Presentazione di PowerPoint

Download Report

Transcript Presentazione di PowerPoint

SCEF- 2003 International Workshop 30 Sep., 1 Oct. 2003 , Rome

http://erg4146.casaccia.enea.it/SCEF

Socio-Cognitive Engineering Foundations and Applications: from Humans to Nations

(

an introduction

) Adam Maria Gadomski High-Intelligence & Decision Research Group, CAMO, ENEA & Sc. Board of ECONA, Italy

[email protected]

http:// erg4146.casaccia.enea.it/

This position paper will be available in the Proceedings of the SCEF-2003 Workshop

SCEF-2003 International Workshop

http://erg4146.casaccia.enea.it/SCEF/

HID

Presentation outline

Foundations, Paradigms, Context SCE Domains, Activities and Strategy Complexity & Objectives Methodology: TOGA Intelligent Organization, Roles & Decision-Making Technology: Intelligent Decision Support Systems Conclusions

About “Foundations”

Foundations

is a basic theoretical framework of a research or engineering field.

It includes: - initial assumptions/axioms/paradigms - its conceptualization tools: methods, methodologies - basic ontology of the domain of interest.

Its objective is to provide tools for computational models development.

Computational model

: formal model which may be implemented on computer and enables computer simulations.

HID

About “ Paradigms ”

Paradigms

are arbitrary chosen.

They are invariant rules/laws which are defined on a highest abstraction/generalization level.

They are either ontological assumptions or methods employed in the domain

One of the key difficulties related to the acceptation of paradigms are their choice criteria (usually implicit).

HID

About “Human Errors”

Human error

: Human action or inaction that can produce unintended results (*) or system failures (**).

(*) [ISO/ ITC Information Technology Vocabulary,96] (**) [ NUREC-1624]

Machine failures Human errors

Complex consequences interrelations

[Gadomski, 2002 ]

Reliability problems Safety problems

HID

Socio-Cognitive Engineering Context:

3 Generation Research

SCE belongs to third research generation in the human culture.

These generations are distinguishable by development: the following specific techno-scientific

First Generation

specialization approach

;

incremental grown

of subject oriented sciences and technologies. They are well isolated and self-limited by: their language (conceptualization systems), observation/measurement tools and engineering approaches.

Second Generation

interdysciplinary approach

;

autonomous cooperation

between different branches of research caused by common interests and by the tentative of an unification of their objectives-oriented and interface terminology.

Third Generation

over-disciplinary approach

;

building new common perspectives

, shared top conceptualization and ontology (redefinition of basic terms from a higher more abstract/universal perspective), in such way that they become valid for many, before separated research fields (something similar to the unification in physics).

At present At a consequence, in 3 rd technology development process

generation of research, the integration of science and

should be seen as

one parallel goal-driven incremental

About SCE Intervention Domain

Domain

of SCE are systems/networks of interacting humans and human like intelligent entities.

human

humans system communication link

s/h technologies robot can be substituted by

R HID H HS HS H

Intelligent Web - individuals - organizations - associations - communities - society - nations

We use an object-based framework for initial conceptualization of any Intelligence-based systems

An Example of ABSTRACT VIEW OF INTERACTING NETWORKS

[ From Sandro Bologna presentation,2003] PHYSICAL BIOLOGICAL TECHNOLOGICAL ORGANIZATIONAL

H

ZOOM

Complex systems

Made of many non-identical

elements

connected by diverse

interactions

.

NETWORK

ELECTRICAL SYSTEM NETWORK

Fonte: Corriere della Sera 30.09.03

About Intervention Domain

Human Errors

Socio-Cognitive Engineering application for

Multi-grid Large Complex Critical Systems/ Infrastructures

(LCCI) (such as electricity, telecommunication, gas networks)

Human component Artificial Highly-Autonomous

(Intelligent Agent)

component

for

Decision-support systems Production/Transmission /Control component

of Physical &Technological Layers

Technological Grid

Copyright

: High-Intelligence & Decision Research Group, CAMO, ENEA, http://erg4146.casaccia.enea.it

Organisation Network

Author: Adam Maria Gadomski , 28/09/2003

Human Factors

About Intervention ACTIVITIES

SCE

contributes to the Vulnerability Analysis and to the Improvement of Robustness of Large Complex Critical Systems

Systems)

(Humans-Technology

Human Errors Social Consequences

Key Intervention Activities

• Users/human Modelling and Simulation • Organization Structures and Decision-Making Modelling and Simulation • Assessment of Social Risk and Impacts • Intrusions and Mismanagement

Development and Simulation of

Autonomous

Artificial Intelligent Organizations

embedded in

Complex Human-Technology Systems

. ©

Copyright

: High-Intelligence & Decision Research Group, CAMO,ENEA, http://erg4146.casaccia.enea.it

Adam M. Gadomski , 28/09/2003

About

Strategy

of Socio-Cognitive Engineering

Socio-Cognitive Engineering

takes under consideration the interests and points of view of: citizens, employers, managers, owners & politicians SCE Integrated Strategy is

human-centered and technology-based Identification of the System of Interest and its contexts Design of System Modification System Validation and Design of self- regulation Management Strategy

Copyright

: High-Intelligence & Decision Research Group, CAMO,ENEA, http://erg4146.casaccia.enea.it

Adam M. Gadomski , 27/09/2003

SOCIO-COGNITIVE ENGINEERING

Integrated Approach

Complexity Domains: Sustainable Development Strategic Factors

(Application of the TOGA Methodology covers here computational modelling task)

Technology Barrier

Technology is nothing without Competences

Knowledge Barrier

Competences are nothing without Motivation Management

Cognitive Barrier Organizational Barrier

Management is inefficient under not adequate Organizational Constrains.

Cultural Barrier

All above are nothing if Socio-Cultural Context are neglected

Adam M. Gadomski, 28/09/2003

SCE

Problem of Real-word Complexity

Different Interrelations Different Study Directions Different Perspectives Different Dependences © Adam Maria Gadomski, http:// erg4146.casaccia.enea.it/ High-Intelligence & Decision Research Group Different Methods

Domain

All in Human-Technology Systems is Complex

Tools

: Conceptualization,Methods, Methodologies

Management TOOLS DOMAIN MANAGEMENT

Complex Activites Complex Complex Complex Context Complex

Models of Socio Cognitive Systems

About SCE Complexity

Complexity in SCE is not only a physical complexity but it includes complexity of mental processes and actions of an intelligent entity.

SCE complexity includes new attributes, such as :

Vagueness, Uncertainty Conflicts, Incomplete knowledge, Variable access to information, Emotions, Irrationality, Ethical preferences, Organizational & Socio-cultural factors.

HID

All of them influence Decisional Processes

SOCIO-COGNITIVE ENGINEERING:

Objectives

Numerous improvements of real Socio-Cognitive Systems (SCS)

on the levels of

efficacy of

and

interactions between

their components

( defined before).

Examples of problems:

- Interaction between individuals and always more complex information and business society, Efficiency and “life cycle” of human organizations, - Relation between decision-making and organization structures - Diagnosis of pathologies of human organizations - Individual Interest and Organization Interest impacts - Strategies of the development: democratic, centralized - Technological Support and Intelligent Artifacts © Copyuright: High-Intelligence & Decision Research Group, CAMO, ENEA , http://erg4146.casaccia.enea.it

Adam M. Gadomski, 28/09/2003

SOCIO-COGNITIVE ENGINEERING for LCCIs

(Large Complex Critical Infrastructures)

Socio-Cognitive Engineering

takes under consideration the interests and points of view of

owners, operators and customers

of LCCIs

: +

LCCIs customers need the reliability and continuous providing of the services as long as possible and at low cost as possible.

LCCIs operators wish to be well informed about the infrastructure state and require its efficient management to satisfy customers expectations

LCCIs owners are focused on the socio-economic aspects of LCCIs.

ENEA’s Competences

Modelling Metodology User & Decision-Maker Models & Architecture TOGA

Copyright

: High-Intelligence & Decision Research Group, CAMO,ENEA, http://erg4146.casaccia.enea.it

Intelligent Organisation Modelling & Simulation

Adam M. Gadomski , 27/09/2003

About

Methodology

How is possible to cope with so complex domain and objectives?

Top-down Object-Based Goal-oriented Approach (TOGA)

TOGA is goal-oriented complex-knowledge ordering computational tool. It assumes the top-down

observation metaphor

, to see complex problems

from a bird eye’s view

; this means to first identify a problem’s most general context constraints which remain always true and mandatory for every successive level of its specification (“fleshing out”).

It is based on formal step-by-step decomposition of the relation: Intelligent Entity Environment

Adam M. Gadomski , 28/09/2003

SCE has to use new thinking methods

TOGA METHODS

The methods are divided on: 1. New Methodology/method of

Cooperation

between a SCE project partners 2. New Methodology/method for

Objectives Achieving

, where the Cooperation Method is focused only on the efficacy of the realization of Objectives Achieving Methodology.

They both, in different proportions, are based on parallel, top-down and goal-oriented application of

main paradigms of physics, systemics, cognitive and social sciences related to a generic intelligence

. The meta-theoretical approach TOGA is assumed as a initial methodological and ontological framework. [ see References].

The top methodology includes in parallel, top-down goal-oriented tools development and their subsequent applications.

Examples

Adam M. Gadomski , 28/09/2003

SCE utility

Examples of human/organizational errors

• US Blackout 2003

“The initiating events appear to have happened under the lazy

eyes of a mismanaged utility, but underlying conditions made a massive U.S. power failure almost inevitable”

[IEEE Spectrum, 23 Sep.2003, Special Report.] • Italia Energetic Blackout 2003.

• Cernobyl Nuclear Disaster, Apr. 26, 1986.

•… • ENEA commissioning, Italy, 2002 © Copyright High-Intelligence & Decision Research Group, CAMO, ENEA , http://erg4146.casaccia.enea.it

Adam M. Gadomski, 28/09/2003

SOCIO-COGNITIVE ENGINEERING (SCE):

Humans Modelling

Application of TOGA (Top down Object-based Goal oriented Approach)

Modelling Frameworks

P I IPK Computational Model

(Information, Preferences, Knowledge)

Problem Specifications

are: Requested & Modified

Information K Models are Knowledge Motivations

create proper

Preferences

which activate adequate Knowledge

Human ERRORs :

Not proper or not sufficient

Information

Lack or not proper Importance Scale

(Preferences, risk ass.)

Not proper or not sufficient instructions, procedures (

Knowledge

) Wrong Cognitive and Organizational Factors (

Motivations

Adam M. Gadomski , 28/09/2003

Disaster Managers: simple model example Example

Agent 2

P K I – information system P – preferences system K – knowledge system

Agent 1

K P I 2 I 1 Infrastructure Network

Real Emergency Domain Agent 3

K I 3 P P I n I K P

Agent Manager Agent N

C Many places for human and organizational errors can be evidenced.

Adam M. Gadomski , 28/09/2003

SCE perspective on a Large Research Center (LRC)

Example

Factors: Top-down Identification

• Research Paradigms System • Research Means • Support Technological Means • Research Results/Products • Socio-Business Laws • Socio-Cognitive Context

Components:

• Human components • Technological components • Informational components • Economical components • Political components In general, from socio-cognitive perspective: it is necessary

to see how a decision-maker sees his domain of activity

Adam M. Gadomski , 28/09/2003

SCE:

Large Research Center (LRC)

Example Application of LRC’s Competences IPK Computational Model Project Specifications

are: Requested & Produced

Information Motivations

create proper

Preferences

which activate adequate Competences Competences LRC’s are

Knowledge Application of SCE Theorem to LRC

( “light” version)

LRC ’s Competences are useful if and only if the motivations of human Intelligent Agents will be developed and properly managed

Human ERROR

(on 1st metalevel):

If motivation-management is not adequate then competences are not activated or corrupted.

Adam M. Gadomski , 28/09/2003

SOCIO-COGNITIVE ENGINEERING

:

Intelligent Organization

General Functional Frame

Mission/Fundation-Goal

Organization

Products

TOGA theory framework

Intelligent Organization

is defined by reciprocally dependent roles of interacting/communicating intelligent subjects which should act in order to archive common goal ( usually defined in the organization statute ).

Organization is specified by the set of

roles

, its

structure

,

decisional mechanisms

and

resources

.

Adam M. Gadomski , 28/09/2003

Dynamic Role model

(computational)

Definitions according TOGA

Role

( competences, duties, privileges )

Competences

: what he/she/it is able to do, possessed models of the domain (

knowledge

)

Duties

: responsibility, tasks and requested

preferences Privileges

: Access to the information. It produces conceptual images of the domain. Access to execution tools (

information

Every role is specified by its own IPK Bases Set: Information Bases Preferences Bases Knowledge Bases

– how situation looks, continuously updated – importance scales/relations, ethics rules – required models & know how Copyright High-Intelligence & Decision Research Group, CAMO, ENEA , http://erg4146.casaccia.enea.it

Adam M. Gadomski , 28/09/2003

Universal Management Paradigm (UMP) Pattern-based Structure: Subjective, Incremental, Recursive

H-INTERFACE

SUPERVISOR

Knowledge Preferences

ADVISOR

expertises tasks

MANAGER

information cooperation

COOPERATING MANAGER

information tasks

INFORMER

IN/EX H -INTERFACE

EXECUTOR UMP includes 6 canonical roles and their interrelations HID

Pathologies of Organizations: Examples

Every human-agent is in 3 roles together :

Organizational role

– requested/defined by the structure (fixed) 2.

Informal role

– applied, structure independent (variable) 3.

Personal/real role

– really realized (variable)

Conflicts of Roles

Dynamics of roles creates lack of congruence between them & conflict of interests Compromise, inefficient, risky decisions

Conflict of Interests/Motivations

Differ

Risk-Benefits

HID

relation for

Social interest Organization interest Personal interest © Adam Maria Gadomski, http:// erg4146.casaccia.enea.it/ High-Intelligence & Decision Research Group

Decision-Making

New Information Knowledge Base

Decision-Making

Preferences Base No action/response Meta-action/Pseudo-action Action adequate to D M’er role and situation

Definitions

[TOGA]

Decision-making :

an individual or group reasoning activity/process implied by the request/necessity of a choice caused by received information or task, or by delivered conclusion about possibility of risks/benefits. It is started when either choice criteria are unknown or alternatives are unknown and finished when choice is performed.

Action-oriented decision-making :

it is a decisional process when alternatives represent possible actions in pre-chosen physical domain.

Mental decision-making :

when the final choice refers not to actions but to conceptual objects related to a preselected domain of activity of intelligent agent.

Group decision-making

: when responsibility for decision is allocated to a group of intelligent agents and is based on shared decision-making process.

Adam M. Gadomski , 28/09/2003

Decision-Making

(computational models)

alternatives reasoning path

?

critical node decision

?

data

d-m

decision

Types of Proper and Pathological Decisions Main classes

: - meta-D-M , - pseudo D-M, -

proper D-M.

Pathologies are related to :

response on source type

( “safety” filters ); -

response on subject

( lack of competences, emotions, out of Interest).

response according domain-preferences

(organizational role): proper D-M.

Controlability & updating of Ethics concept

D-M autonomy increases

then:

Efficacy of Control decreases

Importance of Ethics and personal motivation increases

. This rule indicate importance of

Motivation Management.

Adam M. Gadomski , 28/09/2003

SCE Ontological Tools -TOGA

TOGA provides tools which could be used for Identification/Specification of real-world problems:

Complex domain: SPG Modelling framework Complex interventions: WAG Modelling framework Risky decisions: Risk-based Reasoning Model

Intelligent entity modelling with Human Factors , such as: Emotions Irrationalities Motivations HID Fractal-like Multi- and Meta- Modeling, and Simulations tools are required.

Technology support: IDSS

SOCIO-COGNITIVE ENGINEERING (SCE) :

Intelligent Supports

Reinforcement of the LCCI network by

Internal Artificial Intelligent Agents Organization

Reinforcement of Human Organization by

Intelligent Decision Support Agents’ Grid

Higher Infrastructure Network Autonomy Better Human Control and Supervision Selected Infrastructure Infrastructure Dependencies Copy rights High-Intelligence & Decision Research Group, CAMO, ENEA , http://erg4146.casaccia.enea.it

Adam M. Gadomski , 8/10/2003

IDSS: Intelligent Decision Support Systems

What is it ?

“Software program that integrates human intellectual and computer capacities to improve decision making quality, in semi structured problems situations” [Keen, Scott-Morton, 1996]

DSS IDSS

Provides passive

Informational

Aid and Toolkits Provides active, partially autonomous

Decisional

which involve human-like computational intelligence.

Aid

When IDSS is important?

•

amount of information

necessary for the management is so large, or its time density is so high, that the probability of human errors under time constrains is not negligible.

• coping with

unexpected situation

requires remembering, mental elaboration and immediate application of complex professional knowledge, which if not properly used,

causes fault decisions

More information: http://erg4146.casaccia.enea.it

Adam M. Gadomski , 28/09/2003

INTELLIGENT DECISION SUPPORT FOR RESEARCH MANAGEMENT EXAMPLE

Situation Assessment & Decision Making is based on:

Information

: DOMAIN status,

Knowledge

: rules, procedures, instructions,

Preferences:

role criteria

risk criteria, resources criteria,...

Research Activities

Continuous monitoring

Administrative Activities

Actions

Financial & Decisional requests Periodical monitoring

Strategic INTELLIGENT DECISION SUPPORT SYSTEM

Information

Strategic Activities

MIND

Tasks

link to computer networks

Adam M. Gadomski , 28/09/2003

Conclusions

Nowadays SCE is a response on dramatically grown risk of negative consequences of Human Errors , it is inevitable tool of XXI Century.

Complexity of problems requires new 3 rd Generation Approaches, such as TOGA multi-factor problem representation and parallel modelling, and IDSS development.

- Key problems refer to the understanding and transparency of decision-making processes for their intelligent actors-contributors.

- Socio-Cognitive Engineering requires new specialists on organization, national and international levels.

- EU promotes assessment of possible socio-cognitive impacts, innovation governance and new updated roles for policymakers. - EC coordinates cross-integrations of national initiatives with objective of parallel harmonic and sustainable development of science, technology and society.

The above mentioned tasks have to be supported by theoretical foundations and in consequence, by conscious, wise and socio-ethical responsible decision-making.

HID

Conclusions

Futurology: extrapolations

Future Grow in Arbitrary Units 500 400 300 200 100 0

Physics & Energy Chemistry & Bioengin.

Social & Knowledge Engin.

Current name is

Socio Cognitive Engineering

Extrapolation of the current trends in three basic macro-engineering domains. [US Sources, DARPA, Web, 2000]

HID

Adam Maria Gadomski, http:// erg4146.casaccia.enea.it/ High-Intelligence & Decision Research Group

References

For more information see

:

1 2 A.M. Gadomski

, Lectures on Safety and Reliability of Human-Machine Systems

. Materials of SA-EUNET EU Project, 1998.

A.M. Gadomski

, SOPHOCLES Project – Cyber Virtual Enterprise for Complex Systems Engineering: Cognitive Intelligent Interactions Manager for Advanced e-Design,

Transparent-sheets, 28/08/2001, ENEA. ITEA.

4 A.M.Gadomski.

TOGA: A Methodological and Conceptual Pattern for modeling of Abstract Intelligent Agent

.Proceedings of the "First International Round-Table on Abstract Intelligent Agent". A.M. Gadomski (editor), 25-27 Gen., Rome, 1993, Publisher ENEA, Feb.1994.

6. A.M.Gadomski, "

The Nature of Intelligent Decision Support Systems

". The key paper of the Workshop on "Intelligent Decision Support Systems for Emergency Management ", Halden, 20th-21st October, 1997.

A.M.Gadomski, S. Bologna, G.Di Costanzo, A.Perini, M. Schaerf.

Towards Intelligent Decision Support Systems for Emergency Managers

: The IDA Approach. International Journal of Risk Assessment and Management, 2001. A.M.Gadomski, A.Straszak.

Socio-Cognitive Engineering Paradigms for Business Intelligence Modelling: the TOGA conceptualization.

Proceedings of the 5th Business Information System International Conference – BIS 2002, Poznan, Poland, April 24-25, 2002.

9. A.M.Gadomski,

Socio-Cognitive Scenarios for Business Intelligence Reinforcement: TOGA Approach,

The paper preliminary accepted for publicatiin in "International Quarterly of Cognitive Science “, 2003.

HID

Adam Maria Gadomski, http:// erg4146.casaccia.enea.it/ High-Intelligence & Decision Research Group