|September 4 · Issue #55 · View online |
Hey and welcome to a new issue,
Happy hacking and reading. As always if you enjoy receiving this newsletter you can help us by sharing it with friends and colleagues.
See you later this week!
| AI Chip Boom: This Stealthy AI Hardware Startup Is Worth Almost A Billion |
An in depth look at Cerebras Systems, a new startup focused on hardware for deep learning, that has managed to quietly raise more than 100 million dollars. The AI chip race is currently led by Nvidia, but we’re excited to see new solutions enter the market.
| My Year at Brain |
Colin Raffel shares the experiences he made during his one year as a Google Brain resident. He covers his research topics and some details on how such a residency is managed.
| iOS 11: Are Apple’s new NLP capabilities game changers? |
This blog post checks out if Apple’s announcements regarding their new NLP capabilities hold up.
| The new rules of Build vs. Buy in an AI-first world |
An interesting take from Michael Vaccarino of Clarifai on when to run your own AI and when to rely on external services (like Clarifai). Implementing your own system seems like more fun, but in the end someone has to pay the bills, so make sure the math works out!
| My Neural Network isn't working! What should I do? |
A really great collection of common mistakes made during neural network design and implementation. If something doesn’t work as expected, just work through this list and you’ll hopefully find the issue.
| Deep Learning (DLSS) and Reinforcement Learning (RLSS) Summer School |
The videos from this years Deep and Reinforcement Learning Summer Schools are up and provide a great resource to revisit specific topics in depth or just browse through the available material.
| Transformer: A Novel Neural Network Architecture for Language Understanding |
Google presented Transfomer, a new recurrent neural network architecture based on self-attention, that seems to be well suited for language processing.
| How to Train a Simple Audio Recognition Network |
A new tutorial on audio recognition with Tensorflow was added to the libraries repository and looks like a great entry point if you want to get into audio recognition. It covers audio recognition basics and some more specific knowledge and the code is of course included.
| Launching the Speech Commands Dataset |
A new dataset was made available by Google. This collection of spoken commands is supposed to be a learning example for basic experiments and consists of 60.000 1-second-long snippets that contain one of 30 commands.
| Collection of generative models in Tensorflow |
Tensorflow implementations of various GANs and VAEs to use at your will. There is a Pytorch version
available as well.
| dynet/crayon integration |
This tutorial shows you how to show summaries from any neural network framework in TensorFlows Tensorboard using the crayon library
. This example uses the Dynamic Neural Network Toolkit
, but you can adjust it to match your needs.
| A TensorFlow implementation of Deep Convolutional Generative Adversarial Networks |
| Deep Learning for Video Game Playing |
A paper that reviews recent deep learning advances in the context of how they have been applied to play different types of video games such as first-person shooters, arcade games or real-time strategy games.
| Automatic Semantic Style Transfer using Deep Convolutional Neural Networks and Soft Masks |
Another approach to style transfer, this time using soft masks to further improve the generated images, which leads to some amazing results, especially when used for rather constrained scenarios.
| Explainable Artificial Intelligence: Understanding, Visualizing and Interpreting Deep Learning Models |
This paper summarizes recent developments in the field of understanding and visualizing neural networks and makes a plea for more interpretability in artificial intelligence.
| Design and Analysis of the NIPS 2016 Review Process |
This paper explains, how NIPS handled the immense rise of submissions, reviewers, and attendees in last years review process.