Introduction

The expert group ability of making simultaneously non-causal (non-formal) and good decisions is an issue that is constantly discussed in the framework of group decision-making activities. This is because a formal description of the problem is not always helpful when determining suitable solutions for different problems.

At the same time, formal descriptions of the problems are very important for speeding up the group decision-making process. This is particularly true in fields such as crisis and emergency management, politics, and human relations. In this context, methods for speeding up the group decision-making processes are very important [10]. One of these methods creates special structural conditions that accelerate the group insight achievements.

In this paper, we define insight as a characteristic of the human mind whereby a decision is made by anticipating the future, by comprehending the whole, or by creating a group holistic discourse [11]. However, it is not the result of a logical analysis. Insight is used to describe the phenomenon of a person experiencing a momentary illumination or non-causal enlightenment after a series of fruitless attempts to find a solution.

Intuitive solution investigations have a wide scope. Since Vedic times (ca. 1750–500 BCE), people have investigated and used the mystery phenomena of meditation and enlightenment. For a long time, researchers have been intrigued by the effectiveness of instantaneous decisions [11], and there has been substantial interest in the “Eureka-effect” [9]. The futility of finding solutions to human problems using only formal methods was discussed in [2].

Governments, corporate teams, and groups of employees or experts are interested in accelerating the group decision-making process. E-government and virtual collaboration [1, 13] methods have presented a novel challenge. In networks, participants in decision-making processes cannot instantly reach a mutual understanding and gain a group insight, because there are too many messages that must be passed between participants. They must typically discuss structural elements of the decision-making process such as goals, tasks, models, and other factors.

In a networked team conversation, the participants have numerous unmeasured characteristics such as trust, intention, emotion, thought, expectation, and responsibility. These can reduce the probability of reaching a high level of mutual understanding, which means that a group insight is becoming more difficult to achieve. In particular, it is important to support decision-making processes that use e-expertise [4].

In this paper, we propose and prove that we can significantly speed up group decision-making processes by applying a special convergent structure for the semantic interpretation of concepts in cognitive modeling processes. We propose an implementation of this special information structure, which is generated during the networked brainstorming and group decision-making process. This information structure is based on the inverse problem method and uses a breakthrough thinking (“eureka effect”) approach. We use this to introduce a special type of semantic interpretation that is based on the cognitive modeling method, i.e., cognitive programming.

Insight semantic

Insight’s forces

Experts can instantly estimate events or objects and predict event behavior, which may be beyond the ability of a group of scientists who are conducting a prolonged study. A group of scientists or researchers may find the solution of a complex problem after some time. At some point, there is an uncaused hint or cognitive snap (insight) that creates conditions that lead to instant solutions (for example, consider Archimedes or Gutenberg). We must determine the nature of this insight, the formal structure of the path that leads to an uncaused decision or insight, and the mapping between the unformalized insight and the formalized structure that represents the semantics of the insight.

An insight occurs under certain conditions [9]:

  • A persistent intention to find a good solution;

  • A long and unsuccessful search;

  • Little apparent progress;

  • A precipitating event;

  • A cognitive snap; and

  • A transformation of the existing plan.

A creative search of the solutions and ideas in group or individual discussions or brainstorming can be characterized by the following, as noted in [2].

  • The immensity of the possible research directions, when the search is characterized by a variety of intriguing destinations with a small number of solutions.

  • The search wanders through the space without any tips and hints on the correct course of action or decision.

  • There are no solutions (insularity) within a certain part of the problem.

  • There are constantly encountered deceptive promises of the right decision (deceptiveness).

The following semantic pattern is well known. If the syntax of the problem description is more complex (for example, language with roles), there are less class interpretations, and, consequently, there is a smaller probability of achieving a good, holistic representation of the problem. We can use this to enhance the holistic nature of the semantic interpretation; that is, we should try to choose a more generalized interpretation so that we have a more powerful semantic interpretation. For example, if the metric space of the semantic interpretation is not appropriate, it should be increased to a more conceptual level for the semantic interpretation. Instead of the metric distances between points, we use the intersections of the points’ neighborhoods. So, if we wish to increase a team’s trust in a given differential equation model of a political situation, we should use a more conceptual representation of the model. For example, we may use cognitive modeling and programming (see sections “Congnitive Convergence” and “Congnitive Programming”).

Insight in decision-making processes occurs in the context of three different forces: intentions, external pressure, and experience. Experience is the only one of these three forces that can be well represented using formalized schemes or models.

The semantic of the group insight in this context can be emphasized by mapping the intentions and external pressure in formalized models of the experience (which could be a database or a knowledge base). Projection semantics (formal semantics interpretation) can be expressed by mapping unformalized concepts (phenomenon or events) into mathematical constructions (for example, logic, programs, or sets). The concepts must be mapped into the mathematical constructions or sets of objects (i.e., documents, messages, or comments) for each modeling method. This is particularly true in cognitive modeling, where the model consists of unformalized concepts and their connections.

Restrictions of the formalisms

For networked group decision-making and insight processes, we must solve a classically formulated problem that cannot be solved by classical methods. This may be because of the enormous amount of knowledge, or the paradoxical nature of the problem. Consider a situation that occurred when analyzing natural phenomena. When conventional drawings and imaging methods did not result in an explanation, the quantum approach to semantics appeared [10]. Drawings were significantly widened and deepened, and considered from an infinite number of different angles. This resulted in new fundamental laws for the semantic interpretations of drawings. In the quantum approach, an object’s properties depend on the observer. This is also the case in texts. The signs themselves do not mean anything, and are interpreted differently by different people. The usefulness of quantum semantics was shown in [11], and they are particularly useful for ensuring a holistic representation of a problem.

Text can describe a lot of things and phenomena, even when an expert cannot describe everything. The expert can categorize and classify the situation, and sometimes uses negation (i.e., “not a flower”, “not a bicycle”, or “not a mind”). They can consider feelings and values, by looking at relationships instead of words and morphemes. This is the basis behind the well-known language concept, according to which language is seen as a particle and wave [9]. In this sense, the phoneme and phonetics are simultaneously different and the same, similar to a particle and a wave in quantum mechanics. So, the phoneme is the basic unit of the sound system used to identify and discriminate meaningful units (morphemes), and the phonetics are used to study the holistic sound of languages (i.e., acoustic performances of sound phenomena).

The transmission of information between experts may be discrete (bits, bytes, words, or sentences), or analog (teleport effects, meanings, emotions, images, or intentions). The question arises whether discrete and analog signals completely restore the image or phenomena, and vice versa. Apparently, this issue does not have a good answer, because words and their meanings lie in completely different spaces. Moreover, the measurements cannot be reduced to the meanings of the analog signal, because the senses are much more complex than verbal representations and expert comments. This phenomenon is associated with paradoxes, contradictions, or complementary ideas, and we can use the quantum approach to compensate for its effects. However, there is another physical analogy that can be used to build additional mechanisms for improving an expert’s understanding during networked decision-making and group insight processes.

In digital signal processing, the sampling theorem [7] helps to build a fundamental bridge between continuous signals (in the analog domain) and discrete signals (in the digital domain). This theorem states that a continuous signal with a limited spectrum can be accurately reconstructed (interpolated) by its values at the different frequency intervals, with the condition that there is an upper frequency spectrum signal. Thus, a continuous but frequency spectrum-restricted signal can be exactly represented by a sequence of values defined at discrete points.

Real signals transmitted through communication channels can typically be specified with a frequency range within the essential part of its energy. This range defines the main part of the signal spectrum. Spectrum contraction distorts the signal. It follows that processes with bounded spectra can provide adequate mathematical models of many real signals; however, they are not absolute.

The application of this theorem to networked expert communications hints at fundamental limitations of the discrete linguistic representation of thoughts, meanings, and meditative states of consciousness, as transmitted in virtual collaboration processes. We can only transmit a signal with a limited spectrum using discrete methods. However, thoughts, meanings, meditations, and consciousness are phenomena that do not have restricted spectrums. Therefore, they cannot be encoded using discrete logic. One way of removing restrictions to the logic is to complement it using other techniques such as imaging context, discourse, or the interpretation of signals transmitted among experts in real time.

Visual analytics can help by improving experts’ mutual understanding during virtual collaboration and collective modeling processes. It can be defined as an approach that facilitates analytical reasoning using interactive visual interfaces [6]. Visual analytics during interactions between experts include data collection processes, processing, cognitive modeling, presentations of obtained information, human–computer interaction, and decision-making processes. Visual analytics help to understand and solve the problems by applying advanced automatic data processing techniques.

Visualization tools create a virtual context and virtual environment [1] for the decision-making process. Real-time interactive graphics and 3D-models combined with specialized display technologies immerse the user into the world of the virtual model. This provides opportunities for directly manipulating objects in model space. It supports a multidimensional immersion of networked experts into the decision-making problem, and gives a natural and intuitive interface for interacting with objects in a virtual environment. The result is a holistic discourse. The visualization approach is a metaphysical paradigm called natural interaction. It is based on the assumption that it is easier and more convenient for a person to interact with a computer if the input device is based on human feelings. Then, it is inevitable that the focus will be on analog mechanisms for information processing.

Cognitive convergence

Cognitive modeling

Simultaneously original and satisfactory solutions are a feature of an insight. The same conditions are features of the results of practical applications of expert systems, natural computing (genetic algorithms, swarm intelligence, membrane computing, artificial immune systems, etc.), and methods for solving inverse problems [10, 11]. The selection of the appropriate method for solving the problem is be based on the Analytic Hierarchy Process Method [12] that helps to rank methods with taking into account the problem’s specifics. Expert systems and natural computing have been used to support processes in medicine, mineral exploration, and for socio-economic control. In these cases, a certain inaccuracy in the simulation results is not an obstacle to the applications. The models in these cases do not give an answer; they only help to get the right answer. The real-life answer does not necessarily coincide with the simulation results. Cognitive modeling techniques are often used in these cases, where inaccuracy concepts are more important than quantitative and metric values.

A cognitive model (CM) of the decision problem consists of a directed graph that comprises factors, and connections between the factors. It may be represented by a matrix of interrelationships [11]. In the direct graph, the nodes designate factors and the arrows denote the corresponding relations. A CM involves such phenomena as concepts, analytic hierarchy, objects, denotations, association, synonyms, antonyms, connection, and influences. It is developed using the following sequence of procedures.

  • Gather information, identify the problem.

  • Select an appropriate method for solving the problem.

  • Formulate questionnaires based on fuzzy scales.

  • Find groups of experts.

  • Send questions to the experts.

  • Collect answers from the experts.

  • Build the computational CM.

  • Solve the direct problem, estimate the different scenarios of actions.

  • Solve the inverse problem and find the optimal decision.

  • Formulate the decision.

Direct problem-solving evaluates the temporal dynamics of the output factors that depend on different input impulse combinations. In a CM, the solution is optimized using the inverse cognitive problem-solving approach. The inverse problem-solving method uses soft computing techniques such as genetic algorithms (GA) to determine the optimal input combination that fulfills specified conditions for attaining the desired values for the output. The input factors are the control variables and the output factors are the goal variables. In inverse modeling, the decision-making participants search for the values of input values that result in the required output values.

Convergence semantic

Reality does not have a metric representation, and may be semantically interpreted using fuzzy, topological, and nonmetric spaces. The points, neighborhoods, sets, etc., semantically represent concepts in these spaces. Points may be “separated” from each other using operations on the intersections of their neighborhoods.

Working with the previously mentioned concepts and spaces, the solutions of the inverse problem include a stepwise introduction of qualitative information from experts into the decision-making process, with the hope of gaining an insight. An investigation of this method for achieving an insight reveals the following hyperplane and an operator for the semantic interpretation of the insight achieving path.

  • The purpose space U contains the exact purpose \(u_{T}\) and a set of inaccurate goals with index \(\delta \).

  • The resources of the goal achieving space Z contain the exact solution to \(z_{T}\) (“Eureka!”) and many inaccurate solutions with index \(\rho \).

  • The mapping operator is defined as \(A: Z \rightarrow U\).

  • The inverse operator R maps U to F, which, in turn, is a subset of Z.

In this case, we can narrow the domain Z of operator A to subsets F of Z, and the codomain U of A to a certain subset of U. Studies have shown that the expansion method can successfully solve inverse problems [9, 10]. It uses the direction of the necessary conditions that lead to a sustainable convergence of the solution on non-metric (topological) spaces, thereby forming a topological semantic of the insight.

The decision process is built on the assumption that a (not necessarily unique) solution exists, and is in the neighborhood of the exact solution (which cannot be obtained accurately). The solution, for example, could be represented as a fuzzy concept nomination in a CM. However, an approximate solution may be enough for someone who is striving to achieve an insight.

The basic obstacles to solving inverse problems are instability and incorrectness. It is much harder to reach a decision if Z and U are fuzzy spaces, because they follow regions of weak topologies, which are characterized as being non-metric.

For the decision-making process to converge (i.e., to stably and purposefully converge to a solution), the moderator must follow some rules. They must:

  • Separate the goals, resources (means), and actions;

  • Formulate the fuzzy image of the goals;

  • Categorize the goals using a hierarchy tree;

  • Identify all means by which these goals can be achieved using a finite (foreseeable) number of clusters (subsets);

  • Include any non-obvious and weak factors;

  • Build a “verbal bridge” (path) between means and goals;

  • Build a path to each goal from the means to achieve it; and

  • Estimate every path by solving a direct and inverse problem, and determine a convergent solution.

This list of rules includes necessary conditions for determining a sustainable and purposeful solution to the problem on the topological space that is a semantic interpretation of the CM. The solution will be found after one or two loops.

Thus, we can apply a CM to quickly find a result. To accelerate the insight achieving process, we should create the necessary framework conditions. In particular, a number of participants must collectively formulate the goals, determine the resources, and define an intuitively comprehensible set of concepts and influences (factors and connections between the factors).

The previously mentioned rules follow from the results of investigations into inverse problems on topological spaces. Compactness is one of the main conditions for convergence. So, the spaces Z and U are not necessarily metric. It is known that sustainable solutions require that the mapped spaces must be Hausdorff, and the space F must be bicompact [5]. The domain R does not have to be compact.

For ordinary compact topological spaces, every open cover of a topological space has a finite subcover. When representing the problem using a topological space, this condition can be interpreted by dividing the whole discourse space into a finite number of sets or classes. However, interestingly, this condition is not required for the goal space.

Cognitive programming

Holistic trinity path to targets

Solutions can be formed by human intentions. However, the path to the goal is created by a combination of analysis andromance. Analytics characterize the clarity of the logic and may be enhanced by a computer. The romance or spirit is created by poetry, emotions, and meditation. In each case, the analytics and romance are combined by something, thereby forming a holistic trinity (integrity).

Consider the following typical example. Somebody wants to build an image of the goal (future) through the imposition of the team members’ intentions. The participants try to convince their opponents and colleagues, so that they reach a common understanding of the issue using logical terms. Emotions generate deviations and give the discussion an alternative direction. This delays any agreement.

What is the decision-making team? A team, a collective, a group of experts—all these terms describe a number of people with common business interests, being interconnected by common goals and actions. The existing or growing arsenal of methods include strategic conversations, corporate culture development, consent, improvements to employees’ motivation, and improvements to management. The aim is the success of organizational structures whose elements are human beings.

A team may be large or small, have a networked organization or not, be recently designed, or possess many traditions. A team may have an ordered or chaotic structure, and be purposeful or uncoordinated. A team is always unique. It requires individual management of decision-making and insights. A new situation may call for a new combination of methods or techniques. It admits a simple explanation. The core of any organization consists of unique human beings. Their nature is unique, and, therefore, any group of human beings in different situations is unique.

There may be situations when team members try to find solutions to questions that have no clear answers, or where there are disagreements between the people regarding forecasts of the outcomes of scientific research or start-up work. One team member may prioritize the acquisition of intellectual property, and another may place more importance on evaluating marketing requirements.

The outcomes in these situations largely depend on understanding the question, the hidden interests of team members, and latent gaps in future inconveniences. Logic is not always helpful when dealing with such issues, because the true solution may be found in a different place or circumstance. The solutions to these issues require the following.

  • A representation of the issue that involves a holistic coverage of the various characteristics of the situation: logical, chaotic, semantic, and latent.

  • An accelerated convergence of the decision-making process to the unknown results (fuzzy goals).

  • Team decision-making processes that contribute to the accumulation of its intellectual capacities, reduce fear, and as a consequence, synergize the group thought.

In similar situations, the cognitive programming approach is useful for making decisions and achieving group insights.

There are two key questions regarding the specifics of cognitive programming. What is the cognitive form? And what is it as a holistic analytical category for representing political, economic, social, and technical forecasts?

Cognitive programming form

The logical (formalized) part of cognitive programming relies on classical methods for theoretical and practical cognitive modeling, and the theory of formal systems, algorithms, and traditional programming. Formal aspects of cognitive programming are based on the theory of directed and weighted graphs, and on system dynamics models. But unformalized phenomena cannot be represented by traditional logical constructions. The notion cognitive programming has the close meaning to the notions cognitive software or cognitive modeling. But in this case we introduce the new formalized definition of cognitive modeling that has advanced semantic interpretations of the logical constructions.

A traditional programming problem can be represented by:

$$\begin{aligned} \left\{ {{\begin{array}{ll} f({\varvec{x}})\rightarrow \text{ max }; \\ g({\varvec{x}})\le \mathrm{b}. \\ \end{array} }} \right. \end{aligned}$$
(1)

The solution to the problem in (1) can be used to maximize the value of the function \(f(\varvec{x})\) subject to \(g({\varvec{x}})\le \) b. Linear, nonlinear, dynamic, integer, multi-level, and fuzzy programming problems can be constructed so that they have a similar form. In this form, the logical constructs and transformations typically follow the axioms of logic.

Heuristic and artificial intelligence programming methods are different. The semantic interpretations have a significant role (see section “Insight Semantic”). Thus, each semantic factor represents a meaning (interpretation), given in certain way. Possible conversions of these meanings are used to construct conjunctions, disjunctions, and implications. Such conversions typically lie outside traditional programming concepts.

Semantic interpretations can be discussed in the frameworks of the fields of exegesis, semiotics, hermeneutics, and intelligent information technologies. For example, GAs are successful methods that can search for local extremum. They emulate genes and chromosomes.

Cognitive programming belongs to a class of problems from artificial intelligence, which require complex semantic interpretations. Cognitive programming can and must take into account the previously mentioned trinity. It forms a semantic interpretation of the logic elements, while ensuring the integrity (holistic and entanglement) of the coverage of the phenomenon under investigation. On one hand, it goes beyond traditional formalisms, and on the other it needs a formal explanation that uses, for example, algebraic semantics.

The three main components of the decision-making process and the path for gaining insight are connected by their relationships to each other (morphisms). The nature of these relationships depends on the issues raised, the formal representation methods, and the semantic interpretation of the formalisms.

A formal component model for cognitive programming uses N factors on a finite set of factors F (i.e., \(f_{i}\in F (i = 1, 2, {\ldots },N))\), their direct bonds (influences) \(c_{ji} \in C\) (where \(c_{ji}\) denotes the interference (interrelation) between the factors \(f_{j}\) and \(f_{i})\), and C (which is the set of the factors’ influences). The model is defined as:

$$\begin{aligned} \left\{ {{\begin{array}{ll} f_i ( t)=f_i({t-{1}})+\Psi ({f_i({t-{1}}),c_{ji}}),i,j=1,2,\ldots ,\hbox {N},\\ \mathrm{a}\le f_i \le \mathrm{b},\\ \end{array} }} \right. \end{aligned}$$
(2)

where time is discrete (for example, monthly, quarterly, or yearly), and \(\Psi (f_{i}(t-1), c_{ji})\) represents a morphism that defines changes to \(f_{i}(t)\) at time t-1. The morphism considers the connections \(c_{ji}\) between \(f_{i}(t)\) and \(f_{j}(t)\). Each factor can influence itself (reflexive relationship), and \(c_{ij}\) generally depends on a discrete time and has a semantic interpretation.

The semantic components of cognitive programming are defined by morphisms from the model’sformal components to a set interpretation system. When the set interpretation system of the cognitive model is represented using sets of documents, each interpretive document may be an abstract, book, comment or voice message, visual image, diagram, statistical regularity, survey result, the scale of an expert’s assessment of the relevance factor, etc.

In cognitive programming, a document extends the unformalized field of thought to form emotional, intuitive, or even meditative levels of human consciousness. The underlying phenomena are correlated with the formal part. Consequently, in the general case, we miss the direct recursive (formal logical) links between a set of documents that semantically interpret some factor \(f_{i}(t)\) or interrelation \(c_{ji}\) and the nomination of this factor. A factor (interrelation) cannot always be represented by some reference or statistically selected set of terms in the interpreted texts. However, this morphism can only be expressed by including a human in the cognitive programming process.

Thus, the semantic interpretation of the factor \(f_{i}\in F\) or interrelation \(c_{ij}\in C\) in cognitive programming can be represented by the family Q of subsets X that correspond to these factors and interrelations, and satisfy the following conditions:

(3)

These conditions characterize the family of semantic interpretation sets for cognitive programming. With respect to X, these conditions are satisfied by a topological space that can be extended to the fuzzy case [3]. Moreover, this space may not necessarily be a metric. Such a representation of semantic interpretation can be useful, for example, when finding the necessary conditions for sustainably finding an optimal path to the future, when the goals are not always clearly defined. Such a representation helps to create conditions for determining holistic analytical categories in networked expertise decision-making procedures.

E-expertise system

E-expertise

The idea of e-expertise (electronic or networked expertise) was presented in [4]. This book focused on mechanisms for expert support in decision-making processes using modern information, communication, information analysis, and collective intelligence technologies. The authors discussed the role of e-expertise in decision-making processes, provided a comprehensive overview of modern expert technologies with emphasis on the specifics of electronic expertise, described some new technologies for e-expertise, dealt with the problems of expert finding and grouping by information and communication technologies, and considered the trade-off between expertise stability versus strategic manipulations by experts or moderators pursuing individual goals.

The team decision-making process is frequently represented as a block diagram for choosing a certain result from several alternatives. This is used to evaluate the alternatives with defined criteria that depend on the goals, team experience, current situation, and selected method. Such an approach typically yields an inertial path that is not always appropriate, particularly in strategic situations. A principal (CEO, chief, leader, etc.) often possesses exclusive information and knows a specific situation in greater detail than other team members. They may plan something strategically ambitious that is not in line with the inertial path. When subordinates misunderstand a principal, motivation decreases and risks increase. To reduce risks and motivate subordinates, a principal should initiate a decision-making procedure that incorporates all their subordinates and experts.

A corporate team, including experts, is a collective of individuals with almost coinciding common interests, motivations, goals, and ideas. The interests of a member of a team do not necessarily obey a collective interest (it can be hidden). Moreover, the interests of a group do not represent the sum of the individual interests. In fact, a grouping that does not benefit the individuals is useless.

Unfortunately, many teams lack the knowledge, experience, ideas, or time to produce a holistic description of a problem and reach a good decision. In these conditions, we can mitigate for uncertainties using decision-making processes from a networked expert community. This results in a new paradigm for decision-making, where networked expert procedures (e-expertise) are used to inform the decision. The decision-making process may consist of the following steps [4]:

  • Learning about a problem using internal and external information sources like the results of sociological research, the media, government statistics, etc.

  • Specifying the method (more than 30 methods may be available).

  • Having convergent strategic conversations in a situation center [9], to:

    • construct a weighted tree of goals;

    • produce a cognitive model (see section “Cognitive modeling”);

    • collectively formulate action directions (using, for example, a SWOT-analyses method);

    • determining an optimal action plan (using a GA, for example);

  • Organizing and motivating the implementation.

  • Controlling the implementation.

This means that we must choose an optimal managerial decision under a set of qualitative and quantitative factors that affect the situation, including factors that characterize the feedback from the customers.

For the decision-making process to converge, we must determine the optimal structure of the information that is created during communications between different members, to encourage convergent decision-making and insight. An optimal structure means that we can analyze the processes of informational interactions between team members using uniform terms and concepts, by appropriately considering their specifics. This structure allows us to comprehend the logical and psychological schemes in collective decision-making. Using an insight convergent semantic approach, we suggest primary aspects: the goal-related aspect, the functional (regulation and communication) aspect, and the structural aspect. Furthermore, we can examine some additional aspects: the gnostic-analytical (cognitive) aspect, the priority-problematic aspect, and the resource aspect (ways, means).

Among the factors that are relevant to e-expertise viability, we should mention the normative legal base, financial provision, and historical, cultural, and territorial specifics of a certain state [4]. Experts communicate via a network medium, and sometimes do not see or know each other. Nevertheless, they master different techniques for intuitive and logical justification and skills that are difficult to formalize such as presentiment, prevision, divining, foretelling, etc. These are the skills and senses that are crucially important to e-expertise, decision-making, and gaining insight.

Networked expertise technologies

Networked expertise technologies can be implemented in many ways, e.g.,

  • Electronic expert comments;

  • E-expertise with semantic differential scales;

  • E-expertise for monitoring the current situation;

  • Electronic brainstorming;

  • Networked strategic conversation (up to 25 participants);

  • Networked strategic congress (up to 250 participants); or

  • Crowdsourcing within a self-organized networked expert community.

Electronic form of expert comments This is a procedure for accumulating experts’ opinions on an issue by analyzing their comments, conceptual modeling, and recommendations. Experts must explain interconnected proposals on the structural aspects of a problem such as goal-setting, descriptions of relevant factors that influence the situation, mutual influences of factors, obstacles to achieving goals, and problem solving methods.

E-expertise with semantic differential scales Experts give answers to an open or closed set of questions using estimation scales (semantic differential scales, linguistic scales), which guarantee a high-level of mutual understanding between the participants. Differences are due to linguistic scales for questions and automated aggregations of the experts’ estimations using averaging or hierarchy analysis methods.

E-expertise for monitoring the current situation This is implemented in two modes: solving a specific problem and current monitoring of a subject area. Problem-related monitoring is initiated by a principal. This monitoring result fixes the corresponding action plan to a problem or special decision. In each situation, problem-related monitoring needs special procedures, simulation methods, and polling forms. Permanent monitoring is implemented on a continuing basis. Experts share their appraisals of a situation during given periods. For instance, an expert provides a special-form report at least once a month.

Electronic brainstorming This is supervised by experienced moderators, and is a procedure for rapidly generating nonstandard ideas and proposals. When experts communicate online and see each other only on computer displays, traditional brainstorming methods appear inefficient. Using text and voice messages, networked experts must understand each other without ambiguities in the shortest possible time. Additionally, the procedure must use automatic methods for the semantic treatment of numerous messages so that it is properly controlled. This helps to extract a collective idea from the chaos of text proposals.

Anetworked strategic conversation (see section “E-expertise”) can be conducted under the principal’s chairmanship. Two or three networked moderators drive the conversation if there are less than 20–30 participants. A networked conversation possesses a strategic orientation, i.e., it implies consentient goal-setting, problem formulation, and descriptions of future actions. Owing to special conditions, all distant participants can reach a mutual and unambiguous understanding as fast as possible, by exchanging their messages and multimedia tools. To aid understanding, a moderator uses visual tools for virtual collaboration [13].

Networked strategic congress This e-expertise procedure is a natural extension of networked strategic conversations. There may be up to 250 direct participants in a networked strategic congress. The organization of such congresses may involve over 25–30 networked moderators. A networked strategic congress represents a rather new electronic mechanism for strategic control in socioeconomic systems, with the well-timed conciliation of interests pursued by different public subjects (citizens, societies, corporations, regions, authorities, etc.). Here, one can use special techniques to accelerate the process of reaching a consensus of all congress participants on the desired goals and necessary means.

Crowdsourcing with a self-organized networked expert community This requires mechanisms for expert self-organization, which is affected by the principal’s intentions and the expectations and needs of the population. This mechanism favors fast revelations of original ideas, factors, goals, and proposals, an appreciable reduction of the risks, and prevents the negative consequences of decisions. Exactly self-organized environments have networked leaders and identify talented moderators and experts. They may involve some exploitation of content- and connect-analysis systems. A crowdsourcing system must be integrated with one or more of the above expert technologies.

Networked expertise technologies are described in more detail in [4], which proposed an e-expertise system. The software in this system was created using a cognitive programming approach (see section “Congnitive Programming”), was written in Ruby Mine, and consists of modules (see Fig. 1).

Fig. 1
figure 1

Modules of the E-expertise software system

Full size image

Legal, motivation, training provisions

Further developments of legal provisions are vital for e-expertise. Mechanisms must be developed for networked expert communities, explaining the well-defined statuses and motivations of experts, and information security problems. Of particular importance are: trust, conscientiousness, responsibility, and the mutual understanding of experts. Further developments to the legal provisions of e-expertise require more attention to conscientiousness. The concept of conscientiousness must be applied to the rights and duties of different subjects in an expert activity. Conscientiousness can be described as the morally acceptable behavior of different participants in the expert activity, which agrees with existing public morals regarding integrity, utility, and harm. We must also assess the ethics of an expert’s actions in a specific situation. Formal attributes for this assessment may include certain information regarding an expert’s awareness during their decision-making process. Other attributes can be used to judge the intentions behind an expert’s actions. It seems rational to create a professional ethics code for networked experts.

To be efficient, an expert must take on personal responsibility. Maximum centralization is required for the methodological and practical guidance of expert activities. However, contributions to problem solving require maximum decentralization. The responsibility of experts is predetermined by their legal status and involvement in real decision-making. Experts can make mistakes, e.g., they may underrate an important factor or overrate a common factor.

An expert may sometimes make recommendations but take no responsibility for their implementation. A contract may be used so that an expert has certain obligations to compensate possible losses caused by planned risks or bear financial and property liabilities. However, this will reduce the expert’s creative activities.

An expert possesses social responsibility. There are several types of social responsibilities related to expert activities (for example, legal, professional, and reputation-related social responsibilities). These require clear regulations. The copyright or intellectual property rights for all the materials provided by experts generally belong to the experts. When information is provided by experts with certain compensations, intellectual property rights for resulting products can be transferred to a customer. This can occur when an expert registers his intellectual property according to existing rules.

Information technologies can be used to document the ideas and proposals of experts. This enhances the capabilities of intellectual property protections. Meanwhile, by operating in information space, we presuppose legal supports for many aspects of intellectual property protection.

We must continuously improve the laws that support further growth in e-expertise, and appropriately protect the networked expert’s intellectual property that is the result of networked intellectual activities. Refining the intellectual property protection system encourages the creative activities of experts.

Generally, the existing financial mechanisms and legislature of a country do not lead to an efficiently operating expert community. We must use special approaches. International experience shows that a special type of contract is needed to increase the efficiency of expert activities, i.e., contracts for expert services and expert research work. The underlying reason for this is that expert services are significantly different from other services. First, an order for an expert service may appear spontaneously and rapidly. Second, there are substantially greater uncertainties in the intellectual property, effectiveness, reliability, creativity, utility, riskiness, responsibility, conscientiousness, independence, trust, emotionality, rating assignment, and so on.

Financial mechanisms for expert activities should be focused on: separating a dedicated branch of the law, guaranteeing the responsibility of state authorities when using expert information, creating and ensuring sponsorship funds to independently support expert activities, sharing the reimbursements of costs and risks between a state customer and an expert company (a contractor), and so on.

Moreover, the formation mechanisms of state and corporate needs regarding expert information must be defined on the basis of several information flows: from state authorities, scientific organizations, business communities, and other expert activity participants. This motivates expert activities.

Developments in networked expert activities motivate experts through, for example,

  • Participation in managerial decision-making;

  • Growing personal reputations and abilities;

  • Rating expert assignments;

  • Acquiring exclusive analytical information;

  • Free analytical and computational services;

  • Accreditation, certification and decorations; and

  • Protections to the expert’s rights.

An expert possesses certain knowledge and skills that can be applied to a networked activity. Nevertheless, the networked operations of expert groups require additional knowledge for synchronizing collaborations, methodological support, and improving mutual understandings. This is the training of experts.

The amount of training needed depends on the role of a specific expert within the expert activity. Some knowledge can be updated annually. Additional training can take many forms. The most important training topics concern the organization of group networked analyses of different situations, consistent descriptions of decisions, and networked expert procedures. Programs for this training may include material from disciplines such as philosophy, psychology, political science, sociology, mathematics, physics, management science, solutions of inverse problems, controlled chaos, and quantum semantics. All the study guides and textbooks must be available on expert community portals.

Practical application

Practical approach

Many decision-making problems from e-expertise systems have similar characteristics. The consumer requests, reputation, research and development, technological, political or other factors can affect the success of the decision-making process. The success of large manufacturing projects can, for example, be measured in terms of political, economic, social and technological effects. It is convenient to represent forecasts of the project’s success using a CM (see section “Cognitive modeling”), which estimates factors and connections using numerical or fuzzy scales. We use 5- or 7-point grading scales and during cognitive programming the scales are normalized. The success of a decision always depends on a holistic representation of the problem and the correct structure of the information.

First, we must collect all the information, choose the method (see section “Cognitive modeling”) and Networked Expertise Technology (see section “Networked expertise technologies”), and sketch the cognitive model of the problem using the convergent approach [10, 11]. This provides the necessary structure of the information so that the performance is optimized for scientific, technical, and commercial purposes. The questions to the experts are formed by considering that success is dependent on more than 100 factors that characterize the problem. There are three fundamental factors that are defined by convergent approach and well associated with the holistic trinity idea (see section “Holistic trinity path to targets”):

  • Management (marketing, organization, finance, and so on.);

  • Reputation (merits, popularity, confidence, leadership, and so on.); and

  • The complexity (holistic, integrity) of the problem.

It is essential to assess the project success in terms of the experts’ answers to questionnaires, and consider a finite number of factors, and connections between factors. Examples of questions include

  • What are the growth dynamics of the team’s intellectual capacity;

  • What is the required level of research and development investment; and

  • What is the market volume?

Each question is accompanied by a fuzzy scale and a request for a comment. The experts’ answers are processed and used to create the CM. The CM considers organizational, political, economic, social, and technological factors. To improve the quality of the forecast, the team may attempt to anticipate the risks in the CM. In this CM, the specific combined interrelationships between factors generate a synergy. Any known or unknown factor could have a crucial role, because of its effects on other factors or relationships.

After creating the CM, the analysts attempt to find direct and inverse solutions to the problem. The direct solution answers questions such as “What happens, if we do something?” The direct solution shows the comparative chances of success for different scenarios.

To obtain an optimal solution, we must solve the inverse problem. The investigated situation may behave in an unpredicted or the participants may not be sure of a holistic representation of the problem. Then, the hidden factors must be revealed by networked experts. A GA may be used to solve the inverse problem. The solutions to an inverse problem answer questions such as “What should I do to ensure that the project is successful?”

Fig. 2
figure 2

Cognitive model for making decisions on the space project. The nodes represent factors and the arrows represent the corresponding relationships, which characterize the managerial situation. This picture visualizes the main step of the project, that is, the rocket launch

Full size image

Practical examples

We verified the proposed approach by considering strategies for the complex reconstruction of the territories of Moscow, the development of the Russian information technology market, and the Russian–Israel science and innovation partnership. We also considered higher and professional education, the public health service, social security programs, housing and communal services, property management, population policy, youth policy, and television viewer rating calculation [4, 8, 10]. Our results were successfully implemented to develop several governmental situation centers in the Russian Federation and to create business strategies for some commercial companies.

For example, we used our approach to forecast the successes and risks of a large space project. Our method combines an e-expertise system, convergence semantic, and cognitive programming (see section “Congnitive Convergence”, “Congnitive Programming”, and “E-expertise system”). A group of 17 experts created the cognitive model (Fig. 2).

Fig. 3
figure 3

Distribution recommendations. The result of GA calculation—after the set non-optimal solutions it gives the local optimal solution

Full size image

We formed a list of questions for the experts, on the basis that the success of the project was influenced by several underlying factors. It is important to take into account interference between the factors, because mutual influences generate synergy. Therefore, any insignificant factor may be important because of its relationship with other factors.

We solved the inverse problem using a GA, which generated more than 700 populations (sets) with candidate solutions before determining the optimal solution. The GA provides recommendations for redistributing project development efforts. Figure 3 shows the recommendations, which are represented by points of the optimal solution on normalized fuzzy scales (between \(\pm \)1). Note that the GA provides different local optimal solutions (local extremum) that are not always a global optimum. The manager must make his own decision on the basis of the local optimal solution. This was also confirmed when we applied the GA in practice.

Figure 3 shows that the Principal, Management, R&D, and Consumer factors must be distributed in a different way.

Discussion and conclusion

Discussions concerning approaches for accelerating group decision-making processes using the concept of convergence began in 2008 [10]. However, the convergent method for decision-making that was based on inverse problem solution techniques was suggested at the beginning of 1990. This method has been discussed during many international scientific conferences in Russia, Italy, and Cyprus, among others. Additionally, it has been implemented in real-world applications. This has demonstrated that structuring information in the above-mentioned way can speed-up group decision-making processes. However, discussions have raised questions regarding holistic discourse in cognitive modeling, which have been answered in [11]. The question of how to make non-causal and simultaneously good decisions was always the focus of interest during discussions. This work is a modest attempt to find the answer using the inverse problem solution and the logic behind a breakthrough thinking approach.

The e-expertise system can be used to speed up networked group decision-making insights and consensuses. The system was based on a convergent approach, which exploits the fundamental principles of control thermodynamics, inverse problems, holistic discourse, cognitive modeling and programming, quantum semantics, artificial intelligence, virtual collaborations, unconscious thinking, and breakthrough thinking approaches.

The e-expertise system can be used by different groups of government officials, stakeholders, policy-makers, researchers, innovators and educators, who all share the common concern that public investment in information and communication technology and electronic governance creates public value. It may be useful for supporting targeted forum discussions and creating effective strategies for group actions across national, thematic, development, political, and other borders.