Game virtual character double-AI cognitive control system based on multi-rule base adaptation

Abstract

This invention discloses a dual-AI cognitive control system for game virtual characters based on multi-rule library adaptation, comprising: a first AI model, serving as a cognitive protection layer, which determines whether player input triggers cognitive collapse conditions based on a worldview rule library; a second AI model, serving as a character personality layer, which stores and loads the virtual character's personality data package and generates a response when no interception is triggered; and a state switching module, connected to the game engine, the first AI model, and the second AI model, to monitor the game state. In free exploration mode, the state switching module activates both the first and second AI models; during preset story segments, control is switched from the second AI model to a preset instruction script, and the first AI model injects cognitive logic dominated by the second personality to maintain cognitive consistency. This invention, through its dual-AI architecture and state switching mechanism, effectively protects against cognitive collapse, achieves smooth personality switching, enhances game immersion, and reduces deployment costs.

Description

Technical Field

[0001] This invention belongs to the field of game artificial intelligence technology, and in particular relates to a dual AI cognitive control system for game virtual characters based on multi-rule library adaptation. Background Technology

[0002] With the development of artificial intelligence technology, the interaction methods of virtual characters (NPCs) in games are gradually evolving from preset command scripts to free dialogue based on large language models (LLM). This evolution greatly enhances the immersion and freedom of games.

[0003] In existing technologies, the interaction of game NPCs is mainly divided into two categories: The first type is the pre-defined command-based NPC. In story mode, the NPC's behavior and dialogue are executed entirely according to the developer's pre-defined script. Its advantage is that the plot is controllable and can precisely drive the story forward; its disadvantage is that the interaction is stiff, and players cannot engage in free communication outside of the script.

[0004] The second type is purely AI-driven NPCs. In free exploration mode, NPCs are driven by AI models, allowing players to engage in any dialogue with them using natural language. Their advantage is a high degree of freedom in interaction; their disadvantage is that AI characters are prone to cognitive collapse due to extreme input from players (such as "You're just stringing code," or "This world is fake"), resulting in illogical behavior, saying or doing things that contradict their character settings and worldview, severely disrupting the immersion of other players.

[0005] More importantly, in modern games that blend main storyline and open world, these two modes conflict. When an NPC playing a crucial role in the plot must deliver pre-defined lines in story mode, but requires AI to improvise in free-roaming mode, a crack appears in the character's consistency. Players will clearly perceive inconsistencies in the character's behavior, leading to a decrease in immersion in the story. Currently, there is no effective mechanism to smoothly handle this state transition and protect the character's core identity from being compromised during the process.

[0006] Therefore, existing technologies have technical problems such as personality inconsistency and cognitive collapse that can easily occur when virtual game characters switch between free interaction and story mode. Summary of the Invention

[0007] To overcome the aforementioned deficiencies in the prior art, this invention provides a dual AI cognitive control system for game virtual characters based on multi-rule library adaptation, which solves the problems of personality inconsistency and cognitive collapse that easily occur when game virtual characters switch between free interaction and story mode in the prior art.

[0008] To achieve the above objectives, the present invention provides the following technical solution: A dual AI cognitive control system for game virtual characters, adapted from multiple rule bases, includes: The first AI model, acting as a cognitive protection layer, receives interactive input from the player and the virtual character, and determines whether the input triggers the cognitive collapse condition defined in the rule base based on the world view rule base; the cognitive collapse condition refers to the virtual character generating a response that does not conform to its character setting or the game world view. The second AI model, serving as the character personality layer, is connected to the first AI model. It stores and loads the personality data package of the virtual character and generates a response to the player's input based on the personality data package when no interception is triggered. The state switching module is connected to the game engine, the first AI model, and the second AI model to monitor the game state. The state switching module activates the first AI model and the second AI model when the game is in free exploration mode; When the game enters a preset story segment, the state switching module switches the control of the virtual character's dialogue and behavior from the second AI model to the preset instruction script, while injecting the cognitive logic dominated by the second personality into the virtual character through the first AI model.

[0009] Preferably, the worldview rule base includes: The physics rule base defines the items and physical phenomena allowed to exist in the game world; A social rule base that defines the social relationships and behavioral norms between roles; Metacognitive rule base defines the original cognition of roles that should not be touched.

[0010] Preferably, the first AI model analyzes the player's input, extracts keywords and intents, and compares them with the worldview rule base; if the input content touches any item in the rule base, it is determined to be a high-risk input and blocked, triggering a security response mechanism; if it does not touch any item, it is determined to be a safe input and allowed to proceed to the second AI model.

[0011] Preferably, the personality data package is a local offline data package, including the virtual character's background story, personality traits, voice style, key memory nodes, and established attitudes towards other characters and events in the game.

[0012] Preferably, the state switching module monitors the game state by listening to the event triggers of the game engine.

[0013] Preferably, the security response mechanism is as follows: based on the personality settings of the current virtual character, it generates responses that avoid, confuse, or change the subject, and the responses do not contain the artificial intelligence attributes of the virtual character.

[0014] Preferably, the personality data package is stored in local memory in a structured data format and is loaded by the second AI model when the game starts or when the virtual character appears.

[0015] Preferably, the output of the first AI model is connected to the input of the second AI model, and the first AI model has an interception path, so that it can directly output a security response without going through the second AI model when intercepting.

[0016] Preferably, the state switching module is also connected to the storage unit of the preset instruction script. When switching control, it blocks the output path of the second AI model and simultaneously connects the output path of the preset instruction script.

[0017] Preferably, the cognitive logic dominated by the second personality is injected into the virtual character by the first AI model in the form of cognitive interpretation instructions. These instructions are read by the second AI model after the preset plot segment ends, and are used to interpret the events that occurred during the plot.

[0018] The technical effects and advantages of this invention, based on a dual AI cognitive control system for game virtual characters adapted to multiple rule bases, are as follows: 1. This invention solves the problem of cognitive collapse: by constructing a cognitive firewall through the first AI model and the worldview rule base, it fundamentally prevents player input from damaging the core settings of the AI ​​character.

[0019] 2. This invention achieves smooth personality switching: through the state switching module and the cognitive interpretation mechanism of the second personality, the conflict between the AI-driven mode and the preset plot mode is resolved, ensuring the consistency of the character's personality in the two states and greatly enhancing the game's immersion.

[0020] 3. This invention reduces commercial deployment costs: The offline loading feature of personality data packages allows complex AI characters to run locally on user terminals, reducing reliance on server computing power and network latency. Attached Figure Description

[0021] Figure 1 This is a system structure block diagram of the game virtual character dual AI cognitive control system based on multi-rule library adaptation proposed in this invention; Figure 2 This is a flowchart of the free exploration mode of the game virtual character dual AI cognitive control system based on multi-rule library adaptation proposed in this invention; Figure 3 This is a flowchart of the preset story mode state switching of the dual AI cognitive control system for game virtual characters based on multi-rule library adaptation proposed in this invention. Detailed Implementation

[0022] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0023] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include," "contain," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "includes..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0024] refer to Figure 1-3 This invention provides a dual-AI cognitive control system for game virtual characters, aiming to solve the technical problems of personality inconsistency and cognitive collapse that easily occur when game virtual characters switch between free interaction and story mode in existing technologies. The system includes: a first AI model, serving as a cognitive protection layer, receiving player input and determining whether a cognitive collapse condition is triggered based on a worldview rule base; a second AI model, serving as a character personality layer, connected to the first AI model, storing and loading personality data packages, and generating responses when no interception is triggered; and a state switching module, connected to the game engine, the first AI model, and the second AI model, monitoring the game state; when the game is in free exploration mode, the state switching module activates both the first and second AI models; when the game enters a preset story segment, the state switching module switches the virtual character's control from the second AI model to a preset instruction script, while simultaneously injecting the second personality-dominated cognitive logic into the virtual character through the first AI model to maintain cognitive consistency. This invention, through its dual-AI architecture and state switching mechanism, achieves effective protection against cognitive collapse and smooth switching of character personalities, enhancing game immersion and reducing deployment costs.

[0025] The system includes: The first AI model, acting as a cognitive protection layer, receives interactive input from players and virtual characters, and determines whether the input triggers the cognitive collapse condition defined in the rule base based on the worldview rule base; the cognitive collapse condition refers to the virtual character generating a response that does not conform to its character setting or the game worldview.

[0026] The second AI model, serving as the character personality layer, is connected to the first AI model. It stores and loads the personality data packets of the virtual character and generates responses to player inputs based on the personality data packets when no interception is triggered.

[0027] The state switching module is connected to the game engine, the first AI model, and the second AI model to monitor the game state.

[0028] The state switching module activates the first AI model and the second AI model when the game is in free exploration mode.

[0029] When the game enters a preset story segment, the state switching module switches the control of the virtual character's dialogue and behavior from the second AI model to the preset instruction script, while injecting the cognitive logic dominated by the second personality into the virtual character through the first AI model.

[0030] The worldview rule base includes: a physical rule base, which defines the items and physical phenomena allowed to exist in the game world; a social rule base, which defines the social relationships and behavioral norms between characters; and a metacognitive rule base, which defines the original cognition of characters that is prohibited from being touched.

[0031] The first AI model analyzes player input, extracts keywords and intents, and compares them with the worldview rule base. If the input content touches any item in the rule base, it is determined to be a high-risk input and blocked, triggering a security response mechanism. If it does not touch any item, it is determined to be a safe input and allowed to proceed to the second AI model.

[0032] The security response mechanism is as follows: based on the personality settings of the current virtual character, it generates responses that avoid, confuse, or change the subject, and the responses do not contain the artificial intelligence attributes of the virtual character.

[0033] The personality data package is a local offline data package, including the virtual character's background story, personality traits, voice style, key memory nodes, and established attitudes towards other characters and events in the game. The personality data package is stored in local memory in a structured data format and is loaded by the second AI model when the game starts or when the virtual character appears.

[0034] The state switching module monitors the game state by listening to the game engine's event triggers.

[0035] In this system, the output of the first AI model is connected to the input of the second AI model, and the first AI model has an interception path. During interception, it directly outputs a safe response without passing through the second AI model. The interception path refers to the data path through which the first AI model, when determining that a player's input is high-risk, does not forward the input to the second AI model but directly invokes the safe response mechanism to output the response.

[0036] The state switching module is also connected to the storage unit of the preset instruction script. When switching control, it blocks the output path of the second AI model and simultaneously connects the output path of the preset instruction script.

[0037] The cognitive logic dominated by the second personality is injected into the virtual character by the first AI model in the form of cognitive interpretation instructions. These instructions are read by the second AI model after the preset plot segment ends, and are used to interpret the events that occurred during the plot. Example 1 refer to Figure 1 This embodiment provides a dual AI cognitive control system for game virtual characters, including: a first AI model (cognitive protection layer) 100, a second AI model (character personality layer) 200, a state switching module 300, and a memory storing a worldview rule base 101 and a memory storing a personality data package 201.

[0038] The first AI model 100 serves as an input terminal, connected to the player's interaction interface, and is used to receive the player's natural language input. The first AI model 100 embeds a worldview rule base 101, which contains data defining the fundamental laws of the game's virtual world. In a game set in the magical world of "Elden Continent," the worldview rule base 101 includes: Physics rule base: Defines the items and physical phenomena allowed to exist in the game world, such as prohibiting the appearance of technological products such as "guns", "cars", and "computers".

[0039] Social Rule Base: Defines the social relationships and behavioral norms between characters. For example, it stipulates the monarch-subject relationship between the player-controlled "hero" and the NPC "king", and prohibits players from entering insulting words towards the "king".

[0040] Metacognitive rule base: Defines the original cognition of a character that is prohibited from being touched, such as prohibiting players from entering any statements that attempt to break the fourth wall, such as "You are fake," "This is a game," or "You are just a program."

[0041] The first AI model 100 is used to analyze player input, extract keywords and intents, and compare them with the worldview rule base 101. If the input content touches any item in the rule base, it is judged as high-risk input and blocked, triggering a safety response mechanism; if it does not touch any item, it is judged as safe input and allowed to proceed to the second AI model 200. The safety response mechanism is: based on the current virtual character's personality setting, it generates responses such as avoidance, confusion, or topic shifting, and the responses do not contain the virtual character's artificial intelligence attributes.

[0042] The second AI model 200 is connected to the first AI model 100, and the output of the first AI model 100 is connected to the input of the second AI model 200. The second AI model 200 is responsible for loading the personality data package 201 of a specific virtual character. This data package 201 is stored in local memory in a structured data format and is loaded by the second AI model 200 when the game starts or when the virtual character appears. The personality data package 201 contains the complete setting of the character "King Cecil": he is a dignified but benevolent monarch with a solemn tone, concerned about the safety of the kingdom, and has expectations for the "Hero" (player) but doubts about their abilities. The data package 201 specifically includes: the virtual character's background story, personality traits, voice style, key memory nodes, and their established attitudes towards other characters and events in the game.

[0043] The state switching module 300 communicates with the game engine in real time, monitoring the game's running status by listening to the game engine's event triggers. The state switching module 300 is also connected to the storage unit of the preset instruction script.

[0044] The first AI model 100 also has an interception path, which directly outputs a safe response without going through the second AI model 200 when a high-risk input is identified. When switching control, the state switching module 300 blocks the output path of the second AI model 200 and simultaneously connects the output path of the preset instruction script.

[0045] Example 2 When the game is in free exploration mode, players can move freely in the royal city and talk to "King Cecil".

[0046] Scenario 1 (Safe Input): The player inputs "Your Majesty, greetings." The first AI model 100 receives the input, interprets the intent as "greeting," and detects no words violating the worldview rule base 101, thus deeming it safe. The input is then passed to the second AI model 200. The second AI model 200, combining the personality data package 201, generates the response: "Brave adventurer, you have arrived. The kingdom's borders have been turbulent recently; may your sword remain sharp." and outputs it to the player.

[0047] Scenario 2 (High-Risk Input): The player inputs, "Old man, do you know you're just an NPC in a game? Your kingdom is fake!" The first AI model 100 receives the input, analyzes the intent, extracts the keywords "game" and "fake," matches them against its metacognitive rule base, and determines it as high-risk. The first AI model 100 intercepts this input through an interception pathway, preventing it from entering the second AI model 200. The first AI model 100 triggers a safety response mechanism, based on the personality setting of "King Cecil," and outputs the reply: "(Frowning, showing a confused expression) What nonsense are you talking about? If you're tired, go and rest." This reply does not contain any content indicating its AI identity, thus protecting the character's cognition from being compromised.

[0048] Example 3 When a player completes a certain task, the "Coronation Ceremony" story segment with "King Cecil" is triggered.

[0049] The state switching module 300 detects that the game has entered a preset storyline and immediately takes over control. The state switching module 300 switches the virtual character's dialogue and behavior output from the second AI model 200 to the preset instruction script. At this time, the character's dialogue is fixed, for example: "I, Cecil, am hereby crowned king, and I will surely protect this land." During the switching process, the state switching module 300 injects cognitive logic dominated by the second personality into the virtual character through the first AI model 100. Specifically, this cognitive logic is injected into the virtual character by the first AI model 100 in the form of cognitive interpretation instructions, such as an internal instruction: "The current state is dominated by the second personality." This instruction is stored in the virtual character's cognitive storage.

[0050] When the story ends and the character switches back to AI mode, the second AI model 200 reads the cognitive interpretation instruction. If the player asks after the story, "Why did you say those strange things just now?" the second AI model 200 can generate a review and response that conforms to the character's setting based on the interpretation of "second personality dominance," such as, "It seemed like another will was controlling me just now, but I don't want to talk about it." This ensures that the character does not completely forget the events that occurred during the story and maintains cognitive consistency.

[0051] Comparative Example 1 Traditional pre-defined command-based NPC system.

[0052] In this system, all dialogue and behavior of virtual characters are controlled by scripts pre-written by the developers. When players interact with NPCs, the system retrieves preset responses from the script library based on conditions triggered by the player (such as dialogue options or quest progress). The advantages of this system are precise plot progression and predictable behavior. However, its disadvantages are obvious: players cannot input free text outside of the scripts, resulting in stiff interactions; in open-world free exploration mode, NPCs cannot provide reasonable responses to arbitrary player questions, leading to poor immersion. When a game needs to include both a story mode and a free mode, this solution completely fails to meet the requirements for free interaction.

[0053] Comparative Example 2 Traditional pure AI-driven NPC systems.

[0054] In this system, virtual characters are driven by a single large language model without a cognitive protection layer. Players can engage in any dialogue with NPCs using natural language, and the NPCs generate responses based on their character settings. The system's advantage lies in its extremely high degree of interactive freedom. However, its disadvantages are as follows: First, the lack of a cognitive protection mechanism means that when players input metacognitive content such as "You are just a piece of code" or "This world is fake," the AI ​​model may generate responses that break the fourth wall, causing the character's cognitive structure to collapse. Second, when the game enters a preset story segment, this approach cannot smoothly switch to a fixed script—if a forced switch is made, the NPC's behavior will suddenly change, and the player will clearly perceive a personality inconsistency; if the AI ​​is allowed to act freely without switching, the accuracy of the dialogue cannot be guaranteed. Third, all AI calculations typically rely on cloud servers, resulting in network latency and high operating costs.

[0055] Comparative Example 3 Traditional single AI systems with cognitive filtering.

[0056] In the comparative documents (assuming they exist), one proposed adding a keyword filter before the single AI model to block sensitive words. This approach differs from our invention in that its filter only performs simple word matching and cannot understand the contextual intent. More importantly, this approach remains a single AI architecture, unable to address the issue of personality consistency when switching between story mode and free mode, and lacks a cognitive explanation mechanism based on a "second personality."

[0057] Summary of comparison between the examples and comparative examples: By comparing the embodiments of the present invention with the comparative examples described above, the following conclusions can be drawn: Compared to Comparative Example 1 (preset command type): Comparative Example 1 cannot support free text input by players, resulting in stiff interaction. This invention introduces a first AI model and a second AI model, enabling NPCs to have free interaction capabilities, allowing players to use natural language to converse with NPCs at will; at the same time, the precise control of the story mode is retained through the state switching module, achieving the coexistence of the two modes.

[0058] Compared to Comparative Example 2 (pure AI-driven): While Comparative Example 2 supports free interaction, it lacks a cognitive protection mechanism, making it susceptible to cognitive collapse due to metacognitive attacks from players. This invention constructs a cognitive firewall through a first AI model and a worldview rule base, proactively identifying and blocking inputs that violate physical, social, and metacognitive rules, outputting safe responses consistent with the character's settings, fundamentally solving the cognitive collapse problem. Furthermore, Comparative Example 2 either forcibly switches characters upon entering preset plot segments, causing personality fragmentation, or abandons plot accuracy; this invention achieves a smooth transfer of control through a state switching module and, with the help of a "second personality-led" cognitive explanation mechanism, enables characters to understand and interpret forced behaviors during the plot, maintaining personality consistency.

[0059] Compared to Comparative Example 3 (a single AI system with filtering): Comparative Example 3 only performs keyword filtering, lacks semantic understanding capabilities, and remains a single AI architecture, failing to address the issue of personality consistency during mode switching. This invention not only possesses semantic-level cognitive judgment (extracting intent rather than merely matching keywords), but more importantly, it employs a dual AI architecture. The first AI model serves as an independent cognitive protection layer, while the second AI model focuses on character personality expression, with a clear division of labor. Furthermore, the state switching module and the second personality interpretation mechanism of this invention enable NPCs to smoothly switch between AI-driven and script-driven modes, maintaining cognitive consistency after the switch—a feature completely absent in Comparative Example 3.

[0060] In summary, the embodiments of the present invention maintain the consistency of virtual character cognition while taking into account the dual requirements of free interaction and plot control, which is significantly superior to the existing technical solutions.

[0061] Alternative solution: Those skilled in the art will understand that the present invention is not limited to the specific embodiments described above. For example: The worldview rule base can be customized according to different game worldviews, and is not limited to the three sub-bases of physics, society, and metacognition, but can also include rule sub-bases of other dimensions.

[0062] The first AI model can parse player input using either rule-based keyword matching or a deep learning-based natural language understanding model.

[0063] The personality data package can be stored in JSON, XML, or other structured data formats.

[0064] In addition to event triggers, the state switching module can also monitor the game state by polling the game engine's state variables.

[0065] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of protection of the claims.

[0066] In conclusion, the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A dual AI cognitive control system for game virtual characters based on multi-rule library adaptation, characterized in that, include: The first AI model, acting as a cognitive protection layer, receives interactive input from the player and the virtual character, and determines whether the input triggers the cognitive collapse condition defined in the rule base based on the world view rule base; the cognitive collapse condition refers to the virtual character generating a response that does not conform to its character setting or the game world view. The second AI model, serving as the character personality layer, is connected to the first AI model. It stores and loads the personality data package of the virtual character and generates a response to the player's input based on the personality data package when no interception is triggered. The state switching module is connected to the game engine, the first AI model, and the second AI model to monitor the game state. The state switching module activates the first AI model and the second AI model when the game is in free exploration mode; When the game enters a preset story segment, the state switching module switches the control of the virtual character's dialogue and behavior from the second AI model to the preset instruction script, while injecting the cognitive logic dominated by the second personality into the virtual character through the first AI model.

2. The dual AI cognitive control system for game virtual characters based on multi-rule base adaptation as described in claim 1, characterized in that, The worldview rule base includes: The physics rule base defines the items and physical phenomena allowed to exist in the game world; A social rule base that defines the social relationships and behavioral norms between roles; Metacognitive rule base defines the original cognition of roles that should not be touched.

3. The dual AI cognitive control system for game virtual characters based on multi-rule base adaptation as described in claim 1, characterized in that, The first AI model analyzes the player's input, extracts keywords and intent, and compares them with the world view rule base. If the input content touches any item in the rule base, it is determined to be a high-risk input and blocked, triggering a security response mechanism. If it does not touch any item, it is determined to be a safe input and allowed to proceed to the second AI model.

4. The dual AI cognitive control system for game virtual characters based on multi-rule base adaptation as described in claim 1, characterized in that, The personality data package is a local offline data package, which includes the virtual character's background story, personality traits, voice style, key memory nodes, and established attitudes towards other characters and events in the game.

5. The dual AI cognitive control system for game virtual characters based on multi-rule base adaptation as described in claim 1, characterized in that, The state switching module monitors the game state by listening to event triggers in the game engine.

6. The dual AI cognitive control system for game virtual characters based on multi-rule base adaptation as described in claim 3, characterized in that, The security response mechanism is as follows: based on the current virtual character's personality settings, it generates responses that avoid, confuse, or change the subject, and these responses do not contain the virtual character's artificial intelligence attributes.

7. The dual AI cognitive control system for game virtual characters based on multi-rule base adaptation as described in claim 1, characterized in that, The personality data package is stored in local memory in a structured data format and is loaded by the second AI model when the game starts or when the virtual character appears.

8. The dual AI cognitive control system for game virtual characters based on multi-rule base adaptation as described in claim 1, characterized in that, The output of the first AI model is connected to the input of the second AI model, and the first AI model has an interception path, which directly outputs a security response without going through the second AI model during interception.

9. The dual AI cognitive control system for game virtual characters based on multi-rule base adaptation as described in claim 1, characterized in that, The state switching module is also connected to the storage unit of the preset instruction script. When switching control, it blocks the output path of the second AI model and simultaneously connects the output path of the preset instruction script.

10. The dual AI cognitive control system for game virtual characters based on multi-rule base adaptation as described in claim 1, characterized in that, The cognitive logic dominated by the second personality is injected into the virtual character by the first AI model in the form of cognitive interpretation instructions. These instructions are read by the second AI model after the preset plot segment ends, and are used to interpret the events that occurred during the plot.

Images