While the hype around neural networks may have some ground truth to it, they are not the answer to every single problem.

Here’s a great look at where they may fall short.

Unlike other algorithms, neural networks with their deep learning cannot be programmed directly for the task. Just like a new developing brain, they have the requirement to learn the information. The major advantage of neural networks is its ability to outpace almost every other machine learning algorithm.

NVIDIA CEO Jensen Huang addresses 1,400+ attendees of SC19, the annual supercomputing conference. in Denver

He introduced a reference design for building GPU-accelerated Arm servers, announced the world’s largest GPU-accelerated cloud-based supercomputer on Microsoft Azure, and unveiled NVIDIA Magnum IO storage software to eliminate data transfer bottlenecks for AI, data science, and HPC workloads.

As anyone who has a smart phone can tell you, battery technology has lagged far behind every other type of innovation.

The race is on to build cheaper, longer-lasting, more energy-dense batteries.

One of the most promising technologies in this space is the solid state battery, developed by an absolute legend in the battery world, one of the inventors of the lithium ion battery and recent Nobel Prize winner John B. Goodenough.

Joe Scott discusses in depth.

This episode of the AI show provides a quick overview of new batch inference capability that allows Azure Machine Learning users to get inferences on large scale datasets in a secure, scalable, performant and cost-effective way by fully leveraging the power of cloud.

Learn More:

Batch Inference Documentation

https://aka.ms/batch-inference-documentation

Batch Inference Notebooks

https://aka.ms/batch-inference-notebooks

Tony Gambacorta shows you how to explore your hardware with a serial port.

If you’re interested in hardware but haven’t had a chance to play with any yet, this one’s for you. In this “hello world”-level reversing project we’re checking out a UART (serial port) and using it to access a shell on a *very* soft target. If you decide to try it on your own you’ll find an equipment list, walkthrough references, and some troubleshooting ideas at the link below.

TensorFlow Lite is a framework for running lightweight machine learning models, and it’s perfect for low-power devices like the Raspberry Pi.

This video shows how to set up TensorFlow Lite on the Raspberry Pi for running object detection models to locate and identify objects in real-time webcam feeds, videos, or images. 

Written version of this guide: https://github.com/EdjeElectronics/TensorFlow-Lite-Object-Detection-on-Android-and-Raspberry-Pi/blob/master/Raspberry_Pi_Guide.md

Microsoft Research just posted this video on adversarial machine learning.

As ML is being used for increasingly security sensitive applications and is trained in increasingly unreliable data, the ability for learning algorithms to tolerate worst-case noise has become more and more important.

The reliability of machine learning systems in the presence of adversarial noise has become a major field of study in recent years.

In this talk, I’ll survey a number of recent results in this area, both theoretical and more applied. We will survey recent advances in robust statistics, data poisoning, and adversarial examples for neural networks. The overarching goal is to give provably robust algorithms for these problems, which still perform well in practice.

Talk slides: https://www.microsoft.com/en-us/research/uploads/prod/2019/11/Adversarial-Machine-Learning-SLIDES.pdf