|April 7 · Issue #79 · View online |
Hiya and welcome to another week in deep learning,
As always, happy reading and hacking!
| Apple Hires Google’s A.I. Chief |
Apple has hired Google’s chief of search and artificial intelligence, John Giannandrea in a bid to catch up to the artificial intelligence technology of its rivals.
| Why 2018 Will be the Year Apps Go to the Edge |
An interesting piece on how machine learning and deep learning moves to the edge, running on devices and sensors with minimum latency and enabling more frictionless AI use cases.
| Stanford DAWN Deep Learning Benchmark (DAWNBench) |
Stanford releases DAWNBench, a benchmark suite for end-to-end deep learning training and inference. It provides a reference set of common deep learning workloads for quantifying training time, training cost, inference latency, and inference cost across different optimization strategies, model architectures, software frameworks, clouds, and hardware.
| Deep Learning Studio is Better More Powerful Than Ever (sponsored) |
Open, Free and No-coding deep learning platform got an upgrade. Developers, researchers and students love our platform. Register today (It’s Free)!
| Eric Jang: Aesthetically Pleasing Learning Rates |
Seeing as the best learning rate exhibit the same characteristics as the pixel intensities in paintings, namely temporal correlation and waving up and down, this posts asks the “why not use a learning rate schedule which is at least aesthetically pleasing?”
| Learning to Navigate in Cities Without a Map |
In this the authors present an interactive navigation environment that uses first-person perspective photographs from Google Street View and gamify that environment to train an AI.
| A Birds-eye View of Optimization Algorithms |
A great overview of different optimization algorithm with visualizations from distill.pub, the post also includes examples in scipy
| Highlights from the TensorFlow Developer Summit, 2018 |
Some highlights from the second TensorFlow Developer Summit including interesting use cases ranging from astrophysics, medicine analyzing auditory data in the rainforest to detect logging trucks.
| NVIDIA/nvvl: Hardware Acceleration to Load Sequences of Video Frames |
A promising library from NVIDIA to load random sequences of video frames from compressed video files to facilitate machine learning training.
| Data Scientist |
Deep NLP and Vision for Business at Infinia ML.
| Data Scientist Intern |
Deep Learning for Business at Infinia ML.
| MobileNetV2: The Next Generation of On-Device Computer Vision Networks |
Google introduces MobileNetV2 powering the next generation of mobile vision applications. MobileNetV2 is a significant improvement over MobileNetV1 and pushes the state of the art for mobile visual recognition including classification, object detection and semantic segmentation.
| Deep Extreme Cut |
A great paper which explores the use of extreme points in an object (left-most, right-most, top, bottom pixels) as input to obtain precise object segmentation for images and videos. We do so by adding an extra channel to the image in the input of a convolutional neural network (CNN), which contains a Gaussian centered in each of the extreme points. The CNN learns to transform this information into a segmentation of an object that matches those extreme points.
| Training Tips for the Transformer Model |
Martin Popel has been the most active non-Googler on the Tensor2Tensor repository for months and has posted a series of very interesting experiments about training and convergence in issue comments. This paper summarizes those and can be incredibly helpful.