the deep learning community. For instance, AI powerhouse OpenAI announced that it was standardizing on PyTorch as the default framework to power its deep learning research work. Outside Facebook, this is arguably the biggest endorsement for PyTorch within the deep learning world.
This week Facebook announced the released of PyTorch 1.5. The new version focuses on providing tools and frameworks to make PyTorch workflows production-ready. The most notable aspect of this release has been the collaboration between AWS and Facebook in two projects: TorchServe for model serving and Torch-Elastic Kubernetes for distributed training. This release contributes greatly to make PyTorch a more viable option for many enterprises starting in their machine learning journey.
Now let’s take a look at the core developments in AI research and technology this week:
DeepMind published an analysis of how reinforcement learning agents can game their target tasks producing undesired outcomes.
Evolutionary Meta Learning
Google AI researchers published a paper proposing a new meta-learning method based on evolutionary strategies.
Facebook AI researchers published a paper introducing Quant-Noise, a new method for compressing neural networks without affecting their performance.
Cool AI Tech Releases
Facebook released PyTorch 1.5 which includes several projects like TorchServe and TorchElastic that are based on the collaboration between Facebook and AWS.
Google open sourced the TensorFlow profiler, a new set of tools that you can use to measure the training performance and resource consumption of TensorFlow models.
Apache SINGA 3.0
Apache SINGA is one of the most popular deep learning projects incubated in Asia. The distributed deep learning library just released version 3.0.
AI in the Real World
Researchers from MIT unveiled a method that can reduce the carbon footprint used to train and operate AI models.
CometML’s New Funding Round
CometML is one of our favorite platforms for machine learning management and they just raised a new round of funding.
Peak AI Raises $12 Million
Enterprise AI startup Peak AI announced a new $12 million funding round to help enterprises adopt AI solutions.