Knowledge Flux for Measuring the Intensity of the Ubiquitously

Flowing Intelligence using Denotational Mathematics

JOHN SARIVOUGIOUKAS1, ARISTIDES VAGELATOS2
1
IT Department, 2Continuous Education & Training
1
General Hospital of Athens, 2CTI
Athens
GREECE
1
[email protected], [email protected]

Abstract: - The formally defined mathematical framework of Denotational Mathematics, provides the opportunity
to use existing mathematical tools to define the dimensionless magnitude of knowledge flow. In this paper,
“knowledge flux” is defined in order to provide a quantified expression of the developing knowledge with which
the knowledgeable control of the physical space can be achieved. Moreover, the data flux can be used as a
measure of the interaction between the processes carried in the physical and the corresponding virtual domain.
In addition, the data flux can be used to characterize the efficiency of the symbiotic partner agent software which
facilitates the interaction between the physical and virtual spaces. As a result, the indicative interaction between
the physical and the virtual spaces can be quantified and measured with objective evidence.

Key-Words: Knowledge flux, Context-awareness, Prerequisites & challenges, Knowledge evolvement, Static &
Dynamic Representation of Knowledge, Process & dynamic effects, Dynamic knowledge

1 Introduction developed static and dynamic knowledge along with

The Weiser’s [41] seminal research work placed the applying rules and user preferences [10]. Also,
foundations for an alternative scientific vision using ontology-based knowledge facilitates the inference
the term “ubiquitous computing”. Since then, the process with the development of semantic
technological advancements of internet and mobile frameworks [15] employing specialized taxonomies
devices extended the scientific research frontiers implemented from instantiated classes expressing
beyond the local applications and the entire globe functional properties and data attributes. The
participates in a wider network referred as the employed classes represent the static aspects of the
Internet of Things (IoT). Weiser’s vision led to smart developed knowledge while object orientation
environments handling the developing knowledge to provides the dynamic aspects of knowledge evolution
support decisions making with context-aware [3] over time. Knowledge is treated to convert the
software [1] sensing and adapting according to the contextual data into higher forms of knowledge in
evolving situations [8]. In smart environments, order to extract additional information from the
software is handling and distinguishing the physical available context, and lastly, to extend the holding
and virtual aspects of the formed situations [11] knowledge spectral content [21]. The knowledge that
supported by layered software structures [23]. supports the evolving context can be sensed in
However, there is still a lack of prototypes, various compositional forms such as tacit or explicit,
standardization, and the intermediateness of human declarative or procedural, deep or surface knowledge
interaction and expectations [5]. Sensing the smart [21]. It is accustomed to consider knowledge to be
environment involves data management, reasoning, explicit, tacit, and personal [16]. In addition,
and requires adaptability meeting the requirements autopoietic knowledge is produced by the processed
achieving adequate each time behavior [24]. The data [16]. Moreover, knowledge can be viewed as the
complexity of the formed software systems is due to combination of contextual information and
semantics, knowledge reasoning, and situational experience mixed with contextual inference,
conditions along with the developed relationships in interpretation, and reflection [7] while experience is
the evolving activities of the participating roles [2]. considered to be pumped from past or historic data.
Thus, context can be described with semantic graphs Past contextual data can be used as knowledge bases
consisting of concepts administered with Semantic supporting premises for ordering logic predicates
Web Languages (RDFS) [4] and supported by which can be processed further to develop Bayesian
dedicated ontologies [2]. The ontologies support the networks to support reasoning [20] and decision
making. Such networks can be processed by software necessary data management, the required reasoning,
with resource description frameworks (RDF) as and the needed adaptability to present the adequate
proposed by W3C since they consist of first order each time behavior [24]. Such software systems
logic transactions presenting the dynamic aspects of perform complex operations in order to be capable of
knowledge. The knowledge tracing problem is faced satisfying the contextual and user needs employing
using past performance of the system reflecting on approaches from the area of AI and machine learning
the knowledge level and trying to predict the future areas building stacks of methods. The applying
performance based on the current perception [12]. methods vary from simple First Order Logic to Deep
Therefore, knowledge is changing over time Learning software applications in order to obtain the
according to the developing information created by proper set of data to achieve awareness and support
the supervising system and the prediction of the the user's Activities of Daily Living (ADL) and the
learnt knowledge by the system is still an open everyday ordinary socializing consisted of compound
problem [12] with knowledge to be dynamically tasks [19]. Thus, the complexity of the formed
evolving. software systems can be approached as the resultant
The paper continues with a brief orientation of the following three components: semantics,
regarding the characteristics, the challenges, the knowledge reasoning, and situational conditions
necessary supportive methods, and the explored along with the developed relationships in the
architectures of the underlying framework. evolving activities of the participating roles [2].
According to the authors of [2], context can be
described with semantic graphs consisting of
2 Background Framework concepts administered with Semantic Web
The operation of context-aware software designs Languages (RDFS) [4] and supported by dedicated
requires the integration and support from ad-hoc ontologies [2]. Thus, the management of ontologies
networked sensors and artifacts integrated into supports the developed static and dynamic
systems [11]. The availability of sensors provides a knowledge along with applying rules and user
wealth of raw data which must be adequately stored, preferences [10]. Knowledge connects the high-level
processed, and correlated in order to contribute to the user requirements with the low-level hardware
development of the holding context. Also, ad-hoc capabilities to support the activity of inference for
networked hardware artifacts with embedded context-awareness.
software designs participate in deriving the formation In ubiquitous computing environments, inference
of contexts and supporting the user's activity can be obtained applying various methods such as
implementing Marc Weiser's vision regarding probabilistic networks, clustering, rule or case-based
ubiquitous computing environments [41]. The reasoning, AI, and machine learning. Ontology-based
ubiquitous computing environment can be knowledge facilitates the inference process with the
distinguished into the physical and logical areas [11] development of semantic frameworks [15] consisting
which can be referenced interchangeably pointing at of specialized taxonomies implemented with
layered structures [13] in order to support the instantiated classes expressing functional properties
management of the performing systems consisting of and data attributes. The employed classes represent
hardware and software artifacts. Among the proposed the static aspects of the developed knowledge while
layers, also, the processing layer is included, which object orientation provides the dynamic aspects of
can handle Artificial Intelligence (AI) algorithms to knowledge evolution [3] over time. The employment
provide the required intelligence [23]. The resulting of ontology frameworks provides generic structures
complex applications are characterized by the with loosely coupled taxonomies which are capable
capability to sense and adapt, as well as, discover of providing semantic prescriptions and instructions
resources and augment the formed context [8]. to augment the operating software's semantic context
Therefore, the application of context-aware systems along with interpreting and operating capabilities
tends to attain Weiser's vision of providing the appearing in the processes carried in the physical
availability of well-defined software and hardware environment revealing the advanced concept of
systems supporting the user's performed tasks. process-awareness [25]. Hence, there is a stack
The challenges related to context-awareness are which can be traversed and which is based on context
associated with the lack of prototypes, covering all the layers from the raw data to business
standardization, and the intermediateness of human processes hiding the intermediate layers and
interaction and expectations [5]. Also, the gathered providing the means for interaction to the end-user.
data from the contextual environment must be input Context-awareness requires such a computer
to software systems' applications to perform the architecture that demands loosely coupled software
applications to interoperate under a given scenario. frameworks (RDF) as proposed by W3C since they
Such setups can be met in Service Oriented consist of first order logic transactions presenting the
Architecture (SOA) which are supported by open dynamic aspects of knowledge. Computerized
standards such as WSDL, SOAP, and BPEL and systems face the knowledge tracing problem using
implemented independently from the employed past performance of the system reflecting on the
programming languages. Another architectural knowledge level and trying to predict the future
paradigm concerns the Service Delivery Platform performance based on the current perception [12].
(SDP) which uses networks of Internet Protocol (IP) Therefore, knowledge is changing over time
addresses. Both SOA and SDP architectures support according to the developing information created by
the event driven requirements needed for context- the supervising system and the prediction of the
awareness providing services orchestration learnt knowledge by the system is still an open
capabilities. Thus, the software designer separates the problem [12] with knowledge to be dynamically
concerns and is devoted to the management of the evolving.
carried processes [22]. Such architectural “Knowledge Flux” is a recently introduced term
frameworks provide the ability to re-use software coined by the authors to express the concerns of the
modules and develop only the additionally required quantitative interactions of smart and ubiquitous
each time. Moreover, the availability of the internet environments exchanging knowledge. The level of
provides a wealth of available services that can be intensiveness of knowledge must be measurable in
obtained even in cases where the local hardware order to perform comparisons and manage ubiquitous
presents limitations as in the “Software As A computing environments.
Service” (SAAS) case paradigm. Therefore, on top of The rest of the paper starts with references to the
context-aware applications, there can be developed related literature, it carries on with the development
software frameworks where the software designer is of a model accommodating knowledge, it goes on
concerned with the dynamic aspects of processes and with the determination and formal definition of the
the provided knowledge management. data flux for knowledge transfers, it keeps on with a
The state of knowledge reflected by the contextual discussion, and it finishes up with the drawn
environment, including the end-user, is changing conclusions.
dynamically along with the change in context. Thus,
as context is changing the corresponding state of
knowledge is changing to reflect the dynamically 3 Related Literature
occurring changes. However, the changes in context Knowledge can be expressed using the terms of
are not directly proportional to the supporting concept, object, attribute, and relationships along
knowledge. Moreover, knowledge is treated to with the formally defined Concept Algebra to
convert the contextual data into higher forms of perform algebraic manipulations [34]. The concept is
knowledge in order to extract additional information considered to be the unit of knowledge and it is
from the available context, and lastly, to extend the defined [34] as the tuple of
holding knowledge spectral content [21]. The C= ̂ (𝐴, 𝑂, 𝑅 𝑐 , 𝑅 𝑖 , 𝑅 𝑜 ) (1)
knowledge that supports the evolving context can be where, 𝐴 is the set of attributes of a concept "C",
sensed in various compositional forms such as tacit 𝑂 is the set of objects of a concept "C", 𝑅 𝑐 is the set
or explicit, declarative or procedural, deep or surface
of relationships among the objects and attributes, 𝑅 𝑖
knowledge [21]. It is accustomed to consider
is the set of input relationships from other coexisting
knowledge to be explicit, tacit, and personal [16]. In
concepts, and 𝑅 𝑜 is the set of output relationships with
addition, knowledge is developed by the processing
other coexisting concepts.
of contextual data and the autopoietic knowledge
It is apparent from the above definition that the
which is produced by the processed data [16]. Also,
concepts can dynamically form networks or trees of
knowledge can be viewed as the combination of
concepts. Moreover, intention of a concept is the set
contextual information and experience mixed with
of its semantic uses expressed with the concept's
context, interpretation, and reflection [7] with
attributes and properties while extension of a concept
experience considered to be pumped from past or
is the set of its semantic uses beyond its intended
historic data. Past contextual data can be used as
uses, sharing attributes and properties with other
knowledge bases supporting premises for ordering
formal concepts. Networks of concepts are defined
logic predicates which can be processed further to
within the universe of discourse of knowledge which
develop Bayesian networks to support reasoning [20]
is defined [34], as
and decision making. Such networks can be
𝔘= ̂ (𝔒, 𝔄, 𝔑) (2)
processed by software with resource description
 where, 𝔒 refers to the set of objects, 𝔄 coexistence of people and computers. The man-
corresponds to the set of attributes, and 𝔑 denotes the machine paradigm evolves to the development of the
set of relationships holding in the formed Information Society in spite of the accounted
hyperstructure which is depicted in Fig.1. technical, social, and economic problems and leads
to the development of more advanced, if not just
different, computing paradigms [41]. In the
symbiotic context, the discrete spaces of Physical
Space (PS) and Virtual Space (VS) are coupled and
coexist. In between the two spaces there exist the
Community-Agents supporting four function
models: Perceptual Functions, Social Functions,
Cognitive Functions and Decision Functions [31].
Thus, according to the Symbiotic Computing
paradigm, between the user and the coexisting
computer there is an interacting symbiotic partner
agent software facilitating the interaction of PS and
Figure-1. Model of an abstract concept in the DS spaces, Fig.2. The developed model of that
universe of discourse of knowledge. software is analogous to the Layered Reference
Model of the Brain (LRMB) (On Cognitive
Networks of concepts are formed by the input and Computing) with characteristics borrowed from the
output associations developed among the Cognitive Machines [14].
participating concepts. The networks of concepts are
developed within the locally holding universe of
discourse of knowledge which can be distinguished
and characterized as open or closed. Further, the
networks of concepts can be analyzed and processed
with a layered structured model consisting of three
layers, the concept, the object, and the attributes
layers [34].
It is worth mentioning that analytical mathematics
provides the capability of developing matrices of Figure-2. Symbiotic partner agent model.
simultaneous equations for each of the layers
including the input and output parameters in order to The partner agent can facilitate the transfer of the
describe the behavior of the formed system of developed knowledge between the PS and VS spaces.
concepts. However, such an attempt requires However, the representation of the developed
excessive processing power in order to obtain knowledge can be implemented with software
practical results. Alternatively, the use of modules taking advantage of the functional rules
Denotational Mathematics, as it is described below, provided by Concept Algebra. The rules of Concept
provides the required behavior of complex systems. Algebra can be applied on the Object-Attribute-
Processing of networks of concepts can be Relation model specified in the Concept definition in
performed by a set of algebraic operations formally order to develop software artifacts to represent the
defined by Concept Algebra [34]. Denotational holding knowledge between PS and VS spaces. The
Mathematics provide the means to algebraically software artifacts constitute modules to parse the
support the performed manipulations which are formed network of concepts, compile the syntactic
distinguished into relational, reproductive, and adequacy of the concepts interconnected in the
compositional semantic operations. Also, in order to formed network, develop the knowledge base with
face the concepts’ intentions and extensions, granular the identified intentions and extensions (open or
modeling [21] becomes a necessity in order to closed) of the formed network, and visualize the
overcome the excessive appearance of inherited acquired knowledge in some form, for instance, as an
complexity [34]. application of LRMB [29].
Symbiotic Computing is the revealed necessity of Examining the model of the symbiotic partner
the coexistence between computers and the internet agent, it is apparent that there is a continuous
along with society. Moreover, the ubiquitous exchange of messages - information. Thus, the
computing paradigm enhances cooperation and knowledge state changes when information
containing additional concepts are injected into the human brain's operation and defining a magnitude of
symbiotic partner agent. The resulting new knowledge expressed by
knowledge state can be expressed by the composition 𝑑(𝑂𝐴𝑅)
𝐾𝑛𝑜𝑤𝑙𝑒𝑑𝑔𝑒𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑖𝑎𝑙 = (5)
of the existing network of concepts with the addition 𝑑𝑡

of the received concepts. Hence, the dynamic

representation of knowledge is performed by the
Concept Algebra operation of composition which is 4 The concept of flux
defined [37] by composing the participating The dynamic behavior of networks of concepts is
structures, as follows: following the operations performed by the human
brain to transform received input of concepts or
𝑂𝐴𝑅𝑛𝑒𝑤 𝐒𝐓 = 𝑂𝐴𝑅𝑒𝑥𝑖𝑠𝑡𝑖𝑛𝑔 𝐒𝐓 ⊎ relations among concepts into behaviors. The
𝑂𝐴𝑅𝑎𝑑𝑑𝑖𝑡𝑖𝑜𝑛𝑎𝑙 𝐒𝐓 (3) dynamic behavior of concept networks is described
with the dimensionless magnitude of Flux which is
Hence, the dynamic knowledge, 𝑂𝐴𝑅𝑛𝑒𝑤 𝐒𝐓, is defined as the difference between the data structures
achieved by successive compositions of concepts. exported from the physical or virtual space and the
The performed concepts composition is an adaptive data structures imported into that space. Also, input
process allowing the integration of new concepts to and output Flux can define a quotient to determine
existing networks of concepts [37]. Thus, the state of the activity of the applying system. Thus, when
a concept is changing according to the interactions output over input Flux is equal to one then the system
with the coexisting concepts forming networks of can be characterized as stable or knowledge neutral.
concepts representing the holding knowledge state. However, when the Flux quotient is greater than one
The co-existence of a number of associated and then the system is characterized affected by
interacting networks of concepts provide the knowledge while when the Flux quotient is less than
development of systems of concepts. Each part of the one, the system is less affected by knowledge as it is
system inherits the characteristic aspects of the described in the following formula:
constituting concepts as defined in the OAR model (𝑂𝐴𝑅𝐔𝐃𝐌)𝑜𝑢𝑡𝑝𝑢𝑡
𝐹𝑙𝑢𝑥𝑜𝑢𝑡
along with the behaviors expressed by each network 𝐹𝑙𝑢𝑥 = = =
𝐹𝑙𝑢𝑥𝑖𝑛 (𝑂𝐴𝑅𝐔𝐃𝐌)𝑖𝑛𝑝𝑢𝑡
and the constraints applying to each subsystem of < 1 𝑛𝑒𝑔𝑎𝑡𝑎𝑡𝑖𝑣𝑒𝑙𝑦 𝑎𝑓𝑓𝑒𝑐𝑡𝑖𝑛𝑔
concepts [36]. The associated subsystems of a system
{= 1 𝑠𝑡𝑎𝑏𝑙𝑒 𝑜𝑟 𝑛𝑒𝑢𝑡𝑟𝑎𝑙 (6)
of concepts which is within a proper hyperstructure
> 1 𝑝𝑜𝑠𝑖𝑡𝑖𝑣𝑒𝑙𝑦 𝑎𝑓𝑓𝑒𝑐𝑡𝑖𝑛𝑔
has the capability to perform operations from the
following categories of relational, reproductive, and
From the equation (6), it is apparent that the
compositional operations as well as composition with
dimensionless magnitude of Flux represents the
additional subsystems [36] forming large systems
quantitative affection of knowledge on a given space
within the functioning hyperstructure. Hence, the
as it is given by the instantiation of the Unified Data
knowledge represented by such a system of network
Model 𝑂𝐴𝑅𝐔𝐃𝐌. For example, in a home UbiHealth
of concepts is expressed as
𝑛 𝑛 environment [27], the Physical Space (PS) contains,
𝐾 = 𝔑𝐶𝐴 : 𝑋𝑖=1 𝐶𝑖 → 𝑋𝑗=1 𝐶𝑗 (4)
among others, the data structures of five parameters:
where 𝔑𝐶𝐴 represents the applying set of the patient's glucose level as it is measured by some
operations formally defined in Concept Algebra [30]. device, an implanted insulin pump, the measured
Knowledge can be described by a static snapshot heart rate and blood pressure, as well as the time the
of a properly defined and constructed network of data is transferred to symbiotic partner agent. In this
concepts, the way described in the field of example, the supporting system is going to process
Denotational Mathematics. However, knowledge is the received data structures and return a single data
dynamic and evolves [38]. For instance, the learning structure containing the appropriate value with which
process chooses the appropriate set of concepts to the implanted insulin pump is going to be adjusted at
develop knowledge and criticism allows the the patient's health needs.
developing knowledge to evolve further changing
and building additional relationships upon the
associated - existing relationships among the
concepts in the system of networks of concepts [38].
Knowledge presents dynamic aspects and evolves by
incrementally altering the relationships among the
nodes of the networks of concepts mimicking the
 Processing of the concept networks is performed the big-R notation [35] in the RTPA framework of
by the Real-time Process Algebra (Tian, et al, 2011) Denotational Mathematics.
(RTPA) algebraic environment which is formally
defined with Denotational Mathematics.
Denotational Mathematics in RTPA provides the
separation of concerns regarding the static and
functional behaviors of the underlying concepts in
the physical or virtual spaces with Unified Data
Model (UDM) and Unified Process Model (UPM),
respectively. Thus, RTPA provides the means for
knowledge representation and Real-Time Operating
System plus (RTOS+) [31] provides with
Denotational Mathematics the means to perform
system, resources, and process management
Figure-3. Autonomous Symbiotic partner agent
including multiple levels of interrupts handling.
model.
Therefore, the formal algebraic representation of
advanced concepts and the necessary underlying
The hyperstructure 𝔄 can be manipulated and
operating system form a complete computing
analyzed into three layers: the knowledge, the object,
environment and the efficiency is expressed with
and the attribute layers [33]. Distinguishing the
Flux.
hyperstructure's contents into layers provides the
The symbiotic partner agent is in between the PS
capability of separation of concerns which means that
and VS spaces with the capability to function
an employed software agent can manipulate the
autonomously. In Figure-3, the symbiotic partner
attributes relations with the side effect to influence
agent model is modified in order to acquire those
the contents of the other layers. However, the
modules that provide the necessary autonomy
attributes mathematical manipulation requires the
functionality performing decision-making and the
capability of moving on various levels of granularity
goal-driven processes [38]. The autonomic behavior
taking advantage of the property of inheritance
is already defined as the behavior depended upon the
among concepts to represent in real-time abstract
goal-driven, the decision-driven, and the interrupt-
concepts or instantiations of concepts [21].
driven behaviors [40]. The interrupts are initiated by
The knowledge of the hyperstructure 𝔄 can be
internal or external events and timing requests
processed with the support of the autonomic agent.
affecting the overall behavior of the system
The support of the autonomic agent refers, among
𝐵𝑎𝑢𝑡𝑜𝑛𝑜𝑚𝑖𝑐 in which the symbiotic partner
others, to the consideration of the concepts of rough
participates and it is expressed by
sets defining the membership of relative or implied
𝐵𝑎𝑢𝑡𝑜𝑛𝑜𝑚𝑖𝑐 = ̂ {𝐵𝑔𝑜𝑎𝑙 , 𝐵𝑑𝑒𝑐𝑖𝑠𝑖𝑜𝑛 , 𝐵𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡 } (7) concepts [40]. Thus, the employment of the terms
In Figure-3, the symbiotic partner agent must be defined by the rough sets membership functions can
equipped with the autonomous agent (AA) which can support the intended or extended use of concepts
be incorporated with the rest of the partner's modules performing rough sets manipulations by including-
[17]. The formed system includes the representation excluding terms with sets operations. Moreover, the
of the Generic Abstract Intelligence Model (GAIM) use of rough sets provides the capability to define the
as formally defined in [38]. dependency among sets in order to provide support to
The knowledge developed in both PS and VS the decision-making or the goal-driven aspects of the
spaces can be expressed as the Cartesian product autonomic agent [40]. The decision-making process
among all the participating concepts [33] within the relies on the development of rules with a certainty
holding universe of discourse of knowledge 𝔘. The provided by the membership of the examined
operation of participants in PS and VS spaces forms concepts [40]. Hence, the hyperstructure 𝔄 provides
the local knowledge base which is a hyperstructure 𝔄 the base to develop rules on each of the three layers
denoting hierarchical structures of interconnected using hierarchical structures of rules of varying
concepts with weighted relationships denoting the certainty supporting the intension and extension
strength of their semantic relations [33], it is given by operation on concepts in order to achieve
the following equation: approximations, dependencies, and reductions that
𝑛 𝑛 𝑛 𝑛
𝔎= ̂ 𝑅𝑖=1 𝐶𝑖 x𝑅𝑖=1 𝐶𝑖 = 𝑅𝑖=1 𝑅𝑗=1 (𝐶𝑖 , 𝐶𝑗 ) (8) lead to varying data granularity [40]. The
The mathematical term 𝑅 is an operator that relationships among concepts belonging to
represents repetitive or recursive operations, called hierarchical structures is examined using
probabilities, like in [28], with the known problems relational databases [9], and language representations
related to the unavailability of Bayesian posterior of knowledge [18].
probabilities conditions. Therefore, calculating the Knowledge graphs present similar data structures
concept’s participation with membership functions of to networks of concepts and they share analogously
rough sets provides deterministic behavior to the practical problems such as the size of the constituting
design of a system. networks with large number of interrelated nodes, the
The limitations of the symbiotic partner agent are consistency of the denotations of the participating
related with two open issues which refer to the size nodes, and the operating environments. In addition,
and the homogeneity of knowledge. The size of the the parameters of each node increase the difficulty of
managed knowledge cannot be strictly defined for a the management administration of concepts
general-purpose system. Also, the homogeneity of networks. However, there are techniques to
the concepts participating in the knowledge efficiently traverse the concepts network for
representation cannot be rigorously defined. searching, placing queries or interacting with
However, the internet provides the capability to questions-answers. In the administration of
perform references to well-defined knowledge bases knowledge graphs besides the developed
and exemplars reducing the demands for the size of interrelations among nodes, it is attempted to enrich
the locally managed memory. In addition, the each node of the formed network with the semantic
homogeneity of the concepts participating in the D codes. Moreover, each code can take up K values
hyperstructure 𝔄 can be defined with the proper use forming K-way D-dimensional codes [6] in order to
of the functionalities of intentions and extensions. present the similarity among the comparing nodes.
Therefore, the current limitations of size and Thus, simply traversing the concepts network or
concepts of homogeneity can be partially remedied applying a query on them, an encoder/decoder model
with the use of the internet and the concepts mechanism applies using a discretization / reverse
functionality of intentionality and extensionality. discretization function [26] for representing complex
The hyperstructure 𝔄 supports the development of or advanced forms of knowledge. Embedding
dynamical hierarchical networks of concepts which characteristics is performed in [6] by a function F:V
can be augmented or reduced using the →𝑅 𝑑 , where V represents the semantics of concepts
functionalities of concept intentionality and placed in a table - vocabulary which corresponds to a
extensionality. Knowledge is dynamic and capable of vector in 𝑅 𝑑 [6]. However, as the size of the concept
offering representations of static snapshots of the network is increasing there must be found ways to
instantly holding knowledge, as well as, the preserve efficiency and among them compression [6]
knowledge generation [17]. The knowledge is considered as promising.
generation can be represented by the concept In Denotational Mathematics, with a formally
functionalities of intentionality and extensionality defined environment, the recurrent operations are
which can develop additional concepts, relations defined with the big-R notation in the Real-Time
augmenting or shrinking the represented knowledge. Process Algebra (RTPA). Knowledge consisted of
Therefore, the relational composition of concepts concepts networks is represented [35] by the
leads to the development of complex concept following recurrent expression:
constructs or newly defined concepts. 𝔎= 𝑛
̂ 𝑅𝑘=1 𝔎𝑘 (𝔎𝑘−1 ) = 𝔎𝑛 (𝔎𝑛−1 (… . 𝔎1 (𝔎0 ))) (9)
The development of concepts networks in the and base condition as,
hyperstructure 𝔄 represent the developing 𝑚
𝔎0 = 𝑋𝑙=1 𝐶𝑙 (10)
knowledge and provide the capability to discover or where 𝑋 is the cartesian product among 𝐶𝑙 's.
extract the contained knowledge. The knowledge Denotational Mathematics provides the means to
extraction can be performed by applying queries on develop knowledge bases [35] with which knowledge
the knowledge graphs or semantic networks can be acquired, fused in the base, manipulated and
consisting of interrelated concepts. In [9], the authors retrieved. Hence, there is a sound and complete logic
claim that literature information can be turned into environment to administer knowledge bases [35].
knowledge when the received data is placed in a However, such knowledge bases develop
network and new knowledge can be generated by relationships and references among the concepts over
augmenting the network nodes-data-concepts time undergoing a change of its contents due to its
developing extended knowledge graphs. Therefore, interaction with the environment. Thus, the behavior
knowledge extraction from the concept networks is a of the formed system of the concepts network is
technical challenge that can be addressed with changing according to the built relationships among
Resource Description Framework (RDF), non- the concepts expressing the self-adjustment or the
earned experiences from the interaction of the receives data and produces data that maintain the
surrounding environment resulting in a cognitive level of the developed knowledge of the formed
dynamic system [14]. The applying rules of the system. Figure-1 depicts the interaction between the
concept’s networks spawn the observed intelligence, environmental-physical and the computer contexts
learning, adaptivity, activity or action in real-time presenting the flows of data between the interacting
affected by the surrounding environment [14]. The contexts which is proportional to electromagnetic
operation of a system consisting of a knowledge base flux.
requires the support of a complete computing The flow of data or data flux is the measurable
environment. Such a computing environment or magnitude to represent the total knowledge which is
system must be initialized, the components of the developed in the interaction between a physical and
system are properly synchronized and sense internal a computer context. The data flux can be used to
and external events interrupting the executing flow of describe the induced knowledge to the active system
operations and re-schedules the operations. The such as that of Figure-1. The exchanged data between
internally carried operations must be managed using the physical and computer context contribute to
temporary and permanent memory banks. Thus, such knowledge components of the formed system. The
a system collects operations and builds processes data flux is the magnitude that is formed by the
with which the system requests, creates, performs, matching data structures of the physical and
runs, interrupts, completes, delays, suspends, kills, computer contexts. In other words, given a set B of
and dispatches the available sets of operations or data structures representing the physical context
processes [32]. Therefore, the operation of a projected to the set A of data structures representing
knowledge base requires the support from a Real- the computer context is the data flux which is
Time Operating System to manage the carried expressed as 𝛷 = 𝐵 ∗ 𝐴. The flux Φ represents the
internal processes and synchronize the requests from number of procedures carried out in a system in order
the surrounding environment with a Real-Time to develop the holding knowledge of that system.
Operating System (RTOS+) [32]. Another term borrowed from electromagnetics is the
flux density which in our case is the division of
contextual data structures over the holding
5 Modelling knowledge's data structures. The contextual data
The description of the data and information structures are driven by the carried procedures which
exchanged among the ubiquitous computing govern the operation of the occurring system.
participants and the system requires an adequately When there is observed data flux across contexts then
matching model. A brief description of the model is there are spawned processes. The spawned processes
presented in the following as well as the carried act analogously to the generation of voltage when a
procedures during the performed interactions, and the conductor is moving within a magnetic field. The
mathematical expression of the dynamic developing processes can generate additional processes which
behavior. can be internal or external to the system depending
A system is formed containing the physical upon the nature of a closed or open system. Flux is
context and the context developed into the supporting changing along time at discrete time intervals
computer machine. Within this system there is expressed as Φ(t) = Φ(t1)-Φ(t0) = (ΔΦ1(t1)+
always a continuous interaction between the two ΔΦ2(t1)+ ΔΦ3(t1)+.....) - (ΔΦ1(t0)+ ΔΦ2(t0)+
contexts. From the physical context, data is flowing ΔΦ3(t0)+.....) representing the flux of the
to the computer machine while from the computer constituting knowledge components. Knowledge can
context data is actuating and affecting the operational be represented as the rate of change of flux given by
terms of the physical context. The total flow of data 𝐾 = 𝛥𝛷/𝛥𝑡.
between the physical and the computer contexts
determines the holding knowledge of the formed
system. 6 Determination and Formal
The physical context is formed by the activity Definition of Data Flow
carried by the artifacts and the users in the computer The sensed knowledge flow leads to the
environment. The activity of the artifacts produces determination of the dimensionless magnitude of data
data that is driven to the co-functioning computer. flux. The static and dynamic aspects of flux are
The computer is processing the received data and associated with the corresponding aspects of
produces data for the connected transducers to knowledge.
properly affect the artifacts and the users of the
physical - environmental context. The computer
6.1 Data Flux symbiotic partner agent. At a given instance of time,
The interaction between the PS and VS spaces is the flux 𝐹𝑜𝑢𝑡 represents how the developing
controlled by the intermediate symbiotic partner knowledge in the symbiotic agent affects the carried
agent. The interaction between the PS and VS spaces processes in the PS physical domain. Similarly, the
is sensed by the transmission/reception of data which flux 𝐹𝑖𝑛 represents how the developing processes in
can be performed by the direct connection of each PS the physical domain affects the build of knowledge in
or VS space with the symbiotic partner agent or the symbiotic partner agent. Therefore, the static
through the use of some intermediate memory bank. representation of flux is performed by the description
The interaction between the PS and VS spaces is of the data structures existing in the artifacts of the
verified by the existence of data flow. The physical domain and the memory of the symbiotic
enumeration of the set of data structures from the PS partner agent.
space is denoted by 𝐹𝑖𝑛 while the corresponding set
received is denoted as 𝐹𝑜𝑢𝑡 . When data is exchanged 6.3 Dynamic representation of Flux
between PS and VS spaces then there are two options: The symbiotic partner agent is the software system
(a) data status is changed, and (b) the concept that continuously and constantly supports the
network managed by the symbiotic partner agent is operation of a ubiquitous computing environment
also changed. Thus, we define the magnitude of data such as UbiHealth. This continuous cooperation and
flux 𝐹 given by interaction between the symbiotic partner and the
𝐹 UbiHealth environment is driving the development of
𝐹 ≙ 𝐹𝑜𝑢𝑡 (11)
𝑖𝑛 flux. The presence of flux leads to the development
with 𝐹𝑖𝑛 being the data transferred from the PS to of knowledge which is the result of the changing in
VS while 𝐹𝑜𝑢𝑡 is the returned data from VS to PS. the transfer of structured data. The flux is the rate of
Thus, at any given instant of time 𝑡, the function 𝑓(𝑡) change of the data exchanged between PS and VS.
represents the flux 𝐹 which takes values given by Therefore, the existence of flux leads to the
< 1, 𝑐𝑎𝑟𝑑𝑖𝑛𝑎𝑙𝑖𝑡𝑖𝑒𝑠: |𝐹𝑖𝑛 | > |𝐹𝑜𝑢𝑡 | development of knowledge and hence, the dynamic
𝑓(𝑡) = {= 0, 𝑐𝑎𝑟𝑑𝑖𝑛𝑎𝑙𝑖𝑡𝑖𝑒𝑠: |𝐹𝑖𝑛 | = |𝐹𝑜𝑢𝑡 | change of flux is the built knowledge by the
> 1, 𝑐𝑎𝑟𝑑𝑖𝑛𝑎𝑙𝑖𝑡𝑖𝑒𝑠: |𝐹𝑖𝑛 | < |𝐹𝑜𝑢𝑡 | symbiotic partner agent.
(12) The knowledge represented by the holding status
The comparisons between the sets of 𝐹𝑖𝑛 and of the concept networks changes by the received flux.
𝐹𝑜𝑢𝑡 can be performed at the syntactic or the semantic Thus, the state of the concept network changes upon
levels. In any case, the comparison focuses on the the reception of additional data from the PS space and
magnitude the governing rules of the concept network bring the
𝐹𝑡 = (𝐹𝑖𝑛 ∩ 𝐹𝑜𝑢𝑡 ) (13) network to another state depending upon the previous
The instant value of flux 𝐹𝑡 can be given by the state. Hence, the concepts network state depends
symbiotic partner agent which can be analyzed with entirely on the previous one revealing a recursive
the application of Denotational Mathematics. The use relationship among the states. Moreover, the concept
of Denotational Mathematics facilitates the formal network requires an initialization leading to the
description of the computing environment that conclusion that there must be a supervising real-time
supports the operation of the symbiotic partner agent, operating system that monitors its operation. In
as well as, the functionality of the symbiotic partner addition, the constituting components of the concepts
agent itself. The computing environment can be network undergo selectable changes depending upon
described and implemented with an appropriate the reception of flux which can cause only parts of
modification of the Real-Time Operating System the concepts network to change while others remain
(RTOS+) [32]. In a similar fashion, the formal unchanged. This fact leads to the conclusion that the
description of the symbiotic partner agent can be concept network operations can be described by
performed with the use of the available tools recursive partial functions. For instance, the artifacts
provided by the Real-Time Process Algebra (RTPA) of a UbiHealth environment can provide sets of data
given in [38]. structures that leave the built knowledge unaffected.
Therefore, the employment of partial functions
6.2 Static representation of Flux provides the advantage of using basic functions
The representation of the PS and VS spaces can be which can be closed under complex operations of
achieved with the formal description of the existing composition and primitive recursion leading to the
data structures. Thus, the static representation of the following expression:
flux is described by the intersection of the sets
constituting the fluxes entering and leaving from the
 𝔎0 = 𝑐, 𝑐 ∈ 𝑆𝑖𝑛𝑖𝑡 = {𝑐1 , 𝑐2 , … 𝑐𝑛 } symbiotic partner agent receives the data structures
𝑘 obtained by the IoT installation in order to develop
𝔎 = {𝔎𝑘 𝔎𝑘−1 , 𝑘 ∈ 𝑆𝑉𝑆
0, 𝑘 ∉ 𝑆𝑉𝑆 the built flux. The developed flux feeds the software
(14) modules of the symbiotic partner agent in order to
The 𝔎𝑘 function is a total and computable partial develop adequate representations of concepts
function as it is formed by the collection of partial networks. The functionality of the concept networks
functions and closed under some fundamental adjusts the knowledge status to correspond to the
operations for forming new functions from old ones. needs of the PS. The changes in the built knowledge
Therefore, the mathematical representation of the feed the decision-making modules of the symbiotic
flux is given as a function of the built knowledge: in partner agent to react accordingly to the carried
continuous form processes in the PS space.
𝑑𝐹 𝑑 The operation of the symbiotic partner agent
𝐹 = 𝑑𝑡 = 𝑑𝑡 (𝔎) (15a) software interacts with the works, procedures, and
or in discrete form processes taking place in the PS space by receiving
𝐹 = 𝛥𝐹 = 𝛥(𝔎𝑘 − 𝔎𝑘−1 ). (15b) the produced data. The developing flux passes the
flows of data to the symbiotic partner agent to
develop the corresponding concepts network. The
7 Discussion developing knowledge by the concepts networks
The sensed data of the carried processes in the PS feeds the decision-making module to cause the
space support the built flux that causes the symbiotic partner agent to export the data to those
development of the corresponding knowledge in the data structures required in order to transfer the
VS. Any change in the processes' workflow results in knowledge impacts to the carried processes of the PS
an analogous change in flux and consequently, the space. Therefore, knowledge governs the
corresponding effects on the developed knowledge. functionalities of the processes operating in the
Thus, flux can be used as that magnitude which physical domain.
senses the functionality of the PS and affects the
knowledge of the VS space. Therefore, there always
exists a relationship among the processes carried in 8 Conclusion
PS, the developed flux, and the knowledge The formal mathematical representation of
developing in the VS space. knowledge with the application of Denotational
The carried processes in the PS space and the Mathematics provide the opportunity to represent
development of knowledge in the VS space, they are constituting elements and quantified magnitudes. The
related with direct proportional but not linear formal Denotational Mathematics framework
relationship. The rate of change of the carried provide adequate mathematical tools to perform
processes provides the rate of change in the manipulations on the rigorously defined elements,
developing flux between the PS and VS spaces. In a magnitudes and their relationships which are required
similar fashion, the given instant of flux causes a to express the built knowledge.
definite rate of change in the VS space expressed by Between PS and VS, the developed knowledge is
the development of knowledge. In other words, flux caused by the presence of flux which is defined by
expresses the dynamic changes in the processes of the the update of the data structures transferred and
PS space and the dynamic changes in the built exchanged between the two spaces. Flux is a
knowledge, as it is presented in (16) below: magnitude proportional to knowledge and it is used
𝑑(𝑃 ) 𝑑(𝐾 ) to perform knowledge-able control on the physical
𝐹 = 𝑑𝑡𝑛 = 𝑑𝑡𝑛 (16)
domain. The formal mathematical definition of flux
where 𝑃𝑛 and 𝐾𝑛 represent the sets of the values provides the capability to use mathematical tools to
domains of the computable functions describing the determine the instant knowledge, the processes
processes in the PS and the concepts network in the status, and the capabilities of the symbiotic partner
VS spaces respectively. agent.
The engineering design refers to the design of the The operation of the symbiotic partner agent
symbiotic partner agent software which is in between facilitates the interactions between the PS and VS
the PS and the VS spaces. The symbiotic partner spaces. The interaction of the symbiotic partner agent
agent is equipped with adequate modules which are with the computing environment, e.g. UbiHealth
connected with properly defined internet of things environment, provides the capability of applying
(IoT) development in order to sense the functionality direct knowledge-based control of the carried
of the carried processes in the PS space. The processes.
 The future research plans focus on the wearable computing. Digest of Papers. Third
development of a symbiotic partner agent prototype. International Symposium on Wearable
The prototype is expected to provide the basis for Computers, San Francisco, CA, USA, 21-28.
experimentation on the uses of flux in the [9] Dörpinghaus, J., Stefan, A., Schultz, B., &
development of knowledge and knowledge-based Jacobs, M. (2020). Towards context in large
control on the surrounding computing environments.
scale biomedical knowledge graphs. [Online].
The prototype is going to present the attributes
mentioned above with the static and dynamic Available: http://arxiv.org/abs/2001.08392
facilities provided by Denotational mathematics. The [10] Gandon, F.L., & Sadeh, N.M. (2004). Semantic
symbiotic partner agent will be designed as an web technologies to reconcile privacy and
internet service implementing the quantization of the context awareness, Journal of Web Semantics,
continuous magnitude of Knowledge Flux along with 1(3), 241-260.
the relationship with time, i.e. the consideration of [11] Gellersen, H.W., Schmidt, A. & Beigl, M.
knowledge as a function time. (2002). Multi-Sensor Context-Awareness in
Mobile Devices and Smart Artifacts. Mobile
References: Networks and Applications 7, 341–351.
[1] Abowd G.D., Dey A.K., Brown P.J., Davies N.,
[12] Ghosh, A., Heffernan, N., & Lan, A. S. (2020).
Smith M., & Steggles P. (1999) Towards a
Context-aware attentive knowledge tracing. In
Better Understanding of Context and Context-
Proceedings of the 26th ACM SIGKDD
Awareness. In: Gellersen HW. (eds) Handheld
International Conference on Knowledge
and Ubiquitous Computing. HUC 1999. Lecture
Discovery & Data Mining, 2330-2339.
Notes in Computer Science, vol. 1707. Springer,
[13] Hofer, T., Schwinger, W., Pichler, M.,
Berlin, Heidelberg.
Leonhartsberger, N., Altmann, J. &
[2] Anagnostopoulos, C., A. Tsounis, S.
Retschitzegger, W. (2003). Context-awareness
Hadjiefthymiades. (2007). Context awareness in
on mobile devices - the hydrogen approach. 36th
mobile computing environments. Wireless
Annual Hawaii International Conference on
Personal Communications, 42 (3), 445-464.
System Sciences, Big Island, HI, USA, 10-18.
[3] Armand, A., Filliat, D. & Ibañez-Guzman, J.
[14] Kinsner, W. (2007). Towards cognitive
(2014). Ontology-based context awareness for
machines: Multiscale measures and analysis.
driving assistance systems. IEEE Intelligent
International Journal of Cognitive Informatics
Vehicles Symposium Proceedings, Dearborn,
and Natural Intelligence, 1, 28–38.
MI, USA, pp. 227-233.
[15] Korpipää P., & Mäntyjärvi J. (2003) An
[4] Brickley, D., & Guha, R. (2001). RDF
Ontology for Mobile Device Sensor-Based
Vocabulary Description Language 1.0: RDF
Context Awareness. In: Blackburn P., Ghidini
Schema. W3C Working Draft.
C., Turner R.M., Giunchiglia F. (eds) Modeling
http://www.w3.org/TR/PR-rdf-schema
and Using Context. Lecture Notes in Computer
[5] Bricon-Souf, N., & Newman, C.R. (2007).
Science, 2680. Springer, Berlin, Heidelberg
Context awareness in health care: A review,
[16] Koskinen, K. U., Pihlanto, P., & Vanharanta, H.
International Journal of Medical Informatics,
(2003). Tacit knowledge acquisition and sharing
76(1), 2-12.
in a project work context. International journal
[6] Chen, T., Min, M.R., & Sun, Y. (2018).
of project management, 21(4), 281-290.
Learning k-way d-dimensional discrete codes
[17] Lieto, A., Lebiere, C. & Oltramari, A. (2018).
for compact embedding representations. arXiv
The knowledge level in cognitive architectures:
preprint arXiv:1806.09464.
Current limitations and possible developments,
[7] Dalley, J., & Hamilton, B. (2000). Knowledge,
Cognitive Systems Research, 48, 39-55.
context and learning in the small business.
[18] Liu, W., Zhou, P., Zhao, Z., Wang, Z., Ju, Q.,
International Small Business Journal, 18(3), 51-
Deng, H., & Wang, P. (2020). K-BERT:
59.
Enabling Language Representation with
[8] Dey, A.K., Salber, D., Abowd, G.D. &
Knowledge Graph. Proceedings of the AAAI
Futakawa, M. (1999). The Conference
Conference on Artificial Intelligence, 34(03),
Assistant: combining context-awareness with
 2901-2908. UbiHealth. International Journal of Software
https://doi.org/10.1609/aaai.v34i03.5681. Science and Computational Intelligence
[19] Lukowicz, P., Pentland, S. & Ferscha, A. (IJSSCI), 14(1), 1-14.
(2012). From Context Awareness to Socially [29] Yousheng, T., Wang, Y., & Hu, K. (2009). A
Aware Computing. IEEE Pervasive Computing, knowledge representation tool for autonomous
11(1), 32-41. doi: 10.1109/MPRV.2011.82 machine learning based on concept algebra. In
[20] Pedrycz W. (2021). From Data to Information Transactions on Computational Science V, 143-
Granules: An Environment of Granular 160. Springer, Berlin, Heidelberg.
Computing. IEEE 20th Int’l Conf. on Cognitive [30] Wang, Y. (2012). On denotational mathematics
Informatics and Cognitive Computing foundations for the next generation of
(ICCI*CC’21), Banff, AB., Canada, IEEE CS computers: cognitive computers for knowledge
Press, Oct., p.2. processing. J Adv Math Appl 1(1), 118–129.
[21] Pomerol, J.C., & Brézillon P. (2001). About [31] Wang, Y., Ngolah, C., Zeng, G., Sheu., P.C.,
Some Relationships between Knowledge and Chiy, P., & Tian, Y. (2010). The Formal Design
Context. In: Akman V., Bouquet P., Thomason Model of a Real-Time Operating System
R., Young R. (eds) Modeling and Using (RTOS+): Conceptual and Architectural
Context. Lecture Notes in Computer Science, Frameworks. International Journal of Software
2116. Springer, Berlin, Heidelberg. Science and Computational Intelligence 2(2),
[22] Preuveneers, D., & Berbers., Y. (2008). Internet 105-122.
of Things: A Context Awareness Perspective. [32] Wang, Y., Baciu, G., Yao, Y., Kinsner, W.,
The Internet of Things: From RFID to the Next- Chan, K., Zhang, B., & Hameroff., S. (2010).
Generation Pervasive Networked Systems, Perspectives on cognitive informatics and
Auerbach. cognitive computing. International Journal of
[23] Randell, C., & Muller, H. (2000) Context Cognitive Informatics and Natural Intelligence,
awareness by analysing accelerometer 4(1), 1-29.
data. Digest of Papers. Fourth International [33] Wang, Y., Zatarain, O., & Valipou, M. (2017).
Symposium on Wearable Computers, Atlanta, Building cognitive knowledge bases sharable by
GA, USA, 175-176. humans and cognitive robots. In 2017 IEEE
[24] Ranganathan, A., & Campbell, R.H. (2003). An International Conference on Systems, Man, and
infrastructure for context-awareness based on Cybernetics (SMC), 3189-3194.
first order logic. Personal and Ubiquitous [34] Wang, Y. (2015). Concept algebra: A
Computing, 7(6), 353-364. denotational mathematics for formal knowledge
[25] Rosemann, M., Recker, J., Flender, C., & representation and cognitive robot learning.
Ansell, P. (2006): Understanding Context- Journal of Advanced Mathematics and
Awareness in Business Process Design. Proc. of Applications. 4(1), 61-86.
the 17th Australasian Conference on [35] Wang, Y. (2014). On a novel cognitive
Information Systems. Australia Association for knowledge base (CKB) for cognitive robots and
Information, Adelaide, Australia. machine learning. International Journal of
[26] Sachan, M. (2020). Knowledge Graph Software Science and Computational
Embedding Compression. In Proceedings of the Intelligence. 6(2), 41-62.
58th Annual Meeting of the Association for [36] Wang, Y. (2008). On concept algebra: A
Computational Linguistics, 2681-2691. denotational mathematical structure for
[27] Sarivougioukas, J., Vagelatos, A., Parsopoulos, knowledge and software modeling."
K., & Lagaris, I. (2017). Home International Journal of Cognitive Informatics
UbiHealth, Encyclopedia of Information and Natural Intelligence. 2(2), 1-19.
Science and Technology, Fourth Edition, IGI [37] Wang, Y. (2007). The OAR model of neural
Global. informatics for internal knowledge
[28] Sarivougioukas, J., & Vagelatos, A. (2022). representation in the brain. International Journal
Fused Contextual Data with Threading of Cognitive Informatics and Natural
Technology to Accelerate Processing in Home Intelligence. 1(3), 66-77.
[38] Wang, Y. (2009). Toward a formal knowledge
system theory and its cognitive informatics
foundations. In Transactions on Computational
Science V, 1-19. Springer, Berlin, Heidelberg.
[39] Wang Y., Karray F., Kaynak O., Kwong S.,
Leung H., Plataniotis K.N., Hou M., Rudas I.J.,
Tunstel E., Trajkovic L., and Kacprzyk J.
(2021), Perspectives on the Philosophical,
Cognitive and Mathematical Foundations of
Symbiotic Autonomous Systems (SAS),
Philosophical Transactions of Royal Society
(A), 379(x):1-16, Oxford, UK.
[40] Wang Y., Hou M., Plataniotis K.N., Kwong S.,
Leung H., Tunstel E., Rudas I.J., and Trajkovic
L. (2021), Towards a Theoretical Framework of
Autonomous Systems Underpinned by
Intelligence and Systems Sciences, IEEE/CAS
Journal of Automatica Sinica, 8(1), 52-63.
[41] Weiser, M. (1999). The computer for the 21st
century. ACM Mobile Computing and
Communications Review, 3(3), 3-11.