IoT weekly round-up: Thursday 1st December 2016
It’s been a musical sort of week in the connected world, as computers get to know Beethoven through annotated scores and a neural karaoke program generates music by analyzing a digital image. In other news, spectators at this weekend’s Big Tem Championship game will get a new perspective thanks to GoPro, and a sensor-embedded suit may help monitor stroke patients in recovery. Read on for the latest news.
IoT Community (IoTC) announces IBM as Diamond Corporate Member
The world’s largest group of CxOs and IoT professionals, the IoTC, has announced that IBM will be a headline partner for its IoT Grand Slam virtual conference on 1st December. At the event, Chris O’Connor, General Manager of IoT Offerings at IBM, will present a keynote on how recent IoT advancements drive industries forward, focusing on real-world use cases. IBM Fellow John Cohn will also be participating in a cognitive panel.
Amazon launches Amazon AI platform
At this week’s re:Invent developer event in Las Vegas, Amazon announced the launch of its new AI platform, bringing its machine learning smarts to developers outside the company. Three services make up the initial offering: ‘Rekognition’, which allows you to recognize specific objects and scenes, ‘Amazon Polly’, a text-to-speech service to produce life-like speech across 24 languages, and ‘Lex’, which allows construction of conversational apps able to feature multi-step conversations.
Referees to wear GoPro hats during Big Ten Championship Game
Referees will wear hats mounted with GoPros during the NCAA Football Big Ten Championship Game this weekend between Penn State and Wisconsin. Footage captured from the hats will give a unique viewing experience – as spectators can see the game from the Referee’s point of view – which will be shown online and on FOX during the broadcast. The project is brought to life as part of a deal between FOX Sports Lab and GoPro.
Machine learning algorithms take on Beethoven with MusicNet
MusicNet is a new project attempting to use machine learning to introduce classical music to computers. MusicNet comprises over 300 live recordings of classical music, with an emphasis on the works of Beethoven, Schubert, Brahms, Mozart and Bach. Each live recording is matched to a score – effectively annotating it – with each note indicated exactly as played, thanks to a technique called dynamic time warping which maps the written score to the exact nuances of a live interpretation.
Neural karaoke program brings us the AI-generated Christmas song
A computer program by the name of ‘Neural karaoke’ has generated and performed a Christmas-themed song after analyzing the visual components of a digital image. The program can analyse any digital photo and interpret the data to create a computer-generated song. Before you get too excited – the musical end-product in this instance is bizarre to say the least, and certainly won’t be topping any charts – but it does serve for a light-hearted demonstration of the potential of AI.
Sensor suit helps gauge stroke patient recovery
A high-tech suit embedded with sensors may provide a new way of keeping tabs on stroke patients in recovery. The suit represents a major advance in rehabilitative care and is designed by Bart Klaassen of the University of Twente in the Netherlands. Sensors within the suit monitor patients as they perform tasks in their own homes, such as getting in and out of bed, making a cup of tea or moving about the house.