The Cutting Edge: Week 6

September 03, 2014
by Anthony Tempest

Robo Brain to teach robots

A large-scale computational system that learns from the internet, called Robo Brain, is currently downloading a significant amount of data. This includes approximately 1 billion images, 120,000 YouTube videos, and 100 million how-to documents and appliance manuals. This project aims to provide information that is stored in a robot friendly manner for robots to access as and when they need to. The overall goal is quite a futuristic thought. Once robots become mainstay in our homes, offices, and factories, they’ll need to understand how everything works around them. The current method teaches robots one thing at a time, which is time consuming. This new initiative aims to provide the learning capability in one package with Robo Brain. Read more about it here.

Cancer-hunting ‘nano-robots’

Soon, an army of tiny robots could be deployed in order to hunt down and destroy cancer from within. Promising progress has been made with a nanoparticle called “nanoporphyrin”, which can help diagnose and treat cancers. Cancer has surpassed cardiovascular diseases to become the leading cause of death in Australia. It’s natural that this statistic will drive scientists even more to diagnose and treat this disease. Nanotechnology works by building up tiny structures at the nano-level, each with their own functions and applications. However, building these nanoparticles takes time and are expensive to research and build. This time and expense will be worth it to develop a safe, efficient tumour treatment though. Want to know more about how these nanoparticles work? Click here! 

Google’s drone delivery program

Google’s drone delivery system has recently completed test flights conducted since mid-August. This marks the culmination of the first phase of Project Wing, a program that has been running for two years at Google X, the company’s long-range research lab. Although there have been many rumours about the project, it has only been revealed very recently. During the initial phase of development, Google settled on an unusual design known as a tail sitter for their drones. This means it takes off vertically, but when it comes to flying about, it does so horizontally. When attempting a delivery, it hovers and winches packages to the ground. A bundle of electronic sensors attached to the end of the tether can detect the package hitting the ground and then detaches itself from the delivery. Once the delivery is complete, the drone flies back to its local depot. Learn more about Google’s drone project here.

Facial analysis spots emotions on Google Glass

German researchers at the Fraunhofer Institute for Integrated Circuits IIS have developed facial recognition technology for Google Glass. The result is that it can recognise emotions, gender, and age. There have been issues regarding privacy with internet connected wearables, but this new technology stops short of recognising somebody’s identity. The software, called Sophisticated High-speed Object Recognition Engine (SHORE), processes video on the Google Glass CPU in real-time. Libraries of data built on C++ analyses the face and produces information on the person, which is superimposed over them. In combination with structure-based features and learning algorithms, it is expected that the software can boast extremely high recognition rates. Read more about the Google Glass project here.