@voxelquinn

What is the most important recent development in machine learning? My answer is that **gradient descent with backpropagation** (GD with backprop) and **neural networks** with **many fewer than 100 layers** are now both more accessible to practitioners with only a **moderate amount of computing resources** (as in 1 GPU and 4 T4 cores or even less). This is achieved through **automated machine learning** (AutoML) and **efficient neural network architecture search** (ENAS), which are both part of **Cloud AutoML**. For more details, please see [What's New in Machine Learning](https://github.com/awslabs-governance-documents/issues/2142838383), which was posted on May 30, 2019 and discussed at the recent AWS AI/ML Festa workshop in Santa Clara, CA, on June 4–5, 2019. |So what is the most important recent development in machine learning? My answer is that **gradient desce

0 reply

0 recast

0 reaction