• Founded in 2005
• Offices in Washington DC and Boston
• Customers in US gov't, banking, energy, health/bio, retail
• Strong academic partnerships in US, UK, Europe, and Mexico
• Leader in Semantic Technologies: graph databases, reasoning, planning, services

• We agree that graphs are great
• Fantastic alternative to traditional systems
• Superior for many tasks
• Semantics are a natural complement
  
• As asset for information integration & analysis problems

• Organizations employ people other than programmers
• Not everything needs to be in code
• Especially business logic
• You can't just teach everyone programming either
• So why not semantics?
• Declarative, formal descriptions of nodes, edges and relationships
• A little bit of semantics goes a long way
  
• Yields a powerful combination that is well suited for information integration tasks

• Adding semantics to our graphs gives meaning to our data
• Better yet, it lets us free our business logic from our codebase
• By encoding it in the graph
• Still better? It frees it from programmers (sorry!)
• Can be written in a high-level declarative language
• By programmers and non-programmers
• Also lets us talk about the same thing while using different vocabularies

• Complex organizations face challenging IT problems

• Enterprise Semantics: model-driven information integration for decision support & analysis
• Let an enterprise better utilize what they already have
• Data is an underutilized asset
• Leverage semantic graphs
• Smart data, not strictly big data
  
• Scale is not a necessary condition for utility
  
• Problems are not always solved with more data
  

• Help deal with information overload
• Not just overload, but also management
• Semantics adds meaning to the data so the computer can help
• Isn't this another silo?
  
• Consumers of the data simply issue queries
• But the data can live in place
• Key if it changes frequently
• Easier to bring new data online
• Domains use their own vocabulary, semantics provides the glue
• Facilitates cross-cutting views for information analysis & accessibility

• Business logic is naturally a first class object to be curated, validated, versioned, etc.
• Encode your business logic in a language with a formal semantics
• Capturing expertise
• Inference rules & queries
• Let the computer do the work
• No programming required
• Can write programs, but that's not easier or more maintainable
• Actual experts in the business logic can implement it, not a programmer
• Lets non-programmers do complex information processing without having to write code

• Make implicit information explicit
• Consistency checking
• Find contradictions in your logic
• Entailment
• Answer queries using reasoning
• Integrity Constraint Validation
  
• In some use cases, data consistency is crucial
• Explanations
• Logical debugging
• Can be as complex as your domain requires
  
• Expressivity broken down into profiles based on computational characteristics

• For example, enforcing security (ACLs)

• You want to see if Bob can access Resource1
• This can get hairy quick
• Finding this in a single query, or multiple related queries is not easy

Bob is-a Admin OR Bob created Resource1 OR (Bob hasRole ?x AND ?x canAccess Resource1) OR ... 

• It doesn't get better writing a program
• We can leverage reasoning to make this easier
  

• But if you encode your security policy in semantic graphs

• Bob canAccess Resource1

• That's it!
• Allows a more concise and maintainable query
• Shields you from evolution in the security policies in your application
• Try updating that query (or program) for each change! 

• Data is key for solving problems with Enterprise Semantics
• As I said, smart data
• In some use cases, data consistency is crucial
• So we need to be able to enforce integrity constraints over our data
• Age must be a positive integer
• Supervisors may not supervise more than 10 employees
  
• All US Citizens must have one Social Security Number and it must be unique
• ...

• Custom logic to validate incoming data can be complex & time-consuming
  
• Not to mention telling a user what they did wrong
• No need to enforce some in the application layer, some in the database ...
• Semantic Graphs make this easier
• Again, no reason to write more code
• Can use same language as reasoning
• A rich modeling language for integrity constraints
• The high-level expressive syntax makes authoring complex validations easier
• Can be authored by the experts
• Reasoning is used to satisfy, or violate, your constraints
  

• Things go wrong, it's inevitable
• People fat-finger stuff when doing data entry
• Errors modeling the business logic
• You can sit around and debug them
• Debugging is fun, right?
• Some Programming Required
• Or you can have the computer tell you what is wrong
• In fact, it will tell you why

• Tells you why something is the case
  
• Inference from business logic you don't understand?
• Someone violate an integrity constraint?
• You get a proof
• Aids traceability & debugging
• Pinpoints the exact issue
• Can reference other proofs
• Even tells you how to fix it
• Or in some cases, fixes it for you

• Semantic graphs are still graphs
• Adding semantic capabilities does not diminish their graph-ness
• You can pointer chase if you want
• But all the normal algorithms still apply
• Page Rank
• Clustering
• Vertex Degree Scoring
• Betweenness Centrality
• Shortest path
• and so on...

• All of this is based on W3C standards
• Data format, query language, semantics; none are proprietary
• Don't like your current vendor?  Just switch
• Standard semantics means queries won't return different answers
• No ETL required to move your data
• Makes interoperability easier
• Promotes re-use
• Semantics does not change, data retains original meaning

• VIACOM
• Best Buy
• SAIC
  
• NASA
• Wolters Kluwer
• JPMorgan Chase
• inQuest Technologies
• IBM

• NASA is a big, interesting organization
• 100k+ employees across 12 centers throughout the US
• They have a universe of data
• And one simple problem, finding experts
• COTS solutions were not working
• All the information was already in-house, but collecting dust
• Enter POPS
• Created a model for their existing data
• Built a simple web interface to let them use their own data
• Saved NASA $38M a year!

• The leading semantic graph database
• Design influenced by our approach to Enterprise Semantics
• Optimized support for reasoning, integrity constraints & explanations
• These resolve to query answering
• So our main focus was on a query planner
• Built to support information integration & analysis applications
• But general enough to support any semantic graph use case
• Also wanted best out of box experience possible

• Make it easier for all developers to contribute
• Organizations don't always have experts in semantics or graphs readily available
• Provide a tool that abstracts away these details
• Rely on web development standards
• HTML, Javascript, CSS, etc.
• Start building an application right away without focusing on learning the semantics or the graphs
• Because the value is in solving the problem

• Graph databases offer advantages over traditional systems for many uses
• Semantics complement those advantages
• More than just programmers work at organizations
• Not everything needs to be code
• Key components, such as business logic, can live separate
• And thus, be created & maintained by non-programmers
• So using semantic graphs we retain advantages of graph databases, while adding features valuable to the enterprise
• Promotes adoption & interoperability
• Facilitates smart data & Enterprise Semantics