BIAS FAST ANIPLA INTERNATIONAL CONFERENCE - AUTOMATION WITHIN GLOBAL SCENARIOS, Milan Fair Quarters, 19-20-21 November 2002 Socio-Cognitive Engineering Scenarios for the Reinforcement of Global Business Intelligence: TOGA Approach Context Problem Method Model Solution Example Adam Maria Gadomski, High Intelligence & Decision Research Group, ENEA, Italy http://erg4146.casaccia.enea.it Copyright ENEA, A.M.Gadomski, 2002
Socio-Cognitive Engineering Scenarios for the Reinforcement of Global Business Intelligence Presentation Outline Context Problem Method Model Solution Examples Context Global Business Intelligence What is it? TOGA Methodology Socio-Cognitive Scenario Socio-Cognitive BI Modeling Development of Highly Intelligent Systems (Meta-systemic perspectives ) Current ENEA Research Copyright A.M.Gadomski, HID, ENEA 2002
Socio-Cognitive Engineering Scenarios for the Reinforcement of Global Business Intelligence Research Activity Context Context Problem Method Model Solution Examples The study has been performed in frame of: Scientific Collaboration CIMA (Computer Intelligent Managerial Advisor) between: - High-Intelligence & Decision Research Group, ENEA - Institute of Systems Research (IBS-PAN), Poland Scientific Collaboration HAM ( Human cognitive Agent Modeling) between: - High-Intelligence & Decision Research Group, ENEA - Interuniversity Center for Cognitive Processes in Natural and Artificial Systems (ECONA), Italy The work has also been partially supported by the EUREKA ITEA Project SOPHOCLES (http://erg4146.casaccia.enea.it/sophocles) The activity is focused on the modeling and reinforcement of Business/Organizational Intelligence. Copyright A.M.Gadomski, HID, ENEA, 2002
Socio-Cognitive Engineering Scenarios for the Reinforcement of Global Business Intelligence Automation Context Essence of Automation: Context Problem Method Model Solution Examples - Transition from human intelligence driven actions to formal models and action algorithms. - Reinforcement of human physical capabilities. - Substitution of humans senses and effectors by machines. New Automation Hypothesis: - Substitution of physical by mental - Substitution of humans senses and effectors by human reasoning.
Problem: Objective of the presentation To illustrate how the application of a modern systemic methodology called TOGA (Top-down Object-based Goal-oriented Approach), + socio-cognitive engineering paradigms, + artificial intelligence and + software technologies, may enable modeling, construction and reinforcement of business intelligence.
Global Business Intelligence What is it? Consensus building: preliminary definitions Business a class of human goal-oriented activities where their products have values in the one commonly accepted well ordered scale (in a normed space). Intelligence a complex mental capacity which is visible as an efficacy in unexpected and uncertain circumstances. Global Bussines Intelligence (GBI) is a behavioral (observable) intelligence independent on the type and level of the business activity. It also includes managerial and organizational intelligence. GBI can be seen as an individual and group property. It can be decomposed and allocated to humans and computers. In general, the reinforcement of human GBI requires its generic functional/processual model and the recognition of its well observed properties.
Why Reinforce Global Business Intelligence? Management of not routine events Complexity of the real world, Numerous disasters, Complexity of management tasks which excess human capacities. Human decisional errors and loss of efficacy. Known Solution: Automation of routine tasks Weakness of KS: - Difficulty of Identification of human tasks - Low Usability of Automation - Lack of Human Factor in the design Classical engineering approach: To adopt humans to machine failured in the case of high-risk systems and complex tasks. New Solution: Intelligence-Centered Modelling
Method: Socio-Cognitive Engineering Approach Socio-Cognitive Engineering: New emerging crossdisciplinary approach which integrates: psychology, sociology, engineering and systemic perspectives on large complex realworld aggregates : Society- Human- Organization-- Technology- Environment (SHOTE) Systemic Perspective HO H CSS AD ENV Elementary heterogenious unit in the modern systemic sociocognitive approach H - Human, CSS - Computer Support Systems HO - Human Organization, AD - Domain of Activity ENV - Environment
Information about TOGA Meta-theory TOGA: Top-down Object-based Goal-oriented Approach. It is a systemic socio-cognitive metatheory framework. TOGA provides an ontology and conceptualization rules for a goal-oriented knowledge ordering. Its main structure has been proposed and developed in ENEA since 1988 (A.M.Gadomski). The key TOGA assumptions/axioms are divided on: Conceptualization, Ontological and Methodological Meta-methodological Assumption TOGA is a goal-oriented conceptualization tool focused on the maximalization of utility, applicability and trust but not on an absolute true discovering. It is axioms-based self-growing and self-referred (it is applicable to itself).
TOGA Meta-theory Everything said is said by an observer'. (Maturana & Varela, 1980) Conceptualization Assumptions (How to express?): Everything is formally conceptualizable in frame of Worlds of Abstract Objects composed with abstract objects, relations and changes. Ontological Assumptions (What exists?) Every identification/design problem involves an intelligent subject and its domain of activity (problemdomain). Existence is a relative concept, exists only what is conceptualizable and usable for intelligent entities. Real-world is a source of quasi-infinite number of data.
TOGA Assumptions Methodological Assumptions (How to be efficient?) 1. Paradigms: TOGA employs systemic, physics, sociocognitive and engineering paradigms 2. Only Top-down and Goal-oriented Approach enables complete and congruent conceptualization of real-world problems. 3. Simplest theory is always considered as working true if it has the same utility for the same goal. 4. Bottom-up acquisition of information and knowledge is governed by top-down rules.
TOGA Structure TOGA is composed of: TAO: Theory of Abstract Object KNOCS: Knowledge Conceptualization System MRUS: Methodology and Rules System 1 2 ABSTRACT INTELLIGENT AGENT ENVIRONMENT 3 DOMAIN OF ACTIVITY Copyright A.M.Gadomski, HID, ENEA, 2002
Abstract Intelligent Agent: IPK Essential TOGA concepts employed in the modeling of abstract intelligent agent (AIA) : information, preferences and knowledge (IPK). They are well defined and independent - Information - - How situation looks - Past/Present/Future states of Domain-of-Activity (D-o-A) - Preferences - - A partial ordering of possible states of D-o-A which determines what is more important (A is better then B) - Knowledge - - What agent is able to associate (descriptive knowledge: rules, models) - What agent is able to do in Domain-of- Activity (operational knowledge)
IPK Process Basic elements of reasoning process: Information D = Knowledge D (Information D ), where Information represents a state (S) of a domain D. Preference: IF (IF S2 is better than SX) than Knowledge D : = K (S1 S2) Perception information Preferences system Action Abstract d-o-a information intervention goal Knowledge system
Meta-thinking: Thinking about Thinking Personoid model (an example): Different points of view and meta-levels. Model: Iterative, recursive
Decision-Making Process Main element of intelligent agents mental activity that causes goal-directed human behavior is a decisionmaking process. In the IPK conceptualization, abstract IPK bases are basic necessary carriers of decisional process (D-M), and D-M can be represented as follows: I = Complex_Choice_Operator [I,P,K] I, where I is an information which activates D-M, and New Information I is an information which includes decision. Knowledge Base Decision-Making Preferences Base No action (emergency end) Action adequate to DM er role and D-domain state
Decision-Making Process Role (competences, duties, privilliges ) Competences: what he is able to do, possessed models of the domain (knowledge) Duties: tasks and requested preferences Privilliges: Access to the information. It produces conceptual images of the domain. Access to execution tools (information). An advantage of the personoid model is its applicability to natural and artificial intelligent agents, such as: intelligent unit, distributed organization, corporate systems with human and technological components.
Socio-Cognitive Scenario Cognitive processes in social contexts are employed in business-oriented decision-making Therefore Socio-Cognitive Scenario requires an integration of : -Engineering perspective, in which IPK is applied to the construction of useful artifacts - this leads to the questions of the utility of intelligent problem solving; -Psychological perspective, where a systemic and ecological knowledge is involved in mental processes. This approach tends to explain, forecast and modify human intelligent behaviors. -Sociological perspective; organizational D-M.
Socio-Cognitive Scenario Informal representation of the business world where the sociocognitive IPK and socio-cognitive engineering are taken under consideration.
Socio-Cognitive Business Intelligence Modeling : the UMP of TOGA TOGA Universal Management Paradigm (UMP) SUPERVISOR tasks information Knowledge Preferences ADVISOR expertises information INFORMER MANAGER MANAGER IN/EX H-INTERFACE IPK1 cooperation tasks EXECUTOR COOPERATING MANAGER IPK2 TOGA repetitive dynamic functional frame for every intelligent centralized organization (with subjective roles of intelligent agents)
Socio-Cognitive Business Intelligence Modeling Abstract Intelligent Agent: Role model Causes of Decisional Errors Knowledge Competencies Out of competencies Preferences Responsibilities, Duties Wrong choice criteria Information Access to information Not proper or insufficient information
Business Intelligence Reinforcement Strategy of Business Intelligence Reinforcement is focused on: Understanding Socio-Cognitive context and model of Intelligent Decision-Making, Developing of conscious meta-reasoning mechanisms Capacity building for the management of the IPK support Development of High-Intelligent IPK-based support GBI dramatically depends not only on managers mental IPK and on a IPK support but on: organizational intelligence which includes corporate human-computer intelligence. Copyright ENEA, A.M.Gadomski, 2002
Computer Technology for Reinforcement of BI Context Problem Method Model Solution Examples They support: - Large Data Bases (information providing) - Information & Communication Networks (for IPK) - Events Simulations (What-if) - Decision Support Systems (I, passive, toolkit) - Intelligent DSS - cognitive modeling (knowledge, preferences providing) - Intelligent Tutoring Systems (business games) capacity building, skill improvement and real-time business management.? Copyrights ENEA, Adam M. Gadomski High-Intelligence & Decisions Research Group
High-Intelligence: Anticipatory Research are Today Research Ray Kurzwiel Is it my or his idea? Basic Properties of High-Intelligence: - capacity to use available IPK for goals achieving. - capacity of goals modification according to the new information - capacity of self-learning and meta-reasoning. - emotional component Banks, Industry Policymakers Servicies Society Family MATRIX Solution
Examples: ENEA Research ILLUSTRATIVE ENEA s EXAMPLES High-Intelligence & Decision Research
Examples: ENEA Research Strategy of the High-Intelligence & Decision Research Group MUSTER: Tutoring System for Emergency Management. - Study of the IPK utility. IDA Project - Intelligent Decision Advisor for Emergency Management Decision Process Modeling. SOPHOCLES Project - Development of the highintelligent kernel for Intelligent Cognitive Advisor System level development Platform based on HeterOgeneous models and Concurrent LanguagEs for System applications implementation for System on the Chip. SAFEGUARD Project Intelligent Agent Organization for the protection of Large Critical Complex Infrastructures Problems of roles-decisions-org.structure.
MUSTER Project: Computer-Aided Cooperation Training for Disaster-Managers Role-oriented specific scenario 1 Recognition of errors and explanation (AIA-model based ) TUTOR General emergency scenario & control of training AGENT 1 Role-oriented specific scenario 2 communication MUSTER System - Emergency Domain Models - Intervention Scenarios - Management Strategies - Agents Role-Models - Cooperation Models (multi-agent) - Training Strategies amg AGENT 2 AGENT N Role-oriented specific scenario N Disaster Managers no communication
MUSTER Project: Different perspectives of Disaster Managers Agent 1 K P I 2 Agent 2 P K I 1 Infrastructure Network Real Emergency Domain Agent 3 I 3 K P I K P I information system P preferences system P I n Agent N C.... Agent Manager K knowledge system Copyright ENEA, A.M.Gadomski, 2002
Lecture on Safety and Reliability of Human-Machine Systems Adam M.Gadomski MONDO O SIMULATORE ESTERNO AGENTE SEMPLICE CONSIGLIERE DIRETTO SISTEMA DOMINIO Multi-Agent Structure of Abstract Intelligent Agent SISTEMA PREFERENZE SISTEMA CONOSCENZA AGENTE SEMPLICE GESTORE PREFERENZE S. RAPRESENTAZIONE DELLE PREFERENZE S. RAPRESENTAZIONE DELLA COMOSCENZA AGENTE SEMPLICE PIANIFICATORE S. META -PREFERENZE STRATEGIE CAMBIO PREFERENZE CRITERI COSTRUZIONE PIANI S. META-CONOSCENZA METODI DI PIANIFICAZIONE Copyright ENEA, A.M.Gadomski, 2002
Copyright ENEA, A.M.Gadomski, 2002 SOPHOCLES -DEMO
TOGA Universal Management Paradigm (UMP) Repetitive functional structure SUPERVISOR tasks information ADVISOR Requests/ expertises MANAGER MANAGER Requests/ supports COOPERATING MANAGER information tasks INFORMER EXECUTOR observations actions INTERVENTION DOMAIN SAFEGUARD Project: Artificial Intelligent Agents Organization ENEA, A.M.Gadomski
Examples: ENEA Research More information is available on: http://erg4146.casaccia.enea.it Thank you.