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: Uber’s Open Source Machine Learning Stack
Welcome to our first newsletter of 2020! This time we start where we left off in 2019: talking about machine learning technologies in the real world.
Artificial intelligence(AI) is a trend that differs from many of its predecessor technology movements in the sense that a substantial percentage of the innovation is not coming from startups but from the big corporate labs. Companies like Microsoft, Google, Amazon, Facebook, Uber are actively contributing to the rapid evolution of AI research and technology. Among those technology giants, Uber has quietly become one of the most active contributors to open source machine learning technologies of the last few years.
A few days ago, Uber announced the latest addition to its open source machine learning stack. Manifold is a toolset for debugging and interpreting machine learning models at scale. The release of Manifold joins a deep stack of open source machine learning technologies that include the following:
· Ludwig :A TensorFlow based toolbox that allows to train and test deep learning models without the need to write code.
· Pyro : A deep probabilistic programming language
· Plato: A Framework for Building Conversational Agents at Scale
· Horovod: A framework for parallelizing the training of deep learning models
Uber’s contributions to the open source machine learning community are helping to advance the space and provide data scientists with tools and frameworks that have been tested at Uber’s scale.
Now let’s take a look at the core developments in AI research and technology this week:
Dopamine as an Inspiration for Reinforcement Learning
DeepMind started the year publishing a groundbreaking paper that proposes a new reinforcement learning algorithm inspired by the role of dopamine in the human brain.
Solving Advanced Mathematical Equations using Deep Neural Networks
Facebook AI Research(FAIR) published a paper describing a model that can solve complex math equations using symbolic reasoning.
Quantifying Model Uncertainty
Google AI researchers published a paper that benchmark the uncertainty of state-of-the-art deep learning models as they are exposed to new data distributions.
Cool AI Tech Releases
Facebook released a new version of the PyTorch deep learning framework that includes improvements in mobile, an excitement new mode for parallel training, Java bindings and several other cool features.
AWS Gets Into AutoML with AuGluon
AutoGluon is a new framework that allows developers to author deep learning models with a few lines of code.
Manifold Allows You to Debug Models Visually
Uber open sourced Manifold, a tool for visually testing and debugging machine learning models.
AI in the Real World
AI Startups Raised $18.5 billion in 2019
A new report by the National Venture Capital Association claims that 1,356 AI-related companies in the U.S. raised $18.457 billion in venture capital last year.
Apple Buys AI Startup for $200M
Apple has acquired Xnor.ai, a startup that has powering machine learning models in edge devices.
AI in Hollywood
Hollywood studios are relying more and more on AI models to optimize the chances of success of new films.