Share on Facebook Tweet on Twitter Share on LinkedIn Share by email
Featured Research Videos

Several years ago, the biotech company Novartis came to Microsoft because they had seen the Kinect system for Xbox game consoles and thought the body movement sensors might be helpful for tracking the progression of multiple sclerosis. That, in turn, could help them more consistently judge whether a treatment they were proposing was working. Microsoft researchers were intrigued, and called upon their experts in machine learning to create algorithms that could more accurately measure the subtle changes in movement that can occur as MS progresses. The Assess MS project also brought in Microsoft’s experts in human-computer interaction, who were able to help design a specialized system that would work in a real-world hospital setting, where medical professionals with no specialized technical skills would be asked to operate the machine in cramped exam rooms. The collaboration between Microsoft Research’s Cambridge, UK, lab and Novartis holds the promise of helping doctors more quickly bring better treatments to patients who are suffering from MS, by offering more consistent analysis of whether the treatments are helping patients.




Introducing Microsoft Project Natick, a Microsoft research project to manufacture and operate an underwater datacenter. The initial experimental prototype vessel, christened the Leona Philpot after a popular Xbox game character, was operated on the seafloor approximately one kilometer off the Pacific coast of the United States from August to November of 2015. Project Natick reflects Microsoft’s ongoing quest for cloud datacenter solutions that offer rapid provisioning, lower costs, high responsiveness, and are more environmentally sustainable.



Tutorial: Deep Learning
Tutorial: Deep Learning
02:29:10

Deep Learning allows computational models composed of multiple processing layers to learn representations of data with multiple levels of abstraction. These methods have dramatically improved the state-of-the-art in speech recognition, visual object recognition, object detection, and many other domains such as drug discovery and genomics. Deep learning discovers intricate structure in large datasets by using the back-propagation algorithm to indicate how a machine should change its internal parameters that are used to compute the representation in each layer from the representation in the previous layer. Deep convolutional nets have brought about dramatic improvements in processing images, video, speech and audio, while recurrent nets have shone on sequential data such as text and speech. Representation learning is a set of methods that allows a machine to be fed with raw data and to automatically discover the representations needed for detection or classification. Deep learning methods are representation learning methods with multiple levels of representation, obtained by composing simple but non-linear modules that each transform the representation at one level (starting with the raw input) into a representation at a higher, slightly more abstract level. This tutorial will introduce the fundamentals of deep learning, discuss applications, and close with challenges ahead.


MobileFusion: Create 3D scans with your mobile phone
MobileFusion: Create 3D scans with your mobile phone
00:02:16

MobileFusion is a research project that turns ordinary mobile phones into 3D scanners without any additional hardware. The resulting 3D scans are detailed enough for 3D printing or use in augmented-reality games.

Photo Story automatically generates stylized story compilations for easy sharing
Photo Story automatically generates stylized story compilations for easy sharing
00:01:26

Photo Story, a new Windows Phone app, automatically chooses the best representative photos of an event and organizes them into a themed video, complete with stylized music and editing, that you can easily share via email or social media.

Improving the Halo 5 multiplayer experience
Improving the Halo 5 multiplayer experience
00:02:48

Senior researcher Rob DeLine describes how the Halo team uses Trill, a high-performance in-memory incremental analytics engine, and Tempe, a web service for exploratory data analysis, to monitor and quickly improve the Xbox gaming experience.

All hands, no keyboard: New Handpose technology can track detailed hand motion
All hands, no keyboard: New Handpose technology can track detailed hand motion
00:00:53

Researchers at Microsoft have developed a system that can track -- in real time -- all the sophisticated and nuanced hand motions that people make in their everyday lives. The Handpose system could eventually be used by everyone from law enforcement officials directing robots into dangerous situations to office workers who want to sort through e-mail or read documents with a few flips of the wrist instead of taps on a keyboard.

Skype Translator



We recently previewed Skype Translator to two elementary school classes -- one in Washington and one in Mexico City. A few rounds of "Mystery Classroom" was all it took for these students to discover the potential of Translator to break down language barriers and bring people together.



Krysta Svore on quantum computing and machine learning
Krysta Svore on quantum computing and machine learning
00:01:38

Senior Researcher Krysta Svore comments on the potential effect quantum computing can have on machine learning.