Wie Semantik Inhalte intelligent verbindet microDocs und mehrhs.i4icm.de/fileadmin/user_upload/Ziegler_Semantic_inTC... · 2020. 2. 13. · ©Prof. Dr. W. Ziegler I4ICM Wie Semantik

© P ro f. D r. W. Z i e g l e r I 4 I C M

Wie Semantik Inhalte intelligent verbindetmicroDocs und mehr

11. Februar, 2020 DOKU+MEDIEN FORUM 2020DokuNord, Hamburg

Fakultät Informations- und MedienmanagementHochschule Karlsruhe (HSKA)

Institut für Informations- und Content-Management (I4ICM)

Prof. Dr. Wolfgang Ziegler


Prof. Dr. Wolfgang Ziegler

– Karlsruhe University of Applied Sciences, Germany

„Communication und Media Management“ (HSKA)

» Knowledge, Information, content, data modelling

» Information processes and systems in TC

Institute for Information and Content Management

– Institute for Information and Content Management (I4ICM)

» Research Transfer (PI-Class, CVM, REx, CDP, CoReAn, microDocs)

» System evaluation/introduction, process analysis/engineering,

CMS/CDP optimizing, classification/content engineering


Agenda

• Introduction

• Content Management Concepts

• Content Delivery Concepts

• Digital Information Services

• Intelligence Cascade and Types of Semantic Applications

• Content Access using microDocs


Introduction


Evolution of Perspectives on (Technical) InformationIntroduction

Inte

rnal

Exte

rnal

Department Enterprise

CMS

Process &

System

Integration,

Industry 4.0

IoTECMS/

PLM

• Situational

• Online

• Mobile

• Onsite

Info

rmati

on

Use C

ases

Data Creation/Management

CDP DIS


What has been the focus of content management

and related technologies?

•Most granular Information (topic-based content)

•Metadata enrichment (of topics) according to product

variants

•Automization of CMS processes (generating, publishing)

Introduction

CMS


Collection of variants

depending on parameters

Manuals & data on

configurable products

Introduction

CMS


What is the focus of content management and

related technologies?

Lack of specific information…

Introduction

CMS


Content Delivery Objectives (User side)

• Create and deliver more and better user-centered information:

– situational (according machine and user situation)

– case-based (following predefined use cases)

– product-/variant-specific (as most as possible)

– suitable media

• Create new business cases / Digital Information Services

Introduction

CDP


Use Case Dependencies & Requirements

for Deliverables & Services

• Sales process / Information & Product Search:

– Overview data and summarized tables

– Specific: Data sheet, technical data, dependency information

•Machine Planning: specific envisaged configuration

• Set-up /Installation: specific context and configuration setting

•Repair planning: specific existing configuration

•Operation/customer services: Detailed information

Introduction

DIS


What is the (recent) focus of content management

and related technologies?

•Most granular Information (topic-based content)

•Metadata enrichment (of topics) according to product

variants

•Automization of CMS processes (generating, publishing)

• Support search & delivery (interfaces, facets, )

•Definition of use-cases (user stories, customer journey, …)

and target groups (personas)

Introduction

CMS


How can we…?

•…manage complexity (product, information, processes)

• … enable and empower writers to produce variant specific,

i.e. configuration-dependant information units

•… plan configurations and relevant parameters, also for

authoring & document creation

•… retrieve information in an efficient way

•… connect to use cases (sales, operation, servicing, …)

•… apply new and appropriate media

•… prove use and relevance of content

Introduction


Content Management

Creating (native) intelligent content


CMS drivers and demands from industry

Products from industrial engineering and manufacturing are characterized by „mass

customization“ and „globalization“ :

• Short-time development cycles

• Many changes within development phase

• Often changes within time of use (servicing for manufacturing/machines) and reuse of parts

• Use of standard parts & components (mass production)

• Adapting products to customer needs (customization)

• Adapting products to all addressed markets (globalization issues for export-oriented industries)

• Comply with regulations and standards

Software similar; more software variants (branches), platform & agile development;in many cases individual customizing;

CM Methods


CM Methods Basic CM Concepts in TC

•CMS principles

Controlled reuse of content modules (topics) in

multiple delivery structures, documents or media using metadata

•CMS offer technologies for

– Variant management (product & media variants, configuration)

– Version management (change Management)

– Translation management (internationalization, globalization)

– Cross media & publishing management


Reuse, Aggregation and Publishing

CM Methods Referencing modules/topics • permits controlled processes• avoids uncontrolled redundancies• defines and populates document structures by topics

Reuse by

Cross Media Publishing

(automated)

Referencing

Doc

Mod1Mod3

Mod4

Mod2 Mod5

Document Structures

Modules/

Topics

Data Format <XML>


11,4 %

11,0 %

3,7 %

3,9 %

2,1 %

2,0 %

7,8

50,7 %

7,4 %

14,6 %

16,9 %

6,5 %

5,1 %

2,8 %

2,5 %

7,1 %

39,2 %

5,3

14,6 %

26,3 %

6,4 %

3,5

2,5 %

3,9 %

10,5 %

29,3 %

2,9 %

CMS ist für uns kein Thema, mit dem wir uns befassen

Informationsphase über CMS

Entscheidungsphase pro / contra CMS

Gegen CMS entschieden

Für CMS entschieden: Analyse / Konzeption

Für CMS entschieden: CM-System-Auswahl

Für CMS entschieden: CMS-Implementierung

Für CMS entschieden: CMS-Nutzung

Für CMS entschieden: CMS-Systemwechsel

Umfragejahr 2000

Umfragejahr 2013

Umfragejahr 2018

System ChangeSystem Change

System Use

System Implementation-

System Selection

System Planning

Not an issue

Informing

Deciding

Decided (against CMS)

CMS Phases

Straub/Ziegler: tekom CMS-Studie 2019


What is TC focussing for authoring & linguistics?

•Comprehensive writing

•Rules-based writing

•Consistent terminology

•Writing for international markets (translation, localization)

• Self-contained information: topic-based writing

•Definition of target groups (personas) and use cases (user

stories, customer journey, …)

CM Methods


XML Authoring environments (for content creation) CM Methods


CM Methods Metadata Enrichment by

Semantic Metadata for Modular Content (PI-Class®)

Physical & Virtual Objects

(Product Components)

Content

Content

Content

Content Objects

(Modular Topics)

Operation

Dismount

Repair

Information Classes

Product Classes


Basic Dimensions of Module Classification

(PI-Class®)

Topic: self-contained information unit;

topic concept and content is defined by (intrinsic) PI-classes

CM Methods

Product-Class

Base/ Telescopic Rod

X3B, X3-H1,X5-B, X5-D,…

Information-Class

Operation/Height Adjustm.

User Manual,Service Manual,…

intrinsic

extrinsic

www.pi-fan.de


CMS „Taxonomies“ from Topic Classification CM Methods

Rotor

Display

Heating

X3B

T3B

ContentTopic

Safety

Repair

Functional Description

User Manual

Service Manual

Intrinsic Taxonomies Intrinsic Taxonomies

Extrinsic Hierarchies Extrinsic Hierarchies

Hierarchies, Taxonomies, List, …

Variant properties Functional Metadata

MultidimensionalInformation Space

Metadata

for identifying

and addressing

modules/topics


Use of (semantic) metadata for/from CMS

•Definition of topic content (module concept)

•Content planning (workflow, content development)

•Automized document creation (rules based)

•Variant management

CM Methods



ContentTopic

Tools

Variant properties Functional Metadata (Collections)

Time SpareParts

ErrorCodes

MaintIntervals

…

Extended PI-Class: Multidimensional Information Space



ContentTopic

Variant Properties/Features→ Product Configurations

Geo-metry

PartsNo

Mate-rial

Features

Extended PI-Class: Multidimensional Information Space

Location

Functional Metadata(Collections)


Information Environment and DependenciesCM Methods

Rotor

Display

Heating

X3B

T3B

ContentTopic

Safety

Repair

Functional D.

User Manual

Service Manual

Component

TaxonomiesInformation

Class

Product

Structure

Document

Types


Depency on:

• Stake Holder

• Process Owner & Driver

• Information Sources & Systems

Tech Comm.Standards, iiRDS

Tech Comm.Standards, VDI 2770

Service, Production,Software Dev., … IoT!

ERP, Engineering, PLM, PIM/PDM

Prod.Management,Salesor Tc

Engineering,Development,PLM, ERPor TC


Analyzing CMS processes and use of CM methods

• System development & improvement can be proven by KPI

•Reuse numbers of modules, fragments, media, …

•Reuse rates of deliverables (documents, …)

• Efficiency / cost indicators (Sharing factors per publication)

•Change rates and new content rates (per publication)

•Document fingerprints (module reuse rates within docs)

•Variant managment (number of variants, complexity ….)

Report Exchange (REx)

Method

CM Methods


Adjusting theintensity

The intensity of the fan can be adjusted in seven levels.

• Turn the levelknob until theintensity ofthe fan is setas desired.

Variant Management (Topic variants)

Extrinsic product variant

retrieved by usage

or described by product

types (for retrieval)

CM Methods

©Prof. Dr. Ziegler

Adjusting the intensity

The intensity of the fan can be adjusted in five levels.

• Turn the level knob until theintensity of the fan is set asdesired.

………

T3B, TAB, TB5,T35, T3X5B,…

TPB, TAMP, ….

Extrinsic Classification as variant property


The intensity of the fan can be adjusted continuously.Turn the level knob until the intensity ofthe fan is set as desired.

………

T7B, TFX,..


Variant Management (sub-modular; one topic)

Extrinsic product variant

collection (of all products)

for filtering

CM Methods

©Prof. Dr. Ziegler


The intensity of the fan can be adjusted in five levels.The intensity of the fan can be adjusted continuously.The intensity of the fan can be adjusted in seven levels.

• Turn the level knob until the intensity of the fan is set as desired.

………

T3B, TB5,T445, TX5B,…

TPB, TAMP, …. T7B, TFX,..

Extrinsic Classification as Variant Property


CM Methods Variant Management by Properties (submodular)

Topic planning according to

configuration variants

©Prof. Dr. Ziegler


The intensity of the fan can be adjusted in five levels.The intensity of the fan can be adjusted continuously.The intensity of the fan can be adjusted in seven levels.

• Turn the level knob until the intensity of the fan is set as desired.

………

levels = 5 levels = Cont levels = 7

Product features as variant property


Systematic Variant Analysis

Analyzing the origins of

large numbers of

topic variants

Driver of product complexity

and / or

(necessary?)

content complexity

CM Methods

Number of topic variantsper (PI-) Metadata combination

Number of (PI-) metadatacombinations

There are two (PI-)metadatacombinations having 33 distinct topic variants

(x=33; y=2)

One of two retrieved PI-Combinations:

Mounting Bracket / Base / Telescopic rod+Task / Mounting


Analyzing Content Variants (by content)

Example detected by REx

method:

All topics have the same

(intrinsic) classification

but differ in

(extrinsic) product classes

CM MethodsProdukt A

Produkt B

Produkt C

Produkt D

Produkt E

Produkt F


Summary I (CMS)

• Technology and methods are available in CMS for

topic-based and configuration-specific

information creation, document assembling, publishing and provisioning

(packaging). They rely strongly on taxonomic classification.

• Limiting factors are often data quality / process integration within

companies and human factors (complexity of information structures;

„lost in metadata“ of different configurations) as well as a lack of

information planning! (Therefore, also analytics is needed…)

• Planning, analysis and management of processes (like variant

management) depend on metadata quality

Content

Management


Summary /Vision II (CMS)

•Highly structured content, enriched by semantic metadata has been

created, but often only used for (automated) PDF/print production.

(Even though CMS can produce all types of media)

• Latest delivery technologies benefit from and require such structured

content packages

•New media (e.g. chatbots, training, animations, AR/VR/MR) demand

additional or new types of information structures and content

Content

Management


Intelligent Content Delivery (Methods & Technology)

Making use of native intelligence of content


Webshops as CDPCD Methods

Produktportale II (Amazon) Facets

Filter


Content Delivery

Portals

Basic definition and functionalities

Systems offering web based access to modular, aggregatedcontent or other information for various user groups by related retrieval mechanisms.

Basic functionalities

• Access or import content from relevant data sources and corresponding systems

• Manage and update content within the content lifecycle

• Retrieval functionalities including user interfaces for content searching

• Web-based display of content on a modular or document-based level

• Web services handling requests from other applications and events.

(Definition 2013)


CD Methods Facetted search/request and topic delivery

Oil Pump

Hydraulic system

Testing

Procedure

Z-006

M a c h i n e

Service

D o c u m e n tC o m p o n e n t

I n f o r m a t i o nZ-006, Z-007

Testing the pressure of the oil pump


CD Methods Facetted search/request and topic delivery

Oil Pump

Hydraulic system

Testing

Procedure

Z-006

M a c h i n e

a1 |b3 | … |x5 |y1 |z 5

Konfigurat ionC o m p o n e n t

I n f o r m a t i o nZ-006, Z-007

Testing the pressure of the oil pump

Customer-

dependent

Configuration !


Selection/Generating of

publication depending

on parameters

CD Methods


Content Delivery Portal (PI-Fan)CD Methods

[www.pi-fan.de]

Structured Search

Direct Search

Facets Navigation

Cleaning the rotor

Mounting the rotorProcedures

X-Series

All Components

DocufyTopic Pilot



©Prof. Dr. Ziegler

Navigating the document structure

(before/after facetted search)

Adjusting the tilt

Adjusting the tilt

PI-Fan T3-B

www.pi-fan.de


App Delivery Application including SearchCD Methods


Object Recognition and CDP

Content Request & Delivery

CD Methods CDP Request: Deep Link / Parameter Call


CDP: Facets in DocumentsCD Methods

Schema

Content Delivery

Server

www.pi-fan.de

Navigating the

document structure;

then facetted filter



SchemaContent Delivery Server

[www.pi-fan.de]


CDP environment in industrial applicationsCD Methods

CMS, …

CDP

xMS

xMS

xMS

CMSCMS

Supplier

AdditionalInformation&Sources

User InformationService Information

Off SiteWeb Portal / Mobile

Exchange Format(proprietary vs. standardized/iiRDS)

Machine state(errors, messages,operating conditions)

CDP

OnSite

Online/

Offline

„AR/VR/MR


What ist the (recent/real) role of TC in digitization?

Digital Services by different

• Projects

• Show cases

• Departments

CD Methods

CMS, … CDP

xMS

xMS

xMS

CMSCMS

Supplier

BusinessInformationSources

User InformationBasic / StaticService Information

Off Site (online/offline)Web Portal / Mobile

CDP

OnSite

AR/VR/MR

Machine state(errors, messages, operating conditions)

Chatbots/KI

Service, Production,Software Dev., Engin.… IoT

Sales, Marketing, …

TC


Delivery of TC content: Digital Content Service

CDP as Digital Content Service

• Content provisioningfor data integration

• Web interfaces /API

• Standard Formats(XML, HTML, PDF, iiRDS)

• Requires classified topicbased on variants andconfigurations!

CD Methods

CMS, … CDP

xMS

xMS

xMS

CMSCMS

Supplier

BusinessInformationSources

User InformationBasic / StaticService Information

Off Site (online/offline)Web Portal / Mobile

CDP

OnSite

AR/VR/MR

Machine state(errors, messages, operating conditions)

Chatbots/KI

Service, Production,Software Dev.,… IoT

Sales, Marketing, …

TC

DCS


Technological Use Cases

• Search documents (retrieval manually by facets, full text)

•Read documents (online use; navigation in toc;

adding comments and annotations)

•Download documents (offline use)

• Search topics & read

• Push topics from machine- and IoT-applications

•Collect and download topics & documents (**/remark below)

• Interactivity (Voice, object, behaviour, gesture, role recognition)

CD Methods


Defining User-centered Use Cases

•User situation

•Machine/product situation,

trigger (IoT?)

•General

information need

•Device, on/off-line/on-site,

Media

•Role (user, service, customer)

• Pre-Knowledge, level

Criteria

CD Methods

• Product lifecycle

•Customer journey phase

•Automated detectable

metadata

• Expected search behaviour and

priority (facets vs. full-text)

•Required topic or document;

additional information

•Definition of success


Use Case / User Story (Pre-Sales)

• Product search for planning component

exchange

• Has complex (water cleaning) machine

• Needs information if retrieved

component is suitable; dangers/risks,

restrictions, benefits

• Web-Site

• Planner (Technical staff; can trigger order)

• Knows necessary configuration and

parameters

CD Methods

• Maintenance

• New Customer, informing, procurement

• No metadata detectable

• Wants to select component-type and then

detailed configuration parameters;

wants to compare similar suitable products

• Needs appropriate product with

corresponding specific tech data (specs),

table overview as comparison of variants;

should receive correlated

information about dangers, restrictions

(TC),

benefits (sales)

• Buys component


Use Case / User Story (Installation)

• Installation process of sensor (unexpected

problems)

• Stopped machine; ready for exchange

• Needs concrete information about wiring of

sensor, possible problems and solutions

• App (Android), camera, on-line

• Service technician of customer company

• Knows usual wiring and sensor exchange

procedure

CD Methods

• Maintenance, repair

• Customer

• Errorcodes and related components

• Wants to detect component-type

information for installation

(code/Scanning);

• Needs appropriate installation procedure;

possible problems (description, solutions)

Additional: settings, testing

• Exchange;

exchange protocol, settings and testing

archived


Use Cases / User Stories and CDP kick-off

•Restrict to the most important use cases where benefit and

success can be „measured“

• Start with productive (reference) implementation according

to clearly communicated use cases

•Avoid higly sophisticated use case, unless user story is most

relevant business case

•Develop new digital information services; not another

online-help system no-one uses

• Think of CDP maybe also as an easy to use internal content

communication systeme

Recommendations

CD Methods


CDP and analytics in industrial applications

Analytics

CD Environment

CMS, …

CDP

xMS

xMS

xMS

CMSCMS

Supplier

Additional

Information

&

Sources

User Information

Service Information

Off Site

Web Portal / Mobile

Machine state

(errors, messages,

operating conditions)

CDP

On

Site

Online/

Offline

AnalyticsAnalyticsAnalyticsAnalyticsAnalytics


Overview of Content System Analytics

Analytics

CMS & CDP Analytics

CMS CDP

KPI

Delivery & Feedback

KPI→Metrics:• Reuse Rates

(Abundancy)• Redundancy• Document Sharing

factor• Variant management• Correlations;

Distributions…

Indirect feedback

→Metrics:• visiting time,• Visit frequency• search

behaviour• search terms• …

Direct feedback• Rating• Satisfaction

→ Improve: • Product• Information• Terminology

(Harvesting)

CMS Analytics (REx)

CDP Analytics(CoReAn)

ArtificialIntelligence→Quality

assurance:• Similarity

analysis• Classification

quality…


Digital Archiving Services & CMS/CDPCMS – CDPProof Levels:• Publication• Download• Deployment• Viewing• Search/Retrieval• Aggregation, Dynamic Pub.

CMS CDP


Remark on Archiving Service

Regulations require:

•Archiving of content packages and media (timestamp)

•Archiving of transfer data and proof of content delivery

•Archiving of ustomer aggreements and confirmations

•Documentation of systems & processes

•Archiving lifetime (10 y, 30 y or industry dependent)

•Data security (no access) and transfer security (encryption)

• Product monitoring through problem search and analysis

Digitization means, that content

delivered before as paper, will be

increasingly substituted by new

media and formats.

The increase of electronic

information delivery requires to

rethink and to reorganize for

example also corresponding

archiving processes.

Content Delivery & Regulations


Summary (CDP)• Technology of CDP is available mostly for delivery of document

packages and facetted search for contained topics and documents;

Source of facets are mostly taxonomies from CMS

•Remark**:Dynamic aggregation, variant management is (at the

moment) mostly done in CMS, not in CDP; configuration management

needs new approaches because of its more dynamic and complex

structure

•Delivery use cases for successful applications have to be clearly

explored and defined

•Delivery can be developed in addition as Digital Content Service (DCS)

for various external and internal applications and media

Content

Delivery


Digital Information Services

Business and Use Cases for Content Delivery




Digital Information Service (for machine service)

Service planning and tracking

CDP and CMS information is

connected to service

processes

Source: STAR AG

Retrieval/request by

(PI-)Classification


Access to granular service information & data

Source: STAR AG

Digital ServiceInformation Service

Interactive Data from

CMS & Engineering:

Sensing & archiving

of data setting


DIS as extendedproduct portfolio

Service Information and AR (Hololens)

https://www.youtube.com/watch?v=nyDZ7Q4AFu8

Source: Voith Hydro


Digital services asan extendedproduct portfolio

Service Information and AR (Hololens)

https://www.youtube.com/watch?v=nyDZ7Q4AFu8

Source: Voith Hydro

Interactive data from system sensors;

Content integration from

various sources;


DIS

https://www.youtube.com/watch?v=VGtCQWROytw




Installation, Configuration, TroubleshootingDIS

Reseller support

(Heating, AC)


Installation, Configuration, TroubleshootingDIS

Reseller support

(Heating, AC)


Remote Assist (including video + AR)

Service (von Ardenne)

DIS


Remote Assist (including video + AR)

Remote Service

(von Ardenne)

DIS


VR-based Learning

Training (von Ardenne)

DIS


Augmented Intelligence

The Intelligence Cascade


Levels of Intelligent Content and Data

Native IntelligenceSemantic content and semantic metadata for processautomization, e.g. PI-Classification

Extended/Augmented IntelligenceAdditional relations between (content) objects describede.g. by ontologies or other semantic

Artificial IntelligenceAutomated extraction of metadata and knowledge bystatistical methods (machine learning), …

Intelligence

Cascade


PI-Fan implementation & PI-Classification in CMS

» w w w .pi-fan.de

Native Intelligence

Extrinsic Classification

for Variant Management

On Topic or subtopiclevel


Practice

innovation

IDS c-rex.net

Native Intelligence Content Delivery Portal (PI-Fan)

Adjusting the tilt

www.pi-fan.de


14,0%

2,6%

2,3%

36,5%

19,6%

37,4%

24,0%

3,8%

2,0%

2,4%

19,8%

5,5%

2,7%

68,3%

PI-Klassifikation/ PI-Class (nach Ziegler)

iiRDS (tekom)

Klassenkonzept (nach Closs)

Eigenes Konzept

Klassifikation der Inhalte / Module gemäß Vorgaben der XML-Standardstruktur

Klassifikation der Inhalte / Module gemäß Vorgaben des CMS

Nein, keine weiteren Standardisierungsmethoden für Metadaten und Modularisierung

Standardisierung von Metadaten

Keine CMS Nutzung

CMS Nutzung

Use and Type of Classification with/without CMS

tekom CMS Study 2018

(D-A-CH/ Central Europe)

(approx. 700 – 850 persons)

Content

Management

No use of and no concept for classification

According to concept given by CMS

According to concept given byXML structure (information model)

Custom concept by CMS users

Klassenkonzept (Closs)

iiRDS (tekom)

PI-Class®PI-Classification

Not using CMS

Using CMS

Standardization of Metadata


Limitations of Classical Metadata

(Native Intelligence)

Intelligence

Cascade

• Taxonomies and metadata

hierarchies correspond to

a two-dimensional

(flat) descriptions

of a dree-dimensional world


Intelligence

Cascade

Typical challenges arising from taxonomies

•Multi occurences of product components at different

locations (in taxonomy)

•Relations between product components;

Depencies of topics on combinations of components

•Dependencies of additional variant properties between

each other and on product components (configuration

management)

•Dependencies of information types on other taxonomic

values


More Complexity (and Dimensions)

Relations and rules are

mostly given by

by product management

or by technology

Intelligence

Cascade

FunctionalDescription

Rotor

Display

Heating

X3B

T3B

ContentTopic

Safety

Repair

User Manual

Service Manual

IntrinsicTaxonomies

IntrinsicTaxonomies

ExtrinsicHierarchies

ExtrinsicHierarchies

Hierarchies, Taxonomies, List, …

Variant Features/Properties

Functional Metadata

Multidimensional Information Space including relations



Ontologies and semantics

as communication means

and knowledge

representation

Intelligence

Cascade• Purpose of Augmented Intelligence is to model the complexity of

real world products and information

•Overcome typical shortcomings of the

taxonomic modelling of metadata

• Introduce model of objects, their properties and (conditional)

relations between each other

as semantic network → Ontologies


Ontologies

•Ontologies represent a modelling technique of describing relevant

aspects of a real-world situation (business cases, information,

processes, …)

• The goal is to model all occuring objects in an abstract way as classes

and the real-world objects as their representations (indiv. instances)

•Relations between objects and their properties are modelled within the

ontology on a class level while further rules can apply (between

instances)

•Ontologies are closely related to computer science (OOP), terminology

management (conceptual systems), philosophy, …

Augmented

Intelligence


Augmented Intelligence: Appl. TypeAugmented

IntelligenceInformation

Topics

Product

Component

Modelling tool: ProtegeData by A, Ahmadpour MT HSKA 2019, W. Ziegler

Product

Functions


Augmenting CMS / CDP by Ontologies

Applications for


Augmented

Intelligence

CMS

CDP

Appl. Type I

Appl. Type II


Semantic Middleware (Information Hub)Augmented

Intelligence

CMS CDP

InformationServices

Appl. Type III

Applications for



Augmented

Intelligence

Application Type I (CMS-related)

CMS

CDP


Product model (attached to engineering) as (as far as possible/needed complete) model of

components, their relations, functions and properties

with respect to variants

Ontology Modelling (of PI-Fan) Augmented

Intelligence

© 2017 ONTOLIS GmbH

Source: Ontolis


Augmented

Intelligence

Product Model (Aumented Intelligence) as Ontology

© 2017 ONTOLIS GmbH

Quelle: Ontolis


Application Type I

• Allows planning of product variants in a systematic, rules-

based and visual way of ontology models

•Can be used as a lookup technology or can be integrated via

CMS interfaces

•Connects to product development, configuration

management, PLM and configuration software

• Empowers TC writers to handle product complexity

•Control/support CMS processes (in the future) by interfaces

Augmented

Intelligence


Appl. Type II (CDP-related)Augmented

Intelligence

CMS

CDP


Augmented

Intelligence

Ontology modelling (of PI-Fan)

Source: I-Views (iiRDS metadata class model and topics)


Application Type II

• Allows the modelling of logical connections between information

and topics available in a search system (CDP)

•Can be used as a visual lookup technology (delivery network)

•Can be integrated into CDP in order to improve search results by

logical relationships between information (similarity algorithms)

•Can be connected to enterprise search systems (search expansions)

• Topics/Information can be assigned manually to an ontology or

using AI technologies

Augmented

Intelligence


Application Type IIISemantic Middleware (Information Hub)

Connecting the enterprise information environment

by ontologies/semantic relationsSources can provide: labels, search indexes,taxonomies, ontologies,… data, documents, media,utterances (chats), … , social media content, …

Augmented

Intelligence

CMS CDP

InformationServices


Modelling of (Meta Data)-Mapping between content

from different sources (Sherlock, Fischer IT)

Intelligence

Cascade


Application Type III

• Semantic middleware (information hub) is connecting the enterprise

information environment and contained information

through semantic relations (ontologies)

• Independent of specific applications (CMS, CDP)

• Provide connectors/interfaces to different information systems and

databases

•Can import or manage (meta) data; rules-based operations (coll.)

• Provides (web-)services and interfaces to other applications in order

to benefit from a structured access to all data sources

Augmented

Intelligence


Standardizing Exchange by Ontologies

Content Delivery

Intelligence

Cascade

CMS

CDP

Standardized description and packaging of metadata and content.

Metadata are desciribed by using the formal ontology language RDF and

the logic of extended PI-classification.


Relationship between PI-Class and iiRDS

• The metadata concept of iiRDS (for topics)has been given

by and derived from PI-Class®. The PI-Fan is a reference

model for both.

• The metadada description of iiRDS is written in RDFs-

notation. (Standardized XML-notation)

•RDF is a light-weight ontology-description language, which

is not relevant for authors, CMS users.

• iiRDS uses explicit metadata values, given by standards

• PI-classified topics can be transformed or mapped to iiRDS

iiRDS

Intelligence

Cascade


Metadata Correspondence & Mapping (for Topics)Intelligence

Cascade

Rotor

Display

Heating

X3B

T3B

ContentTopic

Safety

Repair

Functional D.

User Manual

Service Manual

Component

TaxonomiesInformation

Class

Product

StructureDocument

Types


PI-Class (inner area)vs. iiRDS

See also:

https://iirds.tekom.de/fileadmin/iiRDS_specification/20180418-1.0-release/index.html#iirds-metadata-class-diagram

DocumentationMetadata→ProductMetadata→ProductVariant

DocumentationMetadata→ProductMetadata→Component

DocumentationMetadata→ProductMetadata→ProductFeature

DocumentationMetadata→FunctionalMetadata

DocumentationMetadata→ProductMetadata→ProducLifeCyclePhase

InformationType→

DocumentType

InformationType→

TopicType

InformationType→

InformationSubject

https://iirds.tekom.de/fileadmin/iiRDS_specification/20180418-1.0-release/index.html#iirds-metadata-class-diagram


The Role of Artificial Intelligence (AI)

Recent, possible and future applications

• automated text classification

• translation process (Machine Translation)

• language recognition/processing and dialog systems

(chatbots)

• visual object recognition

• content generation

• interaction analytics (data analytics, behaviour, context, …)

Intelligence

Cascade


Artificial

Intelligence

Where AI can help and is used

CMS

CDP

Auto Classifi-cation




Artificial

Intelligence

Where AI can be used in (future) industrial applications

CMS, …

CDP

xMS

xMS

xMS

CMSCMS

Supplier

Additional

Information

&

Sources

User Information

Service Information

Off Site

Web Portal / Mobile

Machine state

(errors, messages,

operating conditions)

CDP

On

Site

Online/

Offline

Contentgeneration

Machinetranslation

Language recognition

Interaction / gesture

analyses

Objectrecognition

Machineanalytics


Future Content Access

•microDocs

• System and user tracking


Retrieval Methods by Content User

Hypothesis & first experiences

• The more unexperienced, the more users will use direct search

(full text search)

•Document structures are often designed as print publications or

as a most complete set of information;

therefore, navigation by document structures (toc) will not be

the appropriate (online) retrieval structure

Future

Content Access


Retrieval Methods by Content User

Hypothesis & first experiences

• Facetted search relevant (only) for experienced users, e.g. service

technicians, product managers, production process, …

• Full facets derived form product (component) taxonomies are quite

complex to access;

taxonomies may be (more) relevant in the system background;

→ How can we benefit even more from facets/ taxonomies/classes

and/or ontologies?

Future

Content Access


Retrieval Structures by CDP Future

Content Access• Single-topic retrieval vs. classical document package retrieval

•Both can be implemented by package delivery

(iiRDS-packages or standard proprietary CDP import mechanisms)


Retrieval structures by CDP Future Content Access

Product-Class

Base/ Telescopic

Rod

X3B, X3-H1,X5-B, X5-D,…

Information-Class

Operation/Height

Adjustm.


Product-Class

Base/ Telescopic

Rod

Information-Class

Operation/Height

Adjustm.

Single topic Document(complete topic assembly)

X3B, X3-H1,X5-B, X5-D,…


Lack of context Abundance of contentWhat is needed?

microDocs

A structured set of topicswith relevant contextand sufficient content

Relevance and sufficencyist defined by use cases

„Delivery zwischen Kontext und Content"

technische kommunikation, Heft 6

S. 58-61 (2019)


microDocsFuture

Content AccessDefinition

A microDoc is a (sub-)set of topics

required by predefined use cases

and connected by a logical concept

as a dynamic publication in search media

Additional comment:

The logical concept, the relevant context and

the amount of required content can be derived

at different levels from semantic models.


microDocsFuture

Content AccessImplementation levels

I. Static documents aggregated on CMS level and packaged

for CDP

II. Dynamic topic aggregation / or filtering in CDP according to

predefined structures using taxonomies (including dynamic

linking)

III. Extraction and linking of content from predefined semantic

relations, rules and properties of ontology classes and/or

instances


microDocs

How topics can derived from

a logical structure/ semantic

network

Future

Content Access

A. Ahmadpour, W. Ziegler HSKA


microDocsFuture

Content AccessImplementation levels

IV. Extraction and linking of content from human-derived

semantic relations, rules and properties

(by search and access analysis)

→Web-Analytics (CoReAn) of retrieval processes

V. Extraction and linking of content from machine-learned

(AI) semantic relations, rules and properties

(by search and analysis)

→ „Predictive Content“


microDocs as MicroServices

•Digital content & information services will be offered for and with

products as selling points

•Companies might offer information services at different levels

(regarding service level, information depth, variant specificity of

parameters )

•microDocs might be provided dynamically and web-based for

different target groups, use cases and other content-consuming

IT/information services

Future

Content Access


Ontology-based (Type I)

Fa. Ontolis

microDoc (Draft)


Rules set in CDP

Fa Expert Comm. Systems

microDoc (Draft)


Classification-based Rules for

Linking

Docufy

microDoc (Draft)


Delivery of TC content: Digital Content Service

microDocs can be includedin Content Services supportingDigital Information Services

CD Methods

CMS, … CDP

CDP

OnSite

TC

DCS

DigitalInformationServices


Use Cases/Processes:• Publishing• Download (Download Portal)• Retrieval + Viewing (CDP)• Dyn. Publication (Config, microDocs)• Download (Dyn. Publication)

Digital Archiving Services & CMS/CDP

CDP-related archiving relies

on tracking/analytics of user

behaviour and of dynamics of

information delivery

(dynamic publications)

Future User AccessProof Levels:• Publication• Download• Deployment• Viewing• Search/Retrieval• Aggregation, Dynamic Pub.

CMS CDP


Semantic Technologies in TC

•Content delivery and corresponding digital information services (DIS)

benefit from native and augmented intelligence of topic-based

information

•Augmented intelligence relies on semantic technologies for products

and corresponding internal and external information

• Semantic systems can empower

–information managers and writers to cope with product complexity

–information systems to deliver most relevant and specific information

to users (e.g. by microDocs and other services like archiving)

Summary


Semantic Technologies in TC

•Augmented intelligence relies on native intelligence

(classification concepts) used for example in CMS

• Semantic systems using augmented intelligence can empower

– Companies with their various departments to interact based on

explicit models (of products and information, functionality etc.)

– Information managers and writers to cope with product complexity

– Information systems to deliver most relevant and specific information

to users (e.g. by microDocs and other services like archiving)

Summary I


[email protected]

[email protected]

Institute for Information and Content Management

Thank you for your attention!