Search

Search Results

• ALGORITHM FOR FILTERING "HINT INJECTIONS" WHEN USING SPATIAL INFORMATION

  S.L. Belyakov , L.А. Izrailev , О.N. Pokusaev

  32-45

  2026-07-07

  Abstract ▼

  The integration of large language models (LLM) into geographic information systems (GIS) opens up new opportunities for spatial analysis, but it is accompanied by specific vulnerabilities such as "hint injection" (prompt injection). Such attacks allow attackers to bypass LLM security mechanisms, manipulate issuance, gain access to confidential information, and violate data integrity. Using space allows you to access an object not directly, but through its spatial relationships with other objects. Existing keyword or template filtering methods do not provide reliable protection due to the constant emergence of new attack scenarios. This determines the relevance of developing adaptive, self-learning algorithms for filtering queries for industrial injections to large language models. The aim of the study is to develop an algorithm for filtering prompt injections for LLM, based on the Case-Based Reasoning (CBR) method. The paper proposes an algorithm for comparing LLM queries with a database of previously known promt injections. The experiment showed that as the database of use cases accumulates, the accuracy of detecting prompt injections increases from 42% to 83%. At the same time, the processing time for a single request increases slightly (from 0.18 to 0.19 seconds with a 23% increase in the database). Approaches to generalizing the precedent base and introspection of the precedent base were also proposed. The proposed algorithm makes it possible to increase the security of LLM-interface systems against prompt injections due to adaptivity and self-learning. The practical significance lies in the possibility of implementing the developed filter into information systems to prevent leaks and manipulation of spatial data. Further research is related to the development of methods for automatic generalization of use cases and the integration of additional contextual analyzers.

• TRANSFORMING THE DECISION-MAKING EXPERIENCE

  S. L. Belyakov, M. L. Belyakova, S.A. Zubkov, N.A. Golova, K.S. Yavorchuk

  2021-01-19

  Abstract ▼

  The problem of transferring the experience of decision-making in situational analysis using
  geoinformation systems is considered. The need for intellectual support from the geoinformation
  system is due to the fact that spatial objects and connections of the real world are extremely dynamic.
  Under these conditions, it is not possible to apply analytical models of processes and phenomena
  due to the incompleteness and inconsistency of the information describing them. Statistical
  models depend on a large number of factors, which vary as the geographical location of the
  situation changes. An alternative is to use the experience of experts who are able to make effective
  decisions in local spatial situations. The lack of control over the reuse of experience is a problem.
  The knowledge gained in developing solutions in one locality can lead to inadequate solutions in
  another locality. The experience of solving a problem in the same area loses its significance over
  time. In this paper, we propose a representation of knowledge in the form of an image consisting
  of a center and acceptable transformations of the center. Image transformation functions that
  perform knowledge transfer are introduced. The properties of transformation functions that carry
  procedural knowledge about the images of situations are analyzed. The use of the identified properties
  for the formation of a test plan for software transformation procedures is considered. Thecriteria for successful transformation are studied. The optimization problem of finding the best
  transformation function in the GIS knowledge base is formulated. A generalized method of transforming
  experience is proposed. An example of the synthesis of transformation methods for selecting
  an operational call center is given. The image of the situation of making a decision about
  choosing a land plot for a service center is transformed into the specified area on the GIS map.

• KNOWLEDGE FOR ARGUMENTATION IN COMPARISON OF SPATIAL SITUATIONS

  S.L. Belyakov, N.А. Golova, К.S. Yavorchuk, I. N. Rosenberg

  2022-05-26

  Abstract ▼

  The traditionally used way to assess the quality of the solution proposed by an intelligent
  system is to explain the course of logical inference. Knowledge about reasoning is used to argue
  the choice of a solution option. The sequence of applied rules, the facts used and the confirmed
  hypotheses are considered arguments that should convince the user of the validity of the formed
  conclusion. The disadvantage of this method of explanation is that it reflects a formally correct,
  but devoid of semantic content, course of reasoning. The argumentation of the solution obtained is
  based on the tracing protocol, which is essentially no different from debugging information when
  tracing programs. The argumentation in this case is far from the meaning of the situation. The
  meaning is understood as a given set of transformations of the situation that preserve the immutability
  of its perception by a human analyst. Knowledge about the semantic content of situations
  should be presented in a special fashion. In this paper, we consider a representation containing a
  precent and its permissible transformations. In this form, spatial situations in geoinformation systems
  are described. For argumentation, it is proposed to use special relations between images of
  situations. The concept of the area of applicability of the image is introduced. The mutual arrangement of the spatial-temporal and semantic shell of images and the areas of their applicability
  is considered as a carrier of the relationship. Information about relationships is extracted from the
  structure of the cartographic database. The relations of inheritance, aggregation, composition,
  generalization and association of classes of objects are considered. Knowledge for argumentation
  is provided by the rules for determining the reliability index of expert conclusion for individual
  relationships and their combinations. A method of automatic rule generation is proposed.
  The relations for comparison of levels of reliability of rules are given.

• INTELLIGENT RECOMMENDER SYSTEM FOR SPATIAL ANALYSIS

  S.L. Belyakov, А.V. Bozhenyuk, N.А. Golova, К.S. Yavorchuk, I.N. Rosenberg

  14-26

  2025-07-30

  Abstract ▼

  The work is devoted to the analysis of mechanisms of formation of recommendations and
  evaluation of the user's reaction to them in the interactive mode of work with the geoinformation
  system. One of the important areas of application of recommender systems is the search and decision-
  making in spatial situations. A peculiarity of this class of problems is the uncertainty of task definition and ambiguity of decision evaluation. Users are often faced with problems that do not
  have a clear formulation. To try to solve them, it is necessary not only to designate the direction of
  solution search, but also to find an adequate sequence of tasks with clearly formulated input and
  output data. Recommendations in such cases are designed in a dialogue with the user-analyst to
  develop a strategy for finding solutions. In this paper we study a smart recommendation system
  using the experience of dialog interaction. We propose a model of adaptation to the mental image
  of the problem, which builds the user, taking into account the levels of situational awareness and
  cognitive load. The peculiarity of the model is the use of visual cartographic objects, which are
  indicators of the state of the mental image. A recommendation is represented by a set of objects
  that are introduced into the field of cartographic analysis. This implicitly induces a certain semantic
  direction of increasing situational awareness. A criterion of satisfaction with the recommendation
  is suggested. A diagram of recommender system states, which describes the selection of context,
  adequate to the problem being solved, is given. The context is understood as an information
  object, capable of providing program functions and data for solving problems of a limited class.
  A sequence of contexts in an analysis session is considered as a precedent of experience. Indicators
  of trend, tendency and rhythm are proposed for possible chains of contexts. The degree of
  semantic proximity of precedents to the current course of search for a solution is estimated by
  these indicators. Their use will increase the speed of adaptation

• INTELLIGENT RECOMMENDER SYSTEM FOR SPATIAL ANALYSIS

  S.L. Belyakov, А. V. Bozhenyuk, N. А. Golova, К.S. Yavorchuk, I.N. Rosenberg

  14-26

  2025-07-31

  Abstract ▼

  The work is devoted to the analysis of mechanisms of formation of recommendations and
  evaluation of the user's reaction to them in the interactive mode of work with the geoinformation
  system. One of the important areas of application of recommender systems is the search and decision-
  making in spatial situations. A peculiarity of this class of problems is the uncertainty of task definition and ambiguity of decision evaluation. Users are often faced with problems that do not
  have a clear formulation. To try to solve them, it is necessary not only to designate the direction of
  solution search, but also to find an adequate sequence of tasks with clearly formulated input and
  output data. Recommendations in such cases are designed in a dialogue with the user-analyst to
  develop a strategy for finding solutions. In this paper we study a smart recommendation system
  using the experience of dialog interaction. We propose a model of adaptation to the mental image
  of the problem, which builds the user, taking into account the levels of situational awareness and
  cognitive load. The peculiarity of the model is the use of visual cartographic objects, which are
  indicators of the state of the mental image. A recommendation is represented by a set of objects
  that are introduced into the field of cartographic analysis. This implicitly induces a certain semantic
  direction of increasing situational awareness. A criterion of satisfaction with the recommendation
  is suggested. A diagram of recommender system states, which describes the selection of context,
  adequate to the problem being solved, is given. The context is understood as an information
  object, capable of providing program functions and data for solving problems of a limited class.
  A sequence of contexts in an analysis session is considered as a precedent of experience. Indicators
  of trend, tendency and rhythm are proposed for possible chains of contexts. The degree of
  semantic proximity of precedents to the current course of search for a solution is estimated by
  these indicators. Their use will increase the speed of adaptation.

• GEOINFORMATION MODELS OF EMERGENCY SITUATIONS WITH SPATIAL GENERALIZATIONS

  S. L. Belyakov, L. А. Izrailev

  2025-01-30

  Abstract ▼

  The main problem of decision making in emergency situations is the reliability of these decisions. Emergency
  situations by virtue of its unpredictable and dynamic nature often have incomplete and inaccurate information.
  The use of accumulated experience allows to find reliable solutions based on known precedents of
  emergency situations. Geographic information systems (GIS) can act as a tool for accumulating experience and
  generating solutions based on it. The cartographic basis of GIS allows analyzing emergency situations, taking
  into account their spatial and temporal characteristics. However, the cartographic representation of precedents
  with adopted solutions describes them too narrowly. There is no idea what properties of the situation are significant
  and whether the precedent solution can be applied in other circumstances. The use of known images and
  their admissible transformations, created on the basis of expert knowledge, can solve this problem. The image
  generalizes a set of similar precedents. The purpose of such generalization is to expand the area of application
  of information from private observations by determining the boundaries of permissible transformations. However,
  the need to attract experts for their creation is a difficult task, since each situation is unique in its own way.
  No less problematic is the transfer of experience from one spatial and temporal domain to another. In this paper
  we consider an approach to automatic image generation. We propose a method of creating a geoinformation
  model of emergency situations, which includes the generalization of precedents on a common location. This
  approach is aimed at improving the reliability of prediction of emergency situations. An experiment was conducted
  to synthesize images based on precedents of road accidents and evaluate their effectiveness compared to
  individual precedents. The use of the developed method of automatic data processing to create images is relevant,
  as it significantly reduces the cost of knowledge acquisition. The use of spatial generalizations also eliminates
  the need for expert knowledge, since the formation of precedent sets is performed by analyzing their geographical
  location.

• METHOD OF SUPPORTING THE STABILITY OF THE POWER SUPPLY NETWORK BASED ON A GEOINFORMATION MODEL

  S.L. Belyakov, А.V. Isaev

  2024-05-28

  Abstract ▼

  The article considers the problem of controlling the distribution of energy power in an area covered
  by an intelligent energy network. The management objective is to stabilize the energy flow in the presence
  of external influences caused by changes in the surrounding environment. Vulnerabilities in the system are
  inherent due to the nature of energy networks and under certain circumstances can lead to anomalies in
  energy supply. External environmental factors vary in content, making it difficult to confidently predict
  current threats. Geoinformation models utilizing image-based knowledge representation are described.
  Their use enables the assessment of the relevance of known threats. Conceptually, an image comprises a
  center and permissible transformations of that center within a certain context. The case is considered
  where the threat to the functioning of the intelligent network is assessed by transforming the image into a
  specified area of space where the intelligent network is located. The key feature of the proposed approach
  is the evaluation of the feasibility of an event occurring in a given space. The operation of transferring the
  situation requires consideration of the topology of the specified area. The attributes of the generating
  infrastructure become more significant than the attributes of the situation itself in this approach. A distinctive
  feature of the proposed approach is the transfer of semantic context represented by permissible transformations
  of the image. The software transformation function is linked to a layer of cartographic representation.
  For a given object in the original precedent, its placement area is determined, with the boundary
  being defined by the object's properties. If the size of the placement area allows for the construction of
  an object of the corresponding class, that object is created. The credibility of the result is evaluated by
  applying expert knowledge about the quality of objects of the considered class. The listed actions are performed
  not only on the geometry of spatial objects but also on their temporal and semantic attributes, akin
  to the concept of image-based representation of geometry. Forming a list of threats given a specific state
  of the external environment constitutes the essence of stability management. The features of algorithmizing
  the image transformation procedure are analyzed, and a method for assessing the credibility of transformation
  is provided. The application of the proposed approach holds promise for intelligent energy supply
  systems, whose behavior is intricately linked to external environmental factors.