Slide 1


Published on

  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Slide 1

  1. 1. Microsoft Semantic Engine<br />Naveen Garg, Duncan Davenport<br />Microsoft Corporation<br />SVR32<br />
  2. 2. Microsoft Semantic Engine<br />Unified Search, Discovery and Insight<br />
  3. 3. The Situation Today<br />Significant Content is Outside Structured Storage (RDBMS, OLAP, BI)<br />Integration of this Content is Prohibitively Expensive (Time, Money, Resources)<br />Extracting Insight, Analytics, and Recommendations is even harder<br />Situation is a Confluence of Search | Predictive Analytics | Large-Scale Collaborative Filtering<br />
  4. 4. The Solution<br />Having all forms of digital information on asingle platform allows people to blend unstructured and structured content and to drive insight and decision making <br />Microsoft Semantic Engine provides a combination of technologies to form a contextual understanding of all digital content<br />
  5. 5. Scenario|Meaning driven insight<br />
  6. 6. SCENARIOS|UNIVERSAL APPEAL<br />Search and Collaboration | Personalized search, discovery and organization<br />Legal | Precedent and subject based search over large scale textual corpuses<br />Life Sciences | Systems biology with large volume data correlation and search<br />Government Services | Intelligence, real-time analytics, visualization, clustering<br />Social Networking | Social graph relevance mining, ranking criteria auto tuning<br />
  7. 7. FEATURES|UNIFY YOUR CONTENT<br />Unified Search, Discovery and Insight<br />Automatic Clustering and Organization <br />Meaning-Driven Indexing, Classification and Storage<br />Scalable Content Processing over all Content Types<br />Instant On Experience for Out of Box Value<br />
  8. 8. DEMO|VIEWS GALLERY<br />Search, Discover and Organize features exposed via sample UX gallery<br />Seamless installation and indexing of desktop, email and web content<br />Fully documented Managed APIs used in UX gallery and JavaScript / C# samples<br />
  9. 9. DESIGN|MEANING-DRIVEN PROCESSING<br />Streams | Descriptors (Properties) | Kinds (Concepts)<br />Streams processed into contextualized and indexed concepts for search | discovery | organization<br />LEGAL DOCUMENT<br />CONCEPT<br />EVIDENCE<br />CONCEPT<br />LEGAL CASE [xxx]<br />CONCEPT CLUSTER<br />KR_CLIENT_225.docx<br />STREAM<br />EXTRACTED PROPERTIES<br />PROPERTY<br />BILLABLE WORK<br />CONCEPT<br />DEPOSITION<br />CONCEPT<br />SEARCH AND SHARE<br />MDP<br />
  10. 10. Engine consists of self-contained set of pluggable services<br />Search and Markup<br />Trend and Predictive Analysis<br />Automatic Organization<br />Recommendation and Discovery<br />Semantic Engine<br />Clustering<br />Text Processing<br />Video Processing<br />MDI (RBV)<br />Image Processing<br />Audio Processing<br />Supervised Machine Learning<br />Conceptual Search<br />Inference<br />Sequence Store <br />(Suffix Tree)<br />Distributed Content Store<br />Ontology and Taxonomy Management<br />DESIGN|ARCHITECTURE<br />
  11. 11. Scale out by adding boxes; standard “web farm” (VIP) configuration<br />Scale out by adding boxes; each box can run all processors or specific processors<br />Store(<content>) Annotate(<kind>)<br />Index(<content>) Organize(<kinds>)<br />Search(<query>) …<br />Text<br />Image<br />Audio<br />Video<br />Video<br />API1<br />API2<br />API3<br />Analysis3<br />Analysis2<br />Analysis1<br />The logical architecture partitions analysis, indexing and storage<br />Staging<br />Core<br />Index<br />Stream<br />Single Logical<br />Partitionable<br />DESIGN|SCALABLE ARCHITECTURE<br />
  12. 12. Designed to be hassle free out of the box<br />Several programming languages and frameworks supported<br />CLR/.NET, JavaScript, TSQL, C++<br />DESIGN|PROGRAMMING<br />
  13. 13. DESIGN|PROGRAMMING<br />Sample of storing a stream in the system<br />Initiates the content processing, classification, and indexing<br />
  14. 14. DESIGN|PROGRAMMING<br />Sample of search and recommendations<br />Returns contextual results from the store and the web<br />
  15. 15. DEMO|WINDOWS 7 SHELL EXTENSION<br />Seamless Integration in Windows Desktop Federated Search<br />Expose Meaning-Driven Indexing and Semantic Actions<br />Zero Learning Curve<br />
  16. 16. Files<br />PlugIns<br />Importers<br />PlugIns<br />Importers<br />Plug-Ins<br />Importers<br />API Layer<br />System Integration Fabric (SIF)<br />KindLink<br />Kind<br />Descriptor<br />Stream<br />Semantic<br />Engine<br />Database<br />DESIGN|ARCHITECTURE DETAILS<br />ListKind<br />
  17. 17. DESIGN|ANATOMY OF A KIND<br />
  18. 18. DESIGN| MODELSPACE<br />
  19. 19. Periodically, MSE checks the User database for Changes<br />All Change data is returned to MSE as one XML block<br />MSE creates Kinds and Descriptors as needed, and Commits the activity<br />MSE data is exposed through custom views keyed to the Users’ Primary Keys<br />DESIGN| PROPERTYSPACE<br />
  20. 20. DEMO|SQL PROPERTY PROMOTION<br />Seamless Integration of Meaning-Driven Indexing in ALL SQL Tables<br />Expose Meaning-Driven Indexing via T-SQL<br />
  21. 21. PARTING THOUGHTS<br />Unified Search, Discovery and Insightover Every Digital Artifact<br />Extensible and Scalable Semantic Platform<br />Zero Learning Curve<br />
  22. 22. YOUR FEEDBACK IS IMPORTANT TO US!<br />Please fill out session evaluation forms online at<br /><br />
  23. 23. Learn More On Channel 9<br />Expand your PDC experience through Channel 9.<br />Explore videos, hands-on labs, sample code and demos through the new Channel 9 training courses.<br /><br />Built by Developers for Developers….<br />