OntologySummit2011 ApplicationCases Synthesis

OntologySummit2011: Application and Use Cases synthesis
Case Study Summaries Each Case Study participant was asked to provide a grid on one slide, outlining the business problem, the solution, key features (or screen shot) and business benefits. The aim of this was to be able to identify what sort of "Ontology" this was in terms of the application framework once this was completed, and what metrics (if any) were avilable to determine the business benefits.

Each Case Study


 * Challenge
 * What business problem the Ontology set out to address
 * Solution
 * What we mean by ontology in this case e.g. application, conceptual model
 * Screen shot or key features
 * Give a flavor of the ontology
 * Benefits
 * What metrics if any were used to demonstrate the benefits of this ontology.

Virtual assistant as a next UI paradigm Apple Siri

Challenge


 * If your computer were really smart
 * It would understand you in your language
 * It would make sense of your environment
 * It would help you solve everyday problems
 * It would be at your service, everywhere

Key Ontology Features


 * Ontology-driven virtual assistant: as a next generation assistance paradigm for smart consumer applications:
 * Ontology-based UI focuses is on task completion, with intent understanding via conversation in context, and where the virtual assistant learns and applies personal information.
 * Key roles for ontology include interface intelligence, unification of domain and declarative task models, semantic autocompletion, and service orchestration

Solution


 * Virtual assistant leveraging semantic technologies
 * Domain and Task models
 * Speech + Semantics + APIs (application interfaces)

Business Benefit


 * Virtual assistants that complete tasks for you
 * Usability
 * Context sensitivity for services

Integration of Multiple Systems from Multiple Companies YefimZhuk, Sallie Mae

Challenge


 * Multiple systems and sources of knowledge in different parts of the enterprise, owned by different communities of practice.
 * Gaining time and commitment from subject matter experts to ensure completeness of the model.
 * Different groups see different shades of meaning and application for similar terms, in different contexts.
 * Needs a unifying approach supporting local views

Key Ontology Features

Solution


 * Facilitation of knowledge gathering using ontology engineering methods.
 * Formal ontology notation for single ontology, while presenting views and facets of this to subject matter experts.
 * Curation of the ontology

Benefits


 * Best use of subject matter experts�� time and resources
 * Curatorship of Enterprise Semantic Architect ensures quality, consistency and completeness of the ontology
 * Collaboration in industry standardization efforts (e.g. EDM Council), via common semantics
 * Ensures that the knowledge captured at Sallie Mae is taken forward to industry-wide standardization efforts which we can then use

Standardization of Terms and Definitions for Financial Services MikeBennett, EDM Council

Challenge


 * Industry standardization of terms and definitions
 * Integration of multiple sources and feeds into disparate database structures
 * Even a small financial firm has 50 �C 100 separate systems each with its own data model
 * Tried: XML (MDDL); UML data models (ISO 20022)
 * Industry response: ��We need semantics

Key Ontology Features


 * Domain model
 * UML Tool based Editor
 * OWL Constructs
 * Spreadsheet View
 * Simple Diagram View

Solution


 * Semantic (conceptual) model of terms, definitions
 * OWL/ODM metamodel with UML tool
 * Adapted for readability
 * Present draft to business SMEs for input
 * Explained format to SMEs as set theory
 * Reviewed via webcast, direct input to model

Benefits


 * SMEs understood the format and contributed new knowledge on e.g. exotic structured finance
 * Answered industry call for standardization of meaning
 * Industry applications including mapping, master data models, messaging
 * Atomic building blocks means flexibility in defining novel financial products
 * Traction from regulators, for tagging of documents at source, reporting, systemic risk oversight

Semantic Tech in Rental Product Marketing JimRhyne, Sandpiper

Challenge


 * Help consumers find offerings
 * Help consumers select offerings

Key Ontology Features


 * Complex, multi-entity ontology
 * Lots of restriction axioms, not subclass hierarchy
 * Classification and realization part of application process

Solution


 * Semantic aided search
 * Semantic aided SEO
 * Rule-based product selection

Business Benefit


 * Current project is a pilot - stay tuned
 * Progress in discussions with Search Engine Providers

Ontology and Rules provide rapid Natural Language Understanding ChuckRehberg, Trigent Software

Challenge


 * Parsing natural language is complex
 * Identify specific text within a large set of a
 * documents that contains the same or similar
 * meaning as a given natural language description
 * of interest.
 * How do we use and grow Ontologies?
 * How do we map Natural Language to Ontology?

Key Ontology Features

but that is a choice of the application pictures, video, et. relationships, roles, generalizations, instantiations, characteristics, attributes, and units of measure
 * Language-independent Ontology
 * Semantic Items may have a language equivalent
 * Designed to semantically represent smell, sound,
 * The Ontology represents concepts, instances,

Solution

questions), parse and map the various valid constructs to semantic items in an Ontology (we call this mapping the ��meaning�� of the text) combinatorics of language constructs that represent the mapping as having an equivalent ��meaning map�� identify possible meaning matches relates to the original text along with hyperlinks
 * Given some Natural Language text (one or more sentences or
 * Generate (non-statistical) ��reader rules�� to recognize all
 * Apply the high speed ��reader rules�� to a large corpus of text to
 * Verify text identified as having the same ��meaning map��
 * Generate a report showing the information found and how it

Business Benefit

return just what you are looking for without the need to read the individual files yourself.
 * Changing the Dictionary has immediate effect
 * Changing the Ontology has immediate effect
 * Ontology grows with use
 * Ontology curation is widely leveraged
 * Sifts through a large amount of text to find and

Ontology and Rules provide Mass Customization of Vehicles ChuckRehberg, Trigent Software

Challenge

selecting the base model and a wide range of attributes (e.g. vehicle length) and features (e.g. number of exits) assemblies, and locations for a given vehicle �C Each vehicle off the assembly line can be one-ofa- kind.
 * Mass Customization of Trucks and Busses
 * Customers describe the desired vehicle by
 * Combinatorics of parts and assemblies
 * More than 480,000 combinations of parts,

previously built, identify the best set of parts, assemblies and component locations for the vehicle (the Vehicle Configuration) at different plants at different times. So, need to select a configuration that can be built at a plant prior to the promised delivery date.
 * Given an order that may never have been
 * Different parts and assemblies will be available

Key Ontology Features


 * Ontology
 * Rules Engine

Solution


 * Solution Ontology
 * Ontology defined both bottom-up and top-down
 * Solution Rules Engine

both abstractions and instances the engineers patented (2008)
 * Domain-specific UI
 * Engineers identify specific combinations in terms of
 * Rules are generated; They are not directly written by
 * Engineers work only in terms of their domain Ontology
 * Employ a fast Rules Engine
 * Over 600K rules with avg. 24 condition elements
 * Truck configured in under 10 seconds on my laptop
 * Worlds fastest most scalable rules engine �C recently

Business Benefit

of new variations (incl. features and attributes) take effect immediately (or at designated times and plants) proven engineering work
 * Ontology allows quick and reliable specification
 * Rules are specified in terms of the Ontology
 * Changes in Ontology and Changes in Rules can
 * Allows flexible change in suppliers and parts
 * New models and variations reuse previously

Content Intelligence and Smart Applications GregBardwell, Innovative Query Inc.

Challenge

to generate insights to improve business outcomes with content.
 * Content Intelligence: the ability

Key Ontology Features


 * Proactively serving up needed information

Solution


 * IQExplore
 * Semantic Analysis with Natural Language Processing

Business Benefit

publishing applications
 * Improved search, discovery and collaboration
 * Pushing the right information to the right users to do their job
 * Improved information and content publishing
 * Mashups of and with content for new classes of BI and
 * Unlocking information for actionable insights

Semantic BI for Blogging Bardwell - see above

Challenge


 * Utilize data obtained from news,
 * social media, and internal sources
 * Optimize and personalize search
 * Work with open sources
 * Respond quickly to chatter

Key Ontology Features


 * see above

Solution


 * NLP and Semantic index for unstructured sources
 * Custom scoring/alerts for results
 * Authoring tools to expedite content creation and analysis tasks

Business Benefit


 * Save time on analysis of content
 * More complete intel from text sources
 * Quicker and more precise responses to social media
 * Better and faster content creation

Valuing the Harvest from using Ontologies RalphHodgson, TopQuadrant

Challenge


 * Complex information spaces
 * Need to turn these into "Layered information spaces" that are fit for purpose
 * Filter to context

Key Ontology Features


 * OWL Ontologies
 * Graphical views and edit
 * Complex query support

Solution


 * Enterprise Vocabulary Management
 * Flexible solutions for managing business vocabularies in support of content delivery, search, navigation, data integration and disambiguation of terms


 * Semantic-XML Message Builder Workbench
 * Enables XML-based data exchanges that are specific to the local context while remaining compliant with industry and enterprise standards


 * Data Integration
 * Federated access to disparate information sources


 * Enterprise Architecture
 * Solutions for IT governance and management

Business Benefit


 * Canonical data - Subject-Predicate-Object Triples
 * Identifiers - Composition Construct for Aggregations
 * Schemas are also expressed in Triples and can be queried using same query language - SPARQL
 * Evolvability�Cschemas, vocabs and datasets can readily evolve

Architectures and Ontologies for Business Value CoryCasanave - Model Driven Solutions

Challenge


 * Fragmented architecture domains
 * Enterprise Architecture
 * Business Intelligence
 * Business Process
 * etc.


 * Business and systems financial architecture for a government agency
 * Understand the business needs in terms of business processes, information and business services
 * Specify the data, technology processes and SOA services of the systems to meet business needs

Key Ontology Features


 * Ontology Architecture views
 * Distinct views for business, systems, technology

Solution


 * Requirements, processes & services are less often captured as ontologies
 * Yet the ontology of a domain must include these viewpoints
 * Better support for other viewpoints with architecturally focused ontologies would provide increased value
 * Links between architectural an ontological tools provides a bridge between these related approaches

Business Benefit


 * Architectures and ontologies are mutually supportive
 * Ontological precision and the ability to federate ontologies brings value to architecture
 * Architectural tools can provide a more friendly way to express ontological information to stakeholders
 * Automating parts of systems from models and ontologies using MDA (model driven architecture) provides the much of the value without runtime overhead
 * The strategic opportunity is to bring all of this information into focus for the enterprise �C we are only starting to do so.

Model-driven Framework for Process Deployment, eXtreme Traceability SanjivaNath, ZAgile

Problem


 * Project Mgmt is Costly
 * Siloed Tools
 * Distributed Environment
 * Lack of Formal Processes
 * Lack of Traceability

Solution 


 * Integration of People, Tools and Processes
 * Application Integration Platform & Connectors
 * Methodology and Process Modeling
 * Integrated BI
 * Model-driven Architecture

Technology


 * Semantic Technology-based Architecture
 * Domain and application-specific Ontologies
 * Jena-based Framework
 * RESTFul Interfaces
 * SPARQL support

Business Benefit


 * Reduced Costs and Increased Visibility
 * Effective Collaboration
 * Efficient Project Tracking
 * Rapid Knowledge Access

Applying Semantics to Enterprise Systems - Proctor and Gamble Case Study DaveMcComb, Semantic Arts

Challenge


 * Large consumer products company
 * Looking for ways to integrate research findings across disciplines
 * Over 10,000 researchers in nearly 100 disciplines
 * Each discipline has its own language
 * Traditional key word search not useful when searching across domains
 * Problem compounded by departure of many key researchers (retirement, re-organization, etc.)

Key Ontology Features


 * Formal Ontology using OWL Constructs
 * UML Toolset
 * GIST Upper Ontology

Solution


 * Enterprise Ontology for the R&D domain.
 * Interviews with retiring researchers.
 * Re-use of terms from GIST upper ontology
 * Semantic Wiki built based on ontology
 * Two additional domains have been modeled (feminine care and baby care) and both reinforce the original abstractions
 * Additional domains planned for this year

Business Benefit


 * Of the nearly 600 classes in the R&D ontology
 * Only 2 were not derived from gist:
 * Brand
 * Invention
 * Most R&D data is findable without needing to know the specialized dialect of each subdomain.

Ontologies and CRM for Telecoms BillGuinn, MikeLurye, SusanMacCall, Amdocs

Challenge


 * Customer Relationship Management
 * Massive scale
 * Inferencing requirements
 * Structured and unstructured data
 * Past, present and future views

Key Ontology Features


 * Allegrograph triple store
 * AIDA Inference Engine
 * Semantic Concept Model

Solution


 * Built a "Guided Interaction Advisor"
 * Pre-built ontology and rule set

Business Benefit


 * Eliminates system and agent diagnosis time
 * Provides consistent and efficient call handling
 * Increases agent and customer satisfaction
 * Anticipated benefits based on 100K actual accounts assessment:
 * AHT reduction of 10-15

Do it Yourself Data Exploration Cambridge Semantics

Challenge


 * When events trigger action, researchers and analysts examine the data. Combining information from multiple spread sheets and databases is tedious and manual. Desktop tools do not know the categories and properties expressed by column (or row) headings. Moreover, for IT to create a new database or data warehouse is time-consuming, costly, and assumes that all requirements are knowable in advance..

Key Ontology Features


 * Sever based semantic model
 * Used to provide drop down selections in spreadsheets
 * Filters and lenses for populating and visualizing data

Solution


 * Knowledge-centric solution for data exploration links source data from spreadsheets, files, or database tables to a standard (semantic) model stored on a server.
 * There�s an app for that.
 * Works on desktop or via browser.
 * Selecting data to add to a spreadsheet is a pull-down menu option. Filters apply easily. Numerous lenses for visualizing data

Business Benefit


 * Focuses on ease of use for end-users with tools they know how to use; minimum IT involvement, if at all.
 * Rapid and low-cost to solution (hours/days), vs. slow and time-consuming for RDBMS, data warehouse, or manual
 * Flexibility in the face of inevitable change: rapid, low-cost, incremental modification vs. time-consuming costly, and difficult revision of conventional stores.
 * "Low-hanging fruit" for many agencies and programs

Better access with semantic search, navigation, query & question answering Recognos Financial

Challenge


 * Mutual fund industry rules change requires consumer friendly interactive access to 250,000 mandated plan documents.
 * While the industry�s trade association has developed a standard taxonomy for key topics, (a) buyers do not know industry jargon, (b) often related data is not adjacent to topic, and (c) buyer lacks a way to hone in on answers to questions.
 * Conventional DB and CMS approaches are labor intensive, error prone and costly to update.

Key Ontology Features


 * Structuring of unstructured data (specifically, documents in the Edgar database) enhances citizen access to information
 * Example is a Mutual Fund Prospectus that was filed with the SEC.
 * Uses a taxonomy associated with Mutual Funds. It is widely used and was developed by the Investment Company Institute.
 * User screen displays the actual filed document with the selected field value highlighted within the body of the document itself.
 * The taxonomy maps to the point in the document where the topic occurs.
 * This mapping is automated. Semantic processing of the text results in clean data (as received directly from the source of the data) with all standard data elements identified and extracted from the document without being touched by human hands (no data entry). Automated semantic extraction maintains the integrity of the data.

Solution


 * Knowledge-centric solution semantically analyzes and indexes the database corpus using deep linguistics and domain knowledge to extract data, link information to topics, and find answers to questions.
 * Consumers can navigate by topic (faceted search) pose questions in natural language, and query data contained in documents as though it were a database.

Business Benefit


 * Concept-based faceted navigation uses semantic analysis of content to reduce cognitive burden for users including extract specific data from tables (e.g., the amount of a specific type of fee). Question answering allows users to express questions in their own words and get the right answer.
 * Automated semantic indexing and analysis is more consistent, accurate, and cost-effective than comparable manual methods. Since, 80% of all data in organizations is unstructured, applications within government and industry are massive.

Knowledge-centric information webs & process interoperability Revelytix

Challenge


 * DoD attempted to build a data warehouse to connect HR systems and information across the Department. After 11 years and $1B dollars expended, had nothing to show for it.
 * "We�ve tried everything else and failed." - DoD CTO for Business Mission

Key Ontology Features


 * Uses a domain ontology in conjunction with a number of other ontologies:
 * Discussion ontology
 * Process ontology
 * Standards ontology
 * Community ontology
 * Analytics ontology
 * These are mapped via relational mapping ontologies, to individual relational databases.

Solution


 * Built a semantic information web that connected existing systems of record using a common domain ontology connected to relational mapping and source (metadata) ontologies
 * After 9 months (and very modest dollars expended), DoD had demonstrated a solution

Business Benefit


 * Semantic information web ontology patterns enable federated search, information sharing, and SQL-like querying across heterogeneous business databases.
 * Basic to very complex analytics and reporting across all systems become end-user generated queries that reference analytics ontology(s) connected to the domain ontology.
 * Development, extension, and upgrades to the �system of systems� is rapid, incremental, iterative, non-invasive and low-risk.

Do-it-yourself semantic agents to discover, aggregate, analyze & report information Connotate

Challenge


 * Agencies need to find, monitor, aggregate and make sense of information from a great many sources across the web as well as internally within government.
 * The manual effort involved can consume 25-45% of an analysts time.
 * Also, it is costly to custom program and update searches and analytics as needs change

Key Ontology Features


 * Several components of the solution:
 * Agent Studio
 * Discover
 * Create Sources
 * Enable integrations
 * Agent Library
 * Mashup
 * Personalize
 * Enhance
 * On demand applications
 * Agent Portal
 * Integrates with a number of other applications (email, spreadsheet, database, alerts, mobile apps etc.).

Solution


 * Intelligent semantic software agents to access, harvest, tag, and standardize information that are easy to create by anyone and can be shared and reused.
 * Train agents to capture site information, content elements, and take action to extract specific data, capture files, define schemas.
 * Agents "speak" HTML, XML, RSS, RDF, PDF, database and Excel.
 * Mash-ups create new data by element and schema, in time periods, across sources and time periods, and put data into context

Business Benefit


 * 360 degree views on topics, issues, etc. combining information from internal and external sources including web pages, blogs, local news, message boards, social media, databases, email, intranets, enterprise applications, etc.
 * Productivity improvements from automated gathering, monitoring, and alerting for needed information events that is 24/7/365 or other frequency

Smart knowledge-driven citizen-centric services BeInformed

Challenge


 * Permitting site synthesizes requirements, processes, and information across multiple jurisdictions and 14 independent institutions into a unified user experience.
 * Immigration site helps new arrivals solve varied problems of relocation. It combines information, and decision logic from 12 agencies into an easy to use single point of service delivery.

Key Ontology Features


 * Integrates process flow, applications and sources, via a layer of knowledge models
 * Citizen centric:
 * no wrong door, all the right answers;
 * no unnecessary questions;
 * correct advice based upon the citizen's personal context;
 * pro-active service;
 * tailor made.

Solution


 * Knowledge-centric solution separates the know from the flow and the function to create declarative applications configured by users with semantic models of legislation, knowledge, processes, data, and UI.
 * The core infrastructure consists of an ontology, which is enriched with business rules.
 * All functions use the same ontology, e.g., semantic search, information access, automated decision making, decision support, and dynamic processes.

Business Benefit


 * �Open knowledge as a service� bridges the gap between government and citizens and facilitates effective cooperation between independent institutions � both public and private.
 * Provides automated decisions and decision support; means for agencies to manage their knowledge / rules; ability to quickly adapt to external events / implement new legislation; improved decision making, guaranteed compliancy, less errors; improved service delivery to the public; and substantial cost reductions.

Policy-driven compliance, risk, and change management ''Visual Knowledge'

Challenge


 * Global financial services firm was $600B behind in M&A because it could not keep up with compliance requirements. Knowledge to track and report regulatory mandates comprehensively across the business was fragmented in separate documents, systems, and data stores, thus slow, prone to error, and difficult to change.
 * "Our only solution is to add more belly buttons, which means committing thousands of people to compliance."

Key Ontology Features


 * Risk, compliance and policy-driven processes include situation awareness, exceptions, fraud, case management, emergency response.
 * Semantic technologies map the external (legal) requirements to policies (expressed as documents) to semantic models (defining information structures, processes, and user responsibilities) to system infrastructure, behaviors, and analytics (measures of performance). This provides visibility and traceability from mandate to manifestation to compliance reporting.
 * Uses semantic technologies to specify, interrelate, and manage knowledge at the level of individual concepts and relationships independently of its source artifact, whether this knowledge comes from:
 * Documents via natural language or visual language understanding, and semantic tagging;
 * Models including file formats, database schema, object models, content semantics, policies, event and process models, context models, domain ontologies, etc.
 * Behaviors including software source libraries, directories, frameworks, methods, subroutines, objects, semantic agents, and infrastructure standards.

Solution


 * Knowledge centric collaborative solution that captures all of the regulatory mandates, maps them to policy documents, then to semantic models defining schemas, processes, and decision-making rules, to deployed operational systems and procedures, to analytics that track, assess, and report human and system behavior and ensure compliance

Business Benefit


 * Development of knowledge-centric compliance solution requires fewer resources, is more rapid, less costly, quicker to show value.
 * Operation of knowledge-centric solution requires less labor, is more reliable and less error prone.
 * Maintenance and upgrades are less costly and time consuming. Assessing impact of changes on documentation, systems, and procedures is more automated. Change management and version control is automated.

-- maintained by the Track-2 champions ... please do not edit