Original script-oriented AI autonomous plot structure adaptive generation system

Abstract

The invention discloses an original script-oriented AI autonomous plot structure adaptive generation system, and relates to the technical field of creation assistance, and the system specifically comprises the following modules: a structure extraction analysis module, an emotional role analysis module, a plot inference module, a scene generation optimization module, a conservation target generation module, an adaptive control module, and a constraint punishment module. According to the method, a multi-level narrative structure and a causal relationship graph are constructed to form an emotion vector and a trajectory curve, an optimal causal path is generated based on a graph neural network, a graph convolution / attention mechanism and a graph generation algorithm, and emotion toning and conservation target driven text generation are performed on scene and dialogue levels. Structural consistency and emotional arcs are optimized in real time in combination with self-adaptive control, plot path weighting and rewriting triggering are performed by utilizing a multi-dimensional emotional space and a neural oscillator network, intelligent structured management of a script is realized, and script creation efficiency and quality are improved.

Description

Technical Field

[0001] This invention relates to the field of creative assistance technology, specifically to an AI-driven autonomous plot structure adaptive generation system for original scripts. Background Technology

[0002] With the rapid development of artificial intelligence technology, AI is being applied more and more widely in content creation, especially in screenwriting. Traditional screenwriting methods mainly rely on the author's inspiration and experience. While this approach can produce high-quality works, it suffers from problems such as low efficiency and high time and effort. At the same time, existing AI generation methods still have limitations in handling complex plots and multi-layered narrative structures, making it difficult to accurately grasp the narrative logic and emotional expression of a screenplay.

[0003] Currently, AI script generation systems based on large language models can generate coherent and logical text content based on given prompts, but they still face challenges in generating scripts of stable quality. In particular, they perform poorly in handling scene transitions and visual coherence, struggling to achieve intelligent segmentation and transition effects based on content semantics. Furthermore, existing technologies often lack the flexibility for user interaction when generating longer scripts, failing to fully incorporate the creator's skills and insights.

[0004] The existing technology has the following shortcomings:

[0005] 1. Traditional screenwriting methods mainly rely on the author's inspiration and experience. Although they can produce high-quality works, they suffer from problems such as low creative efficiency and being time-consuming and laborious.

[0006] 2. Existing AI generation methods have limitations when dealing with complex plots and multi-layered narrative structures, making it difficult to accurately grasp the narrative logic and emotional expression of the script, thus affecting the quality and consistency of the generated results;

[0007] 3. AI script generation systems based on large language models still face challenges in generating scripts of stable quality, especially in handling scene transitions and visual coherence, and are difficult to achieve intelligent segmentation and transition effect generation based on content semantics.

[0008] 4. Existing technologies often lack the flexibility to interact with users when generating longer scripts, and cannot fully incorporate the creative skills and insights of human creators, thus limiting the intelligence level of the generation system and the user experience.

[0009] 5. Existing script generation systems lack effective graphical representation and reasoning mechanisms when understanding and processing multi-layered narrative structures, making it difficult to accurately analyze and predict the development trend and key nodes of complex plots;

[0010] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0011] The purpose of this invention is to provide an AI-driven autonomous plot structure adaptive generation system for original scripts, in order to solve the problems mentioned in the background art.

[0012] To achieve the above objectives, the present invention provides the following technical solution: an AI-driven autonomous plot structure adaptive generation system for original scripts, specifically including the following modules: a structure extraction and analysis module, an emotional character analysis module, a plot inference module, a scene generation optimization module, a conservation goal generation module, an adaptive control module, and a constraint penalty module;

[0013] Structure Extraction and Analysis Module: By performing semantic analysis on the script text, a multi-layered narrative structure is constructed, and a causal relationship diagram between events is established;

[0014] Emotional Role Analysis Module: Constructs a multi-dimensional emotional feature space, extracts emotional vectors, and calculates emotional trajectory curves;

[0015] Plot inference module: It uses graph neural networks to learn features of nodes and edges, and adjusts edge weights through controllable variables to generate the optimal causal relationship path;

[0016] Scene generation optimization module: supplements and optimizes scenes and plots through a multi-dimensional emotional feature space and a preset narrative rule library;

[0017] Conservation Goal Generation Module: Assigns dialogue text based on role weights and generates the final script dialogue draft;

[0018] Adaptive control module: Through an adaptive controller combined with a dynamic adjustment mechanism for controllable variables, the script dialogue draft is refined and optimized.

[0019] Constraint and punishment module: By constructing a multi-dimensional emotional feature space and applying a neural oscillator network, the emotional changes in the script are optimized and controlled.

[0020] As a preferred embodiment of the AI-driven autonomous plot structure adaptive generation system for original scripts described in this invention, wherein:

[0021] Preprocess the given script text to construct a lexical unit tree;

[0022] A dependency parser is used to perform semantic analysis on the lexical unit tree to extract semantic units, specifically including: subject, verb, object, and complement.

[0023] Create a causal relationship diagram between events in the script;

[0024] The preprocessing specifically includes word segmentation and part-of-speech tagging of the script text;

[0025] Based on lexical unit trees, a multi-level narrative structure containing script events, characters, script scenes, and relationships is constructed, and a graph generation algorithm is used to generate a causal relationship graph between events.

[0026] As a preferred embodiment of the AI-driven autonomous plot structure adaptive generation system for original scripts described in this invention, wherein:

[0027] Extract the characters' motivations, conflicts, and relationships from the script to construct a multi-dimensional emotional character space;

[0028] Emotional features are extracted using a recurrent neural network to obtain the emotional vector of the time series.

[0029] The emotional trajectory is fitted using linear regression to calculate the emotional trajectory curve;

[0030] Define multiple emotional dimensions for each character, specifically including: positive emotions, sadness, and anger;

[0031] Each emotional dimension is mapped onto a multidimensional emotional feature space.

[0032] As a preferred embodiment of the AI-driven autonomous plot structure adaptive generation system for original scripts described in this invention, wherein:

[0033] Each plot point is treated as a node in a causal graph using a graph neural network, and the causal relationships between events in the script are treated as edges in the graph.

[0034] Graph convolutional networks are used to extract features from nodes and edges to obtain high-level representations of the plot;

[0035] Causal relationships are weighted using graph attention networks;

[0036] The optimal causal path is generated based on edge weights using a graph generation algorithm.

[0037] As a preferred embodiment of the AI-driven autonomous plot structure adaptive generation system for original scripts described in this invention, wherein:

[0038] Automatically generate scene description templates based on the location, time, and resource constraints required by the plot;

[0039] By applying emotional color correction to scenes using a multi-dimensional emotional feature space, scenes whose emotions match the needs of the plot can be obtained.

[0040] Based on a pre-defined narrative rule library, the scenes and plots are supplemented and expanded.

[0041] As a preferred embodiment of the AI-driven autonomous plot structure adaptive generation system for original scripts described in this invention, wherein:

[0042] Each role is defined with its own goals and obstacles through a conservation-driven mechanism, and dialogue text is generated using a recurrent neural network.

[0043] Each line is weighted according to its role to obtain the final draft of the script dialogue.

[0044] As a preferred embodiment of the AI-driven autonomous plot structure adaptive generation system for original scripts described in this invention, wherein:

[0045] The adaptive controller monitors the structural consistency and emotional arc of the script dialogue draft in real time.

[0046] Adjust the speed of information presentation according to audience preferences;

[0047] The output of each module is adjusted by weighting controllable variables, specifically: the emotional intensity variable adjusts the output of the emotional module, and the suspense variable controls the plot turning points.

[0048] As a preferred embodiment of the AI-driven autonomous plot structure adaptive generation system for original scripts described in this invention, wherein:

[0049] The possibilities for plot development are represented as multiple points in a multidimensional emotional space;

[0050] By using a neural oscillator network to perform a weighted summation of the points, the final plot development path can be obtained;

[0051] The plot development path is optimized by updating the weights of the neural oscillator using the backpropagation algorithm.

[0052] If a plot point violates causality or a character's motivation is invalid, a rewrite or replacement will be triggered.

[0053] On the other hand, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, it implements the steps of the AI ​​autonomous plot structure adaptive generation system for original scripts as described above.

[0054] On the other hand, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements the steps of the AI ​​autonomous plot structure adaptive generation system for original scripts as described above.

[0055] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0056] 1. By automatically extracting and maintaining multi-layered narrative structures, intelligent and structured management of scripts is achieved, improving the efficiency and quality of script creation and solving the problem of low efficiency caused by the reliance on human inspiration and experience in traditional script creation;

[0057] 2. Based on character motivation, goal conflict, and causal relationship of events, the logic of the progression is generated, overcoming the limitations of existing AI generation methods in handling complex plots and multi-layered narrative structures, and ensuring the consistency and logic of the generated results;

[0058] 3. By inferring the causal graph of the plot through graph neural network, we can realize the visualization and intelligent analysis of multi-level narrative structure, improve the system's ability to understand and predict complex plots, and solve the problem of the lack of effective graph representation and reasoning mechanism in the existing technology;

[0059] 4. By controlling the generated results through controllable variables, the script plot can be finely adjusted, the diversity and flexibility of the generated results can be improved, and the problem of lack of user interaction flexibility when generating long scripts in the existing technology can be overcome.

[0060] 5. Based on the location, time, and resource constraints required by the scenario, the system automatically generates scene description templates and fills in specific details to achieve intelligent scene generation, thus addressing the shortcomings of existing technologies in handling scene transitions and visual coherence. Attached Figure Description

[0061] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0062] Figure 1 This is a flowchart of the AI-driven autonomous plot structure adaptive generation system for original scripts, as described in this invention.

[0063] Figure 2 This is a schematic diagram of the modules of the AI-driven autonomous plot structure adaptive generation system for original scripts according to the present invention. Detailed Implementation

[0064] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0065] Example 1, referring to Figure 1 and Figure 2 This is the first embodiment of the present invention. This embodiment provides an AI autonomous plot structure adaptive generation system for original scripts, which specifically includes the following modules: structure extraction and analysis module, emotional character analysis module, plot inference module, scene generation optimization module, conservation goal generation module, adaptive control module, and constraint penalty module.

[0066] Structure Extraction and Analysis Module: By performing semantic analysis on the script text, a multi-layered narrative structure is constructed, and a causal relationship diagram between events is established;

[0067] Preprocess the given script text to construct a lexical unit tree;

[0068] A dependency parser is used to perform semantic analysis on the lexical unit tree to extract semantic units, specifically including: subject, verb, object, and complement.

[0069] Create a causal relationship diagram between events in the script;

[0070] The preprocessing specifically includes word segmentation and part-of-speech tagging of the script text;

[0071] Based on lexical unit trees, a multi-level narrative structure containing script events, characters, script scenes, and relationships is constructed, and a graph generation algorithm is used to generate a causal relationship graph between events.

[0072] It should be further noted that the lexical unit tree is a tree structure with "word" as the basic unit. Its nodes contain words, parts of speech, and semantics. Its hierarchical structure is represented in this invention as follows:

[0073] (1) Root node: The entire script;

[0074] (2) First layer: Scene;

[0075] (3) Second layer: Sentence;

[0076] (4) Third layer: Lexical unit, words in the same sentence form a linear chain in order, and subtrees are formed for parallel structures;

[0077] It should be further noted that the semantic analysis extracts semantic units from the sentence-level output of the lexical unit tree and obtains the dependency relation book through the dependency parser. The dependency relation mapping specifically includes: subject-predicate relation, predicate / core verb, direct object of verb, and complement.

[0078] Combine tuples, triples, or quadruples using grammatical rules;

[0079] The grammar rules are as follows in this application:

[0080] If the dependency tree contains a subject pointed to by a noun and a verb pointed to by a root word, then record the tuple (subject and core verb).

[0081] If there is an object that a direct object points to, then it expands into a triple (subject, core verb, and direct object).

[0082] If a complement exists and its related object exists, it is expanded into a quadruple (subject, core verb, direct object, and complement).

[0083] It should be further explained that the semantic unit of the triple is used as the basic event unit, the events in the dialogue are treated as independent nodes, and the edges are used as the interaction causality or sequential logic in the dialogue. The causal relationship inference rules are defined based on explicit causal signal words, temporal relationships, goal inferences and relationship clues between roles. The directed acyclic graph is used as the initial model to represent the causal relationship graph. If there is a cyclic scene, the cycle is retained but restricted during the analysis.

[0084] Emotional Role Analysis Module: Constructs a multi-dimensional emotional feature space, extracts emotional vectors, and calculates emotional trajectory curves;

[0085] Extract the characters' motivations, conflicts, and relationships from the script to construct a multi-dimensional emotional character space;

[0086] Emotional features are extracted using a recurrent neural network to obtain the emotional vector of the time series.

[0087] The emotional trajectory is fitted using linear regression to calculate the emotional trajectory curve;

[0088] Define multiple emotional dimensions for each character, specifically including: positive emotions, sadness, and anger;

[0089] Map each emotional dimension to a multidimensional emotional feature space;

[0090] It should also be noted that the multidimensional emotional feature space is defined as positive emotions, sadness, anger, and optional extended dimensions;

[0091] The emotional trajectory curve is obtained by generating a vector sequence over time for each character in each emotional dimension. The time series of each dimension is fitted by linear regression to obtain a smooth trajectory curve. For a certain dimension, nonlinear trends are captured by multinomial regression.

[0092] Plot inference module: It uses graph neural networks to learn features of nodes and edges, and adjusts edge weights through controllable variables to generate the optimal causal relationship path;

[0093] Each plot point is treated as a node in a causal graph using a graph neural network, and the causal relationships between events in the script are treated as edges in the graph.

[0094] Graph convolutional networks are used to extract features from nodes and edges to obtain high-level representations of the plot;

[0095] Causal relationships are weighted using graph attention networks;

[0096] The optimal causal path is generated based on edge weights using a graph generation algorithm.

[0097] Furthermore, it should be noted that the initial node features of each event are obtained through textual information, the node representation is updated through multi-layer graph convolution, a high-level semantic vector of the node is generated, the weights of the edges are adaptively learned to obtain the edge-level representation, and the controllable variables are mapped to the edge weights through conditional gating; based on the edge weights and node representations, the path is used as a sequence input to the downstream task through a graph generation algorithm to find the optimal causal path.

[0098] Scene generation optimization module: supplements and optimizes scenes and plots through a multi-dimensional emotional feature space and a preset narrative rule library;

[0099] Automatically generate scene description templates based on the location, time, and resource constraints required by the plot;

[0100] By applying emotional color correction to scenes using a multi-dimensional emotional feature space, scenes whose emotions match the needs of the plot can be obtained.

[0101] Based on a pre-defined narrative rule library, the scenes and plots are supplemented and optimized.

[0102] Conservation Goal Generation Module: Assigns dialogue text based on role weights and generates the final script dialogue draft;

[0103] Each role is defined with its own goals and obstacles through a conservation-driven mechanism, and dialogue text is generated using a recurrent neural network.

[0104] Each line is weighted according to its role to obtain the final draft of the script dialogue;

[0105] It should be further explained that the input consists of dialogue, character goals / obstacles, and current situational state, and the output is turn-by-turn dialogue text. During generation, a conserved goal vector is introduced as an additional attention key value to calculate the contribution of each sentence of text. The final draft is obtained by combining the character weight vector.

[0106] Adaptive control module: Through an adaptive controller combined with a dynamic adjustment mechanism for controllable variables, the script dialogue draft is refined and optimized.

[0107] The adaptive controller monitors the structural consistency and emotional arc of the script dialogue draft in real time.

[0108] Adjust the speed of information presentation according to audience preferences;

[0109] The output of each module is adjusted by weighting controllable variables, specifically: the emotional intensity variable adjusts the output of the emotional module, and the suspense variable controls the plot turning points.

[0110] Constraint and punishment module: By constructing a multi-dimensional emotional feature space and applying a neural oscillator network, the emotional changes in the script are optimized and controlled;

[0111] The possibilities for plot development are represented as multiple points in a multidimensional emotional space;

[0112] By using a neural oscillator network to perform a weighted summation of the points, the final plot development path can be obtained;

[0113] The plot development path is optimized by updating the weights of the neural oscillator using the backpropagation algorithm.

[0114] If a plot point violates causality or a character's motivation is invalid, a rewrite or replacement will be triggered.

[0115] Example 2

[0116] The following is another embodiment of the present invention, which provides an AI autonomous plot structure adaptive generation system for original scripts. In order to verify the beneficial effects of the present invention, a simulation experiment is conducted for scientific demonstration.

[0117] The purpose of the experiment is:

[0118] (1) Verify the effectiveness and synergistic effect of the structure extraction and analysis module, emotional role analysis module, plot inference module, scene generation optimization module, conservation goal generation module, adaptive control module, and constraint penalty module in original script generation;

[0119] (2) Evaluate the extent to which the generated script improves causal consistency, emotional coherence, scene richness, and audience readability;

[0120] (3) Compare the impact of adaptive control and constraint penalty mechanisms on the quality of the final draft.

[0121] The data source is an original script text corpus, specifically including: three original script texts (dramatic texts or full-length scripts), each approximately 20,000-40,000 words, ensuring a structure containing multiple scenes and rich dialogue;

[0122] The data sources for which permission is granted include: public domain scripts, practice texts from collaborating theater companies, and licensed texts for research purposes.

[0123] Auxiliary annotation data specifically includes:

[0124] Dependency tagging: High-quality subsets are obtained by using English or Chinese dependency tagging provided by SpaCy / Stanza, combined with manual verification;

[0125] Scene and character annotation: Manually annotate some paragraphs to clarify the relationship between "scene number, character, dialogue paragraph, and event sentence";

[0126] Emotional Tags: Assign emotional tags (positive, sad, angry, etc.) to character dialogue and give an emotional intensity score.

[0127] Refer to the sequence of events, causal clues, and character motivations in the script as a benchmark for evaluating causal paths;

[0128] The questionnaire collected audience ratings on the emotional curve, plot tension, and scene richness.

[0129] Separate and compare the individual modules:

[0130] Group A: A complete AI-driven autonomous plot structure adaptive generation system for original scripts (all 7 modules are involved in operation);

[0131] Group B: Versions that have removed the emotional role analysis module;

[0132] Group C: Versions that remove the adaptive control module;

[0133] Group D: Version of the module that removes constraint penalties.

[0134] Preprocessing: The script text is segmented into words, tagged with parts of speech, and a lexical unit tree is constructed;

[0135] Structural extraction analysis: Constructing two / three / four tuples based on dependency analysis and generating an initial causal relationship graph;

[0136] Emotional role analysis: Extracting time-series emotional vectors through recurrent neural networks, and performing multi-dimensional emotional mapping and trajectory fitting;

[0137] Plot inference: Graph neural networks are used to learn node / edge features, adaptively adjust edge weights, and output the optimal causal path;

[0138] Scene generation optimization: Generate scene description templates based on multi-dimensional emotional space and narrative rule base, and adjust emotional color tone;

[0139] Conservative Goal Generation: Taking dialogue input, goal / obstacle, and current context state as input, outputs a round-by-round dialogue and attaches a conservation goal vector as an attention key;

[0140] Adaptive control: dynamically adjusts controllable variables, monitors structural consistency and emotional curve in real time, and adjusts the output weights of modules;

[0141] Constraint and punishment: The emotional points are weighted and summed through a neural oscillator network, and the weights are updated. If the causality or motivation is violated, a rewrite is triggered.

[0142] Evaluation process:

[0143] Automatic evaluation: quantitative scoring of causal path correctness, emotional consistency index, scene diversity, dialogue coherence, etc.

[0144] Human evaluation: Two groups of viewers compared their scores, including emotional consistency, clarity of character motivations, script readability, and scene credibility.

[0145] Compare the differences in output quality between groups A, B, C, and D on the same original text.

[0146] The experimental data are as follows:

[0147] Original screenplay texts: 3 parts, each approximately 30,000 words;

[0148] A subset of annotations includes dependency annotations and scene / role annotations for 2,000 sentences.

[0149] Audience review: 200 audience members, spanning multiple age groups;

[0150] The input segment is as follows:

[0151] Scene 1, Character A's dialogue: "I must solve this mystery tonight, or it will be too late before dawn."

[0152] Correspondence Dependency Analysis:

[0153] Core verb: to solve;

[0154] Direct object: This mystery;

[0155] Addendum: Otherwise, it will be too late before dawn.

[0156] The expected output is as follows:

[0157] Event Triad: <Subject: A, Core Verb: Solve, Object: This Mystery>;

[0158] The sequence of emotion vectors over time is: [(positive 0.2, sadness -0.1, anger 0.5), (positive 0.1, sadness 0.0, anger 0.6)];

[0159] Cause-and-effect path: A → Solving the mystery → Change in character B's motivation;

[0160] Scene template: Night scene, dim lighting, stacked clues;

[0161] Draft Dialogue: Adjust the pace of revealing potential information according to the conserved objective vector.

[0162] The results and analysis are as follows:

[0163] The complete system group A showed significant improvements over baseline groups B / C / D in the following aspects:

[0164] Causal path coverage increased by 18%–28%;

[0165] The average RMSE of emotional trajectory decreased by 12% to 22%;

[0166] The diversity of scene descriptions has increased by 25% (due to the usage rate of different templates);

[0167] In human assessments, the score for emotional consistency improved by 0.6 to 0.8 points (out of 5).

[0168] By automatically extracting and maintaining multi-layered narrative structures, intelligent and structured management of scripts is achieved, improving the efficiency and quality of script creation and solving the problem of low efficiency caused by the reliance on human inspiration and experience in traditional script creation.

[0169] Based on character motivation, goal conflict, and event causal relationship, the logic of the generation is generated to overcome the limitations of existing AI generation methods in handling complex plots and multi-layered narrative structures, and to ensure the consistency and logic of the generated results.

[0170] By using graph neural networks to infer the causal graph of the plot, we can achieve a visual representation and intelligent analysis of multi-level narrative structures, improve the system's ability to understand and predict complex plots, and solve the problem of the lack of effective graph representation and reasoning mechanisms in existing technologies.

[0171] By controlling the generated results with controllable variables, the script plot can be finely adjusted, the diversity and flexibility of the generated results can be improved, and the problem of lack of user interaction flexibility when generating long scripts in the existing technology can be overcome.

[0172] Based on the location, time, and resource constraints required by the scenario, the system automatically generates scene description templates and fills in specific details, achieving intelligent scene generation and solving the shortcomings of existing technologies in handling scene transitions and visual coherence.

[0173] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. An AI autonomous plot structure self-adaptive generation system for original scripts, characterized in that, Specifically include the following modules: structure extraction analysis module, sentiment role analysis module, plot inference module, scene generation optimization module, conservation target generation module, adaptive control module, constraint penalty module; The structure extraction analysis module: through semantic analysis of the script text, a multi-level narrative structure is constructed, and a causal relationship graph between events is established; The sentiment role analysis module: a multi-dimensional sentiment feature space is constructed, sentiment vectors are extracted, and a sentiment trajectory curve is calculated; The plot inference module: graph neural networks are used to learn the features of nodes and edges, and the edge weights are adjusted through controllable variables to generate the optimal causal relationship path; The scene generation optimization module: through the multi-dimensional sentiment feature space and the preset narrative rule library, the scene and the plot are supplemented and optimized; The conservation target generation module: according to the role weight allocation of the dialogue text, the final script dialogue draft is generated; The adaptive control module: through the adaptive controller combined with the dynamic adjustment mechanism of the controllable variable, the fine optimization of the script dialogue draft is realized; The constraint penalty module: through the construction of a multi-dimensional sentiment feature space and the application of a neural oscillator network, the script emotion change is optimized and controlled.

2. The AI autonomous plot structure self-adaptive generation system for original script orientation according to claim 1, characterized in that: The structure extraction analysis module specifically includes: Pretreatment of the given script text, construction of a lexical unit tree; Using a dependency syntax analyzer to perform semantic analysis on the lexical unit tree, extracting semantic units, including: subject, verb, object, and complement; Establishing a causal relationship graph between script events; Based on the lexical unit tree, a multi-level narrative structure containing script events, characters, script scenes, and relationships is constructed, and a graph generation algorithm is used to generate a causal relationship graph between events.

3. The AI autonomous plot structure self-adaptive generation system for original script orientation according to claim 1, characterized in that: The sentiment role analysis module specifically includes: Extracting the character motivation, target conflict, and relationship of the script, constructing a multi-dimensional sentiment feature space; Extracting sentiment features through a recurrent neural network to obtain a time series of sentiment vectors; Fitting the sentiment trajectory through linear regression to calculate the sentiment trajectory curve; Defining multiple sentiment dimensions for each character; Mapping each sentiment dimension to the multi-dimensional sentiment feature space.

4. The AI autonomous plot structure self-adaptive generation system for original script orientation according to claim 1, characterized in that: The plot inference module specifically includes: Through a graph neural network, each plot is treated as a node in the causal relationship graph, and the causal relationship between script events is treated as an edge in the graph; Using a graph convolution network to extract features of nodes and edges to obtain a high-level representation of the plot; Weighting the causal relationship through a graph attention network; Generating the optimal causal relationship path according to the edge weight through a graph generation algorithm.

5. The AI autonomous plot structure self-adaptive generation system for original script orientation according to claim 1, characterized in that: The scene generation optimization module specifically includes: Automatically generating a scene description template according to the required location, time, and resource constraints of the plot; Through the multi-dimensional sentiment feature space, the scene is sentimentally retouched to obtain a scene that meets the emotional needs of the plot; According to the preset narrative rule library, the scene and the plot are supplemented and expanded.

6. The AI autonomous plot structure self-adaptive generation system for original script orientation according to claim 1, characterized in that: The adaptive control module specifically includes: Real-time monitoring of the structural consistency and emotional arc of the script dialogue draft through an adaptive controller; Adjusting the information revealing speed according to the audience preference; The output of each module is weighted and adjusted by controllable variables, including: the emotional intensity variable adjusts the output of the emotion module, and the suspense degree variable controls the plot turning point.

7. The AI autonomous plot structure self-adaptive generation system for original script orientation according to claim 1, characterized in that: The constraint penalty module specifically includes: The possibility of plot development is represented as a plurality of points in a multi-dimensional emotional space; The points are weighted and summed using a neural oscillator network to obtain a final plot development path; The weights of the neural oscillators are updated using a backpropagation algorithm to optimize the plot development path; If a plot violates causality or character motivation, rewriting or replacement is triggered. 8.The AI autonomous plot structure adaptive generation system for original script orientation of claim 1, wherein 。 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is characterized in that: The processor executes the computer program to realize the modules of the original script-oriented AI autonomous plot structure adaptive generation system in any one of claims 1-8.

10. A computer readable storage medium having stored thereon a computer program, characterized in that: The computer program is executed by the processor to realize the modules of the original script-oriented AI autonomous plot structure adaptive generation system in any one of claims 1-8.

Images