Microsoft Research Connections Blog
Next at Microsoft
Social Media Collective
Windows on Theory
Posted by Rob Knies
The launch of Windows 8 late last year provided developers with new opportunities to construct paradigm-shifting apps that can stand out in a busy application ecosystem via their ability to capitalize on touchscreen technologies.Such an evolution doesn’t come along too often in the software industry, and developers have responded in a big way. In its first two months, the Windows Store, which came online at the same time Windows 8 was offered for general availability, enticed visitors to download more than 100 million apps. The Windows Phone Store has surpassed 1 billion downloads. And computing usage of Windows Azure has doubled.This new direction for developing and delivering great ideas will gain even more momentum at San Francisco’s Moscone Center from June 26 to 28 when Microsoft hosts Build 2013, a chance for software engineers to witness presentations from the developers who produce the company’s products and services.
Sometimes, it seems like we’re awash in video choices: broadcast, cable, satellite, Internet, PC, tablet, smartphone. It can seem overwhelming.Sometimes—and stop me if you’ve heard this one before—it seems like, with all these choices, none of them is offering anything particularly compelling.
Computing today is generating and capturing a wealth of data previously unimaginable. Such information has great promise for unlocking some of society’s most elusive secrets, but how can those secrets be unearthed and identified?That pursuit provided the impetus behind Big Data Analytics 2013, a first-ever workshop held at Microsoft Research Cambridge on May 23-24. More than 130 participants from academia and industry—including a strong contingent from the hosting lab, Microsoft Research Redmond, Microsoft Research Silicon Valley, and Advanced Technology Labs Europe—gathered to discuss and identify the most important and challenging directions for the evolution of algorithms and systems for big data.“The organization of the workshop was prompted by a surge of interest and activity in the area of big-data analytics,” says Milan Vojnovic, co-organizer of the event and senior researcher in the Cambridge Systems and Networking group, “including platforms for various kinds of processing, such as batch processing and querying of massive data sets, real-time analytics, streaming computations, and analytics on special data structures such as graphical data.
Moshe Tennenholtz is an accomplished man. An Israel-based principal researcher with Microsoft Research New England, he has performed pioneering work bridging computer science, artificial intelligence, and game theory. He also has co-founded several e-commerce companies. Given such a varied, successful background, there’s little these days that can faze him.Yet when he learned he had been named winner of the 2012 Allen Newell Award from the Association for Computing Machinery (ACM) and the Association for the Advancement of Artificial Intelligence, he couldn’t have been more surprised.“It was announced to me by phone by the chair of the committee [Eric Grimson, chancellor of the Massachusetts Institute of Technology],” Tennenholtz says. “I didn’t even know what he wanted to talk to me about.”
We live in a society obsessed with speed. Whether it’s download times on a mobile phone or Usain Bolt’s time in the 100 meters, the faster the better. We also live during an era when accuracy has become not just preferable but essential. The technological marvels of the 21st century demand it.Speed=good. Accuracy=good. Put them together, and you’ve got a leap forward, such as recent advancements in Bing Voice Search for Windows Phone that enable customers to get faster, more accurate results than ever before.Those improvements come, in part, from contributions delivered via Microsoft Research’s work on deep neural networks (DNNs). Such networks are a computational framework for automatic pattern recognition that is inspired by the basic circuits of the human brain. Refinements in mathematical formulas, coupled with greater computational power and large data sets, enable DNNs to learn and edge noticeably closer than traditional speech technologies to humans’ ability to recognize speech and images.