A look into Microsoft Research - Society for Information Management

cabbagecommitteeΤεχνίτη Νοημοσύνη και Ρομποτική

24 Οκτ 2013 (πριν από 3 χρόνια και 7 μήνες)

49 εμφανίσεις

FUTURE TRENDS:

INFORMATION MANAGEMENT

IN 2015

Lewis Shepherd

Chief Technology Officer

Microsoft Institute for Advanced Technology in Governments

www.ShepherdsPi.com


A Move from DIA to Microsoft

America’s Long Tradition of
Government Supporting Research

Simon Cameron,

U.S. Senator from Pennsylvania


A leading voice during 1861 debate over
Smithsonian Institution funding, during
run
-
up to War

“I am tired of all this thing called
science....

We have spent millions in that
sort of thing for the last few
years, and it is time it should be
stopped.”


Later in 1861:

Named by Abraham Lincoln to head the
19
th

Century’s military
-
industrial
complex (Secretary of War)


1866:

Re
-
elected to U.S. Senate


1862:

Ousted for corruption, censured by the
House of Representatives for “contract
manipulations”

America’s Long Tradition of
Government Supporting Research

Why I Joined this Small West
-
Coast Startup

Platform and
Online Services
Division

Entertainment &
Devices Division

Business
Division

Research &
Development

Security

Ai ds


Vaccine

Quantum Computing

& Cryptography

Robotics

Col laboration

R&D Budget 2008: $8 Billion


2009: $9 Billion

MSR Growth


Research lab locations :


Redmond, Washington


(Sep, 1991)


San Francisco, California


(Jun, 1995)


Cambridge, United Kingdom

(July, 1997)


Mountain View, California


(July, 2001)


Bangalore, India



(Jan, 2005)

MSR Cambridge

MSR India

Research PhD’s

“Institute” in Organizational Context

The Breadth of MS Research


Platform Elements


Networking, Distributed systems, Operating
systems


Cellphone

and other Devices


Sensor networks


Security, Protection against Malware,
Identity


Connecting Developer and IT


Languages, tools, compilers


Revolutionizing software and services


Web


Search and Advertising


Knowledge management


Cybersecurity


Data and Documents


Database Architectures, Data Mining


Machine learning, Fighting SPAM


Meta data extraction, authoring



User Interfaces, Social Computing, and
Collaboration


New UI


Speech, Ink, Gesture, Natural
Language, Large Displays, Surface
Computing


Meetings and Collaboration


Modeling of People and Groups


Technologies for Emerging Segments


Media


Graphics and Multimedia


Digital Photography and Video


Science


AIDS Vaccine, Quantum Computing


eScience



Bioinformatics, Astronomy


Algorithms, Cryptography


Economic models


Rapidly Changing Technology

Predicting is hard



Computational power


Multi/Many
-
core CPUs


Graphics 3x per year


Storage 2x per year


Networking 4x per year


New devices


Ubiquitous connectivity


Nano

Technology


The Web


Microsoft Inc. as an Enterprise Example

435 million unique users

6 billion instant messages
(IMs) per day

280 billion page views/day

29 billion emails sent/day

141,000 end users


260,000 computers


550 Buildings in 98 countries


358,000 SharePoint sites


2,500 internal applications


3 million internal emails/day

20 million incoming emails
per day (97% filter)


42,000,000 remote
connections per month


Technology Themes Benefiting from
Increased Speed and Scale


Cloud Computing


Security & Mobile


Human
-
Computer Interaction


Immersive Data


Geospatial /Robotics/Social Networks


Semantic Computing



2015: What will continue to be
important to I.M. leaders?


Right information


Right people


Right time




Right information


Right people


Right time





ALL information



ALL people



ALL the time



All Information



Explosion of data
-
rich Social Media



Semantic techniques & semantic computing



Hybrid machine/human translation


Semantic Computing


Leveraging our positions as world’s largest
hosting company, world’s largest email
provider, world’s largest IM platform


Finding Meaning in our Search



& our Hosted Social Platforms


Microsoft Live Search index =25 TB content,
3500 queries/sec




Mining Social Media: the
Twitterverse

http://www.TweetGrid.com


A robust acquisition and mining platform supporting research
and product exploration in social
-
media analysis.


Platform acquires social
-
media data, such as Blogs, Usenet, and Twitter in real
time or near real time and provides a stream of this content.


Stream is consumed by real
-
time mining components that can be assembled
into compelling desktop applications.


Platform provides a content store that gives access to the textual content of the
media, as well as statistics and other meta
-
data describing the publications and
authors serving and creating the content.


Real
-
time mining application analyzes social media for references to news
articles, allowing ranking of news & opinions as discussed online.

Facilitates discovery, browsing, and sharing of online info,
keeping the user up
-
to
-
date and informed about events &
attention being paid to them


Social Streams: Real
-
Time Social Media Mining

“Political Streams”


OSINT Early Warning System

Online Info/Blogosphere
Viz

& Data Mining

http://socialstreams.livelabs.com/



Research Desktop Activities allows users to add
semantic tags & labels to related documents,
images, e
-
mails, and other items.


Using the semantic labels, users can easily activate
a particular task or switch between


multiple tasks.


Dedicated information spaces:




Personal Library to collect books, manuscripts,
relevant articles and media


Notes to enable simple storage and access to
content snippets, URLs, and other bits of
information that can easily be misplaced or can
be difficult to find.




Tools and services that can be used in various
contexts. Users can easily analyze individual
books or collections of publications, create a
co
-
author network.

à
Easier discover y of tr ends i n data.


Semantically Enabled “Research Desktop”

Research Desktop augments the
standard desktop environment with
concepts and designs that enabl e
new ways of working and managing
resources. RD provides semantic
enabl ing within four key areas:
Acti vi ti es, Tool s, Li brary
and
Notes
.


Creating overlays over the World Wide Web


Storing a location
-
transparent digital
memory that works to personalize,
semantically relate, and socially enhance
experiences of the web


Designed to integrate with traditional
desktop


Will first be packaged and projected via the
browser
--

but that is just a delivery channel


About putting the human back in the center
of the experience of technology


Every aspect of our UI and technology is
subordinated to creating experiences that
enhance human community building and
interaction.


Consumer/Web
-
User Semantics

Research Desktop demo

Translation Research


MSR’s Natural Language Processing group has
developed a hybrid Human/Machine
Translation (MT) system


Has both data
-
driven and rule
-
based components


Learns translation mappings automatically from
bilingual sentence pairs (Microsoft product TMs)


Allows semi
-
automated human
-
in
-
the
-
loop with a wiki



Google’s MT group is using speed of first
-
pass


API allows integration by 3
rd
-
parties

Translation Research


MSR’s Natural Language Processing group has
developed a hybrid Machine Translation (MT)
system


Has both data
-
driven and rule
-
based components


Learns translation mappings automatically from
bilingual sentence pairs (Microsoft product TMs)


The system has been used successfully by the
internal Customer Support group to translate
knowledge
-
base articles


Its use is being extended to localization work for
selected Microsoft products

Google
-
Powered “Nice Translator”

www.NiceTranslator.com

MS
-
Powered Real
-
Time Translation in Browser

All People



Social Networks



Presence



Assured Identity


”Planetary
-
Scale Views on a Large Instant
-
Messaging Network”


Eric Horvitz


One month of traffic on MS Messenger
(May 2007)


Dataset

contained “summary properties” of
30 billion conversations

among 240 million
people


The communication graph constructed
includes 180 million nodes, 1.3 billion
undirected edges

Human Terrain Analysis:

“Largest Social Network Ever Analyzed”

Visualizing the Human Terrain

Looks like Tom Friedman was right…


“This is the first time a planetary
-
scale social
network has been available to validate the well
-
known
“6 degrees of separation” finding by
Travers and Milgram [1969]. The earlier work
employed a sample of 64 people
and found that
the average number of hops for a letter to travel
from Nebraska to Boston was 6.2 (mode 5,
median 5), which is popularly known as the “6
degrees of separation” among people.”


“We used
a population sample that is more than
two million times larger
than the group studied
earlier and confirmed the classic finding.”

The More Things Change…


“We find that people tend to communicate
more with each other when they have
similar age, language, and location”


“Cross
-
gender conversations are both more
frequent and of longer duration than
conversations with the same gender.”


Some Findings:

MashupOS
: Security in Cloud Services


Today’s mashups turn the browser into a multi
-
user system


Mutually distrusting domains become co
-
users


No control on content integrated from different domains


MashupOS

will apply
operating system principles
to mashups


Service
-
based resource isolation


Protected, data
-
only, message
-
based


communication between services



Invokes well
-
understood Secure OS principles

to provide a stable security foundation

to replace today's mashup anarchy

All the Time



Cloud Computing



Live Mesh



Software + Services

-
> Immersive “augmented reality”

“Cloud Computing”:
Different Definitions

Technically, they may all be “Off
-
premise, Virtualized, Scalable (up
and

down)”


But

Different Business Models


Utility computing
-

Virtual hosting (e.g.
Rackspace

Cloud)


Cloud storage
-

Data hosting (e.g.
Flickr
, Amazon S3)


SaaS

-

Hosted services, email,
Ims

(e.g. Salesforce.com)


PaaS

(“Platform as a Service”) hosted apps.(e.g. Google Apps Engine)



We are bringing these elements together


into a cohesive platform: Windows Azure

Azure: Enterprise
-
Class
Cloud Services

Windows Azure Services Platform

Internet
-
scale cloud computing and services platform




Hosted in Microsoft data centers.



Provides a range of functionality to build applications
that span from individual
mashup

to enterprise scenarios.



Includes a cloud operating system and a set of
developer services.



Fully interoperable through the support of industry
standards and web protocols such as REST and SOAP.



You can use the Azure services individually or together,
either to build new applications or to extend existing ones.

Blogging, social networking


Data processing/transformation


Content upload, sharing, discovery


Storage, computation, messaging


Identity


Mashups: composing data and
applications

Azure: The Web as an Application Platform

SensorMap

Functionality: Map navigation

Data: sensor
-
generated
temperature, video camera
feed, traffic feeds, etc.

Microsoft Institute
-
sponsored
development: Semantic Virtual Earth

Integrates real
-
time
“real
-
world” data
from VE, into rich 3D
immersive
simulation

PhotoSynth: Beyond “Image
-
Stitching”

A technology that analyzes
related images and links
them together appropriately,
to re
-
create physical
environments in a navigable
virtual space.

GeoSynth
: the Semantic
Metaverse



Hostable

behind a secure enterprise firewall



Useful on huge datasets (e.g.
Flickr
, individual hard
-
drives)



PhotoSynth

captures all metadata with images



Will enable semantic image browsing , searching ,
geolocation


Small, colorful codes that can be printed,
displayed, emailed, disseminated anywhere


Simple software works on any smartphone
(yes, even Apple’s iPhone)

Microsoft Tag

Link to online
information

Simplify
personal or
business
contacts

http://www.microsoft.com/tag

Tying it All Together: Live Mesh

Access to all your data, anywhere

http://www.mesh.com


http://www.mesh.com


© 2007 Microsof t Corporation. All rights reserv ed. Microsof t, Windows, Windows Vista and other product names are or may be re
gis
tered trademarks and/or trademarks in the U.S. and/or other countries.

The inf ormation herein is f or inf ormational purposes only and represents the current v iew of Microsof t Corporation as of the
dat
e of this presentation. Because Microsof t must respond to changing market
conditions, it should not be interpreted to be a commitment on the part of Microsof t, and Microsof t cannot guarantee the accu
rac
y of any inf ormation prov ided af ter the date of this presentation.

MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.


more information

http://research.microsoft.com


lewiss@microsoft.com

www.ShepherdsPi.com