Situation simulation method and device based on large language model

Abstract

The invention discloses a situation simulation method and device based on a large language model. The method comprises the following steps: constructing a situation simulation information base; the simulation information base comprises entity names, entity attributes, entity positions and inter-entity communication relations; training a large language model by utilizing the situation simulation information base to obtain an optimized large language model; acquiring situation simulation demand information of a user, and processing the situation simulation demand information of the user by utilizing the optimized large language model to obtain a situation simulation result; and rendering the situation simulation result to obtain a situation display result. The method has important theoretical significance and application value for a presentation mode and an interaction design party of situation information in situation simulation.

Description

Technical Field

[0001] The present invention relates to the technical field of situation simulation, and in particular to a situation simulation method and device based on a large language model. Background Art

[0002] In the information-based environment, with the increase in dimensions and the acceleration of the pace, clearly and accurately grasping the situation and timely feedback of the simulated situation information are the key factors for victory. The situation simulation system is an important platform for commanders to conduct drills. With the development of geographic information systems and graphics engine technologies, certain progress has been made in the technical implementation of two- and three-dimensional situation simulation systems, but a reasonable design system has not yet been formed in terms of the presentation method and interaction design of situation information. Therefore, it is of great significance to study the display technology of situation simulation systems from the perspective of human-computer interaction. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a situation simulation method and device based on a large language model, which constructs a situation simulation information library, processes the situation simulation information library by using the large language model to obtain a situation simulation result, and renders the situation simulation result to obtain a situation display result. The method of the present invention can realize a reasonable presentation method and interaction design of situation information, and has important theoretical significance and application value for the presentation method and interaction design of situation information in situation simulation.

[0004] To solve the above technical problem, in the first aspect of the embodiments of the present invention, a situation simulation method based on a large language model is disclosed, and the method includes:

[0005] S1, constructing a situation simulation information library; the simulation information library includes entity names, entity attributes, entity positions, and communication relationships between entities;

[0006] S2, using the situation simulation information library to train a large language model to obtain an optimized large language model;

[0007] S3, obtaining the situation simulation requirement information of the user, and using the optimized large language model to process the situation simulation requirement information of the user to obtain a situation simulation result;

[0008] S4, rendering the situation simulation result to obtain a situation display result.

[0009] As an optional implementation manner, in the first aspect of the embodiments of the present invention, the using the situation simulation information library to train a large language model to obtain an optimized large language model includes:

[0010] S21, processing the situation simulation information library to obtain situation simulation data information;

[0011] The situation simulation data information includes the graph structure set G of entities, the entity screen coordinate information, and the entity movement trajectory information;

[0012] S22. Use the situation simulation data information to train the large language model to obtain an optimized large language model.

[0013] As an optional implementation manner, in the first aspect of the embodiments of the present invention, the processing of the situation simulation information library to obtain situation simulation data information includes:

[0014] S211. Process the situation simulation information library to obtain the graph structure set G of entities;

[0015] The expression of the graph structure set G of the entities is:

[0016] G = {V, E, Σ, L}

[0017] Among them, V is a non-empty set of vertices, and each vertex represents an entity. is the set of edges, Σ is the set of vertex labels, and L is the label assignment function. For all have L(v) representing the label of vertex v, where L(v) ∈ Σ;

[0018] S212. Process the graph structure set G to obtain the entity screen coordinate information;

[0019] S213. Process the communication relationship between entities to obtain the entity movement trajectory information.

[0020] As an optional implementation manner, in the first aspect of the embodiments of the present invention, the processing of the communication relationship between entities to obtain the entity movement trajectory information includes:

[0021] S2131. Process the communication relationship between entities to obtain the environmental information, initial position information, and target position information of the entities;

[0022] S2132. Process the environmental information, initial position information, and target position information of the entities to obtain the network training parameter information;

[0023] The network training parameter information includes the action a to be taken currently, the next state s obtained by executing the action a , ,

[0024] ,

[0023] , t+1 ,

[0025] , the immediate reward r, and the end signal;

[0024] S2133. Use the network training parameter information to train a preset neural network model to obtain an optimized neural network model;

[0025] S2134. Use the optimized neural network model to process the communication relationship between entities to be processed, and obtain entity movement trajectory information.

[0026] As an optional implementation manner, in the first aspect of the embodiments of the present invention, the training of the large language model using the situation simulation data information to obtain an optimized large language model includes:

[0027] S221. Divide the situation simulation data information to obtain a labeled sample set and an unlabeled sample set;

[0028] S222. Use the labeled sample set to train the large language model to obtain a first large language model;

[0029] S223. Use the first large language model to process the unlabeled sample set to obtain a pseudo-labeled sample set;

[0030] S224. Integrate the labeled sample set and the pseudo-labeled sample set to obtain a training sample set;

[0031] S225. Use the training sample set to train the large language model to obtain an optimized large language model.

[0032] As an optional implementation manner, in the first aspect of the embodiments of the present invention, the processing of the unlabeled sample set using the first large language model to obtain a pseudo-labeled sample set includes:

[0033] S2231. Extract the unlabeled sample set to obtain a first unlabeled sample set;

[0034] S2232. Perform confidence processing on the first unlabeled sample set to obtain a second unlabeled sample set;

[0035] S2233. Perform difference test on the second unlabeled sample set to obtain a pseudo-labeled sample set.

[0036] As an optional implementation manner, in the first aspect of the embodiments of the present invention, the rendering of the situation simulation result to obtain a situation display result includes:

[0037] S41. Perform visualization processing on the situation simulation result to obtain a situation simulation diagram;

[0038] S42. Denoise the situation simulation diagram to obtain a denoised situation simulation diagram;

[0039] S43. Perform visual enhancement on the denoised situation simulation diagram to obtain a situation display result.

[0040] The second aspect of the embodiments of the present invention discloses a situation simulation device based on a large language model, and the device includes:

[0041] An information library construction module, configured to construct a situation simulation information library; the simulation information library includes entity names, entity attributes, entity locations, and communication relationships between entities;

[0042] A model training module, configured to use the situation simulation information library to train a large language model to obtain an optimized large language model;

[0043] A situation simulation module, configured to obtain situation simulation requirement information of a user, and use the optimized large language model to process the situation simulation requirement information of the user to obtain a situation simulation result;

[0044] A situation display module, configured to render the situation simulation result to obtain a situation display result.

[0045] As an optional implementation manner, in the second aspect of the embodiments of the present invention, the using the situation simulation information library to train a large language model to obtain an optimized large language model includes:

[0046] S21, processing the situation simulation information library to obtain situation simulation data information;

[0047] The situation simulation data information includes a graph structure set G of entities, entity screen coordinate information, and entity movement trajectory information;

[0048] S22, using the situation simulation data information to train a large language model to obtain an optimized large language model.

[0049] As an optional implementation manner, in the second aspect of the embodiments of the present invention, the processing the situation simulation information library to obtain situation simulation data information includes:

[0050] S211, processing the situation simulation information library to obtain a graph structure set G of entities;

[0051] The expression of the graph structure set G of the entities is:

[0052] G = {V, E, Σ, L}

[0053] Wherein, V is a non-empty set of vertices, and each vertex represents an entity, is a set of edges, Σ is a set of vertex labels, and L is a label assignment function. For all have L(v) representing the label of vertex v, where L(v) ∈ Σ;

[0054] S212. Process the set of graph structures G to obtain entity screen coordinate information;

[0055] S213. Process the communication relationships between entities to obtain entity movement trajectory information.

[0056] As an optional implementation, in the second aspect of the embodiments of the present invention, the processing of the communication relationships between entities to obtain entity movement trajectory information includes:

[0057] S2131. Process the communication relationships between entities to obtain the environmental information, initial position information, and target position information of the entities;

[0058] S2132. Process the environmental information, initial position information, and target position information of the entities to obtain network training parameter information;

[0059] The network training parameter information includes the action a to be taken currently, the next state s obtained by executing the action a t+1 , the immediate reward r, and the end signal;

[0060] S2133. Use the network training parameter information to train a preset neural network model to obtain an optimized neural network model;

[0061] S2134. Use the optimized neural network model to process the communication relationships between entities to be processed to obtain entity movement trajectory information.

[0062] As an optional implementation, in the second aspect of the embodiments of the present invention, the training of the large language model using the situation simulation data information to obtain an optimized large language model includes:

[0063] S221. Divide the situation simulation data information to obtain a labeled sample set and an unlabeled sample set;

[0064] S222. Use the labeled sample set to train the large language model to obtain a first large language model;

[0065] S223. Use the first large language model to process the unlabeled sample set to obtain a pseudo-labeled sample set;

[0066] S224. Integrate the labeled sample set and the pseudo-labeled sample set to obtain a training sample set;

[0067] S225. Use the training sample set to train the large language model to obtain an optimized large language model.

[0068] As an alternative implementation, in the second aspect of the embodiments of the present invention, the process of using the first large language model to process the unlabeled sample set to obtain a pseudo-labeled sample set includes:

[0069] S2231. Extract the unlabeled sample set to obtain a first unlabeled sample set;

[0070] S2232. Perform confidence processing on the first unlabeled sample set to obtain a second unlabeled sample set;

[0071] S2233. Perform a difference test on the second unlabeled sample set to obtain a pseudo-labeled sample set.

[0072] As an alternative implementation, in the second aspect of the embodiments of the present invention, the process of rendering the situation simulation result to obtain a situation display result includes:

[0073] S41. Perform visualization processing on the situation simulation result to obtain a situation simulation diagram;

[0074] S42. Denoise the situation simulation diagram to obtain a denoised situation simulation diagram;

[0075] S43. Perform visual enhancement on the denoised situation simulation diagram to obtain a situation display result.

[0076] The third aspect of the present invention discloses another situation simulation device based on a large language model, and the device includes:

[0077] A memory storing executable program code;

[0078] A processor coupled to the memory;

[0079] The processor calls the executable program code stored in the memory and executes some or all of the steps in the situation simulation method based on a large language model disclosed in the first aspect of the embodiments of the present invention.

[0080] The fourth aspect of the present invention discloses a computer-readable storage medium, and the computer-readable storage medium stores computer instructions that, when called, are used to execute some or all of the steps in the situation simulation method based on a large language model disclosed in the first aspect of the embodiments of the present invention.

[0081] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0082] The present invention discloses a situation simulation method and device based on a large language model. By constructing a situation simulation information library, using the large language model to process the situation simulation information library, obtaining a situation simulation result, and performing processing such as denoising and visual enhancement on the situation simulation result, a situation display result is obtained. The method of the present invention has important theoretical significance and application value for the presentation method and interaction design of situation information in situation simulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0084] Figure 1 is a flowchart showing a situation simulation method based on a large language model disclosed in an embodiment of the present invention;

[0085] Figure 2 is a structural diagram showing a situation simulation device based on a large language model disclosed in an embodiment of the present invention;

[0086] Figure 3 is a structural diagram showing another situation simulation device based on a large language model disclosed in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0087] In order to enable those skilled in the art to better understand the solution of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0088] The terms "first", "second", etc. in the specification and claims of the present invention and the above drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, device, product, or equipment that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products, or equipment.

[0089] Reference to "embodiments" in this document means that the specific features, structures, or characteristics described in connection with the embodiments can be included in at least one embodiment of the present invention. The phrase appears in various places in the specification and does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0090] The present invention discloses a situation simulation method and device based on a large language model. The method includes: constructing a situation simulation information library; the simulation information library includes entity names, entity attributes, entity locations, and communication relationships between entities; using the situation simulation information library to train the large language model to obtain an optimized large language model; obtaining the situation simulation requirement information of the user, and using the optimized large language model to process the situation simulation requirement information of the user to obtain a situation simulation result; rendering the situation simulation result to obtain a situation display result. The method of the present invention has important theoretical significance and application value for the presentation method and interaction design of situation information in situation simulation. The following will be described in detail respectively.

[0091] Embodiment 1

[0092] Please refer to Figure 1 , Figure 1 which is a flowchart of a situation simulation method based on a large language model disclosed in an embodiment of the present invention. Among them, Figure 1 The described situation simulation method based on a large language model is applied to the field of situation simulation technology, and the embodiments of the present invention do not make limitations. As Figure 1 shown, the situation simulation method based on a large language model can include the following operations:

[0093] S1, constructing a situation simulation information library; the simulation information library includes entity names, entity attributes, entity locations, and communication relationships between entities;

[0094] S2, using the situation simulation information library to train the large language model to obtain an optimized large language model;

[0095] S3, obtaining the situation simulation requirement information of the user, and using the optimized large language model to process the situation simulation requirement information of the user to obtain a situation simulation result;

[0096] S4, rendering the situation simulation result to obtain a situation display result.

[0097] Optionally, the using the situation simulation information library to train the large language model to obtain an optimized large language model includes:

[0098] S21. Process the situation simulation information library to obtain situation simulation data information;

[0099] The situation simulation data information includes a set of graph structures G of entities, entity screen coordinate information, and entity movement trajectory information;

[0100] S22. Use the situation simulation data information to train a large language model to obtain an optimized large language model.

[0101] Optionally, the process of processing the situation simulation information library to obtain situation simulation data information includes:

[0102] S211. Process the situation simulation information library to obtain a set of graph structures G of entities;

[0103] The expression of the set of graph structures G of entities is:

[0104] G = {V, E, Σ, L}

[0105] Among them, V is a non-empty set of vertices, and each vertex represents an entity. is a set of edges, Σ is a set of vertex labels, and L is a label assignment function. For all have L(v) representing the label of vertex v, where L(v) ∈ Σ;

[0106] S212. Process the set of graph structures G to obtain entity screen coordinate information;

[0107] Obtain the geographical coordinates of the entity. The definition rule of geographical coordinates is:

[0108] The X-axis (representing longitude) increases to the right, and the Y-axis (latitude) increases upward;

[0109] The definition rule of screen coordinates is:

[0110] The X-axis increases to the right, and the Y-axis increases downward;

[0111] The difference between geographical coordinates and screen coordinates is only that the increasing direction of the Y-axis is opposite.

[0112] Given the height (y) and width (h) of the screen, and the regional range of geographical coordinates is (maxlon, minlon, maxlat, minlat); the longitude and latitude represented by each pixel can be calculated. The calculation formula is:

[0113] scaleX = h / ((maxlon - minlon) * 3600) --- The number of longitude seconds represented by each pixel on the X-axis

[0114] scaleY = y / ((maxlat - minlat) * 3600) --- The number of seconds of latitude represented by each pixel on the Y-axis

[0115] The coordinates of any point (lon, lat) on the geographical coordinate area on the screen are:

[0116] screenX = lon * 3600 / scaleX --- Screen coordinate X-axis coordinate

[0117] screenY = lat * 3600 / scaleY --- Screen coordinate Y-axis coordinate

[0118] minx = minlon * 3600 / scaleX --- The leftmost side of the area is placed at the leftmost end

[0119] minY = minlat * 3600 / scaleY --- The upper side of the area is placed at the uppermost end

[0120] When the geographical range area fills the entire screen, using the two parameters screenX and screenY, the formula for calculating any point in the area is:

[0121] X = screenX - minX = (lon - minlon) * 3600 / scaleX

[0122] Since the direction of latitude is opposite to the screen Y-axis

[0123] screenMaxlat = (mxlat - minlat) * 3600 / scaleY

[0124] screenlat = (lat - minlat) * 3600 / scaleY

[0125] Y = screenMaxlat - screenlat = (maxlat - lat)scaleY

[0126] The formula for converting longitude and latitude to screen coordinates is:

[0127] X = (lon - minlon) * 3600 / scaleX

[0128] Y = (maxlat - lat) * 3600 / scaleY

[0129] The formula for converting screen coordinates to longitude and latitude coordinates is:

[0130] lon = X * scaleX / 3600 + minlon

[0131] lat = maxlat - y * scaleY / 3600

[0132] S213. Process the communication relationship between the entities to obtain entity movement trajectory information.

[0133] Optionally, the process of processing the communication relationship between the entities to obtain entity movement trajectory information includes:

[0134] S2131. Process the communication relationship between the entities to obtain the environmental information, initial position information, and target position information of the entities;

[0135] S2132. Process the environmental information, initial position information, and target position information of the entities to obtain network training parameter information;

[0136] The network training parameter information includes the action a to be taken currently, the next state s obtained by executing the action a t+1 , the immediate reward r, and the end signal;

[0137] S2133. Use the network training parameter information to train a preset neural network model to obtain an optimized neural network model;

[0138] S2134. Use the optimized neural network model to process the communication relationship between the entities to be processed to obtain entity movement trajectory information.

[0139] The specific steps include:

[0140] 1. Environment setting: Set the map for entity - to - entity communication. The goal is to find the optimal path to reach the target point. The state of the entity is defined as a triple (x, y, α), where x and y are the real - time coordinates of the entity in the simulation map, and α is the yaw angle between the entity and the target point. Let the initial coordinates of the entity be (x0, y0), then the initial actual heading angle is initialized as θ = arctan(x0, y0). Let the current coordinates be (x, y) and the target coordinates be (p, q), then the calculation formula for the desired heading β is:

[0141] β = arctan(x - p, y - q)-arctan(x, y)

[0142] The actual heading angle θ of the entity changes with each execution of an action. Let the corresponding direction angle of the action a be a θ , then the update formula for θ is: θ new = θ old + a θ

[0143] The yaw angle α is defined as: α = β - θ

[0144] The steering angle Δθ is defined as: Δθ = new -θ old

[0145] At time t, the value of the steering angle Δθ is the action angle a taken by the entity at this time. θ .

[0146] 2. Set the reward function;

[0147] 1) Initialize the neural network parameters ω, θ, initialize the state sequence cache M and the training loop counter C.

[0148] 2) Initialize the input sequence input and the starting state s0.

[0149] 3) Use the Actor value network to obtain the action a to be taken currently, execute the action a to obtain the environmental feedback: the next state s t+1 , the immediate reward r, the end signal Do, and the collision signal Co.

[0150] 4) Judge whether Do is True. If so, end this training and jump to 2).

[0151] 5) Store s t , a, r, s t+1 as the state-action sequence in the state sequence cache M for sampling from it to train the neural network.

[0152] 6) Sample the current state sequence input input_current and the next state sequence input input_next from the state sequence cache M and input them into the Critic value network and the Critic target network respectively to obtain the current state value V(s t ) and the next state value V(s t+1 ).

[0153] 7) Calculate the first difference information and use the first difference information to update the parameters ω of the Critic value network. The loss function of the Critic value network selects the mean square error function. The expression of the first difference information is as follows, where i is the sample number:

[0154]

[0155] The loss function of the Critic network is defined as the square of the first difference information error, that is:

[0156]

[0157] The update formula for the Critic network parameters is:

[0158]

[0159] Input the current state sequence input into the Actor value network to obtain the probability distribution of the action output corresponding to the current state sequence. Update the parameters θ of the Actor value network using the first difference information. The loss function expression of the Actor network is as follows:

[0160]

[0161] The update formula for the Actor network parameters is:

[0162]

[0163] Every 30 steps of actions are executed by the agent, and the parameters of the Critic value network and the Critic target network are synchronized.

[0164] Judge whether Co is True. If it is, go back to the previous state and jump to 3).

[0165] Judge whether C reaches the maximum value. If it does, end the training. If it does, jump to 2).

[0166] Optionally, using the situation simulation data information to train the large language model to obtain an optimized large language model includes:

[0167] S221, divide the situation simulation data information to obtain a labeled sample set and an unlabeled sample set;

[0168] The division ratio is set by experiments. In this embodiment, the labeled sample set is 30% and the unlabeled sample set is 50%;

[0169] S222, use the labeled sample set to train the large language model to obtain the first large language model;

[0170] The large language model is a deep learning model trained using a large amount of text data, which can generate natural language text or understand the meaning of language text. The formula used is the text2vec model, which mainly vectorizes the text. The text2vec model can represent the entire text as a vector sequence, thus retaining the semantic information of the text. Convert the question to be asked into a vector, and at this time, perform vector similarity comparison.

[0171] S223, use the first large language model to process the unlabeled sample set to obtain a pseudo-labeled sample set;

[0172] S224, integrate the labeled sample set and the pseudo-labeled sample set to obtain a training sample set;

[0173] Specifically: Combine the labeled sample set and the pseudo-labeled sample set together;

[0174] S225. Use the training sample set to train the large language model to obtain an optimized large language model.

[0175] Optionally, using the first large language model to process the unlabeled sample set to obtain a pseudo-labeled sample set includes:

[0176] S2231. Extract the unlabeled sample set to obtain a first unlabeled sample set;

[0177] Extract 10% of the unlabeled sample set to obtain a first unlabeled sample set;

[0178] S2232. Perform confidence processing on the first unlabeled sample set to obtain a second unlabeled sample set;

[0179] Specifically, use the first large language model to select high-confidence samples S from the first unlabeled sample set and assign them pseudo-labels to obtain a second unlabeled sample set;

[0180] Among them, high-confidence samples are unlabeled samples with a comparison membership degree greater than the membership degree threshold;

[0181] The comparison membership degree is:

[0182] N ij = max(P(C j |x i ; φ)) / max(P(C j′ |x i ; φ))

[0183] In the formula: N ij is the comparison membership degree between sample x i and sample x j , j′∈{1, 2, …, c|j′≠j}, i∈{1, 2, …, m}, j∈{1, 2, …, c}, P(C j |x i ; φ) is the probability that sample x i belongs to category C j , c is the number of categories, P(C j′ |x i ; φ) is the probability that sample x i belongs to category C j′ , m is the number of unlabeled samples, and φ is the membership function;

[0184] S2233. Perform a difference test on the second unlabeled sample set to obtain a pseudo-labeled sample set.

[0185] The difference test is to remove sample points with a local outlier factor greater than the threshold. The local outlier factor is defined as:

[0186] τ(x i ) = 1 / lrd k (x i )

[0187] where τ(x i ) is the local outlier factor of sample x i .

[0188]

[0189] N k (x i ) is the set excluding sample x i .

[0190] Optionally, rendering the situation simulation result to obtain a situation display result includes:

[0191] S41, performing visualization processing on the situation simulation result to obtain a situation simulation diagram;

[0192] S42, denoising the situation simulation diagram to obtain a denoised situation simulation diagram;

[0193] The denoising method is:

[0194] Processing the situation simulation diagram to obtain a real component and an imaginary component;

[0195] Performing wavelet transform on the real component to obtain first wavelet parameter information;

[0196] Performing wavelet transform on the imaginary component to obtain second wavelet parameter information;

[0197] Performing filtering processing on the first wavelet parameter information to obtain third wavelet parameter information;

[0198] Performing filtering processing on the second wavelet parameter information to obtain fourth wavelet parameter information;

[0199] The method of filtering processing is:

[0200] Setting the target interval [α, β] of the initial threshold δ th , where 0 ≤ α < β ≤ max{S(m, n)}, and S(m, n) is the wavelet parameter information;

[0201] Dividing the target interval [α, β] into L equal parts;

[0202] Setting the initial threshold δ th to α, and performing incremental iteration at the step size of to obtain the threshold

[0203] Using a threshold Calculate the wavelet parameter information after filtering for each iteration

[0204]

[0205] Where S(m,n) represents the original wavelet parameter information, represents the wavelet parameter information after filtering, and a, b represent the time and frequency parameters in the threshold;

[0206] Calculate the difference between the wavelet parameter information obtained in each iteration and the previous iteration is the number of iterations;

[0207] Take the difference Perform curve fitting to obtain curve C. When an inflection point appears on curve C, the corresponding threshold is the optimal threshold;

[0208] Use the optimal threshold to perform optimal threshold filtering on the wavelet parameter information S(m,n) to obtain the wavelet parameter information after filtering;

[0209] Reconstruct the third wavelet parameter information and the fourth wavelet parameter information to obtain a reconstructed time-frequency image;

[0210] Perform time-frequency inverse transformation on the reconstructed time-frequency image to obtain a denoised situation simulation diagram;

[0211] S43. Perform visual enhancement on the denoised situation simulation diagram to obtain a situation display result.

[0212] The method of visual enhancement is:

[0213] S431. Process the denoised situation simulation diagram to obtain a reference image;

[0214] Specifically, downsample the denoised situation simulation diagram to obtain a reference image;

[0215] S432. Perform normalization processing on the denoised situation simulation diagram and the reference image to obtain a normalized denoised situation simulation diagram and a normalized reference image;

[0216] S433. Segment the denoised situation simulation diagram to obtain N visible light image blocks;

[0217] S434. Segment the normalized reference image to obtain N reference image blocks;

[0218] The N visible light image blocks and the N reference image blocks are of equal size;

[0219] S435. Calculate the similarity between the N visible light image blocks and the N reference image blocks to obtain a similarity value.

[0220] S436. According to the similarity value, perform migration processing on the N visible light image blocks and the N reference image blocks to obtain a visually enhanced image, which is the situation display result.

[0221] It can be seen that the present invention discloses a situation simulation method and device based on a large language model. By constructing a situation simulation information library, using the large language model to process the situation simulation information library, obtaining a situation simulation result, and performing denoising, visual enhancement, etc. on the situation simulation result, a situation display result is obtained. The method of the present invention has important theoretical significance and application value for the presentation method and interaction design of situation information in situation simulation.

[0222] Embodiment 2

[0223] Please refer to Figure 2 , Figure 2 which is a schematic structural diagram of a situation simulation device based on a large language model disclosed in an embodiment of the present invention. Among them, Figure 2 The described situation simulation device based on a large language model is applied to the field of situation simulation technology, and the embodiments of the present invention do not make limitations. As Figure 2 shown, the situation simulation device based on a large language model may include the following operations:

[0224] S301. An information library construction module, configured to construct a situation simulation information library; the simulation information library includes entity names, entity attributes, entity locations, and communication relationships between entities.

[0225] S302. A model training module, configured to use the situation simulation information library to train a large language model to obtain an optimized large language model.

[0226] S303. A situation simulation module, configured to obtain the user's situation simulation requirement information, and use the optimized large language model to process the user's situation simulation requirement information to obtain a situation simulation result.

[0227] S304. A situation display module, configured to render the situation simulation result to obtain a situation display result.

[0228] Embodiment 3

[0229] Please refer to Figure 3 , Figure 3 which is a schematic structural diagram of another situation simulation device based on a large language model disclosed in an embodiment of the present invention. Among them, Figure 3 The described situation simulation device based on a large language model is applied to the field of situation simulation technology, and the embodiments of the present invention do not make limitations. AsFigure 3 As shown, the situation simulation device based on the large language model may include the following operations:

[0230] A memory 401 storing executable program code;

[0231] A processor 402 coupled to the memory 401;

[0232] The processor 402 calls the executable program code stored in the memory 401 to execute the steps in the situation simulation method based on the large language model described in the first embodiment.

[0233] Embodiment 4

[0234] The embodiment of the present invention discloses a computer-readable storage medium storing a computer program for electronic data exchange, wherein the computer program enables a computer to execute the steps in the situation simulation method based on the large language model described in the first embodiment.

[0235] The device embodiments described above are only illustrative. The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed to multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative labor.

[0236] Through the specific descriptions of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, and the storage medium includes read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc memories, magnetic disk memories, tape memories, or any other medium that can be used to carry or store data and is computer-readable.

[0237] Finally, it should be noted that: What is disclosed in an embodiment of a situation simulation method and device based on a large language model of the present invention is only a preferred embodiment of the present invention, and is only used to illustrate the technical solution of the present invention, rather than limiting it; Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: They can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; And these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A situation simulation method based on large language models, characterized in that, The method includes: S1. Construct a situation simulation information library; the simulation information library includes entity names, entity attributes, entity locations, and communication relationships between entities; S2. Use the situation simulation information library to train a large language model to obtain an optimized large language model; S3. Obtain the situation simulation requirement information of the user, and use the optimized large language model to process the situation simulation requirement information of the user to obtain a situation simulation result; S4. Render the situation simulation result to obtain a situation display result.

2. The situation simulation method based on a large language model according to claim 1, wherein The step of using the situation simulation information library to train a large language model to obtain an optimized large language model includes: S21. Process the situation simulation information library to obtain situation simulation data information; The situation simulation data information includes a set of graph structures G of entities, entity screen coordinate information, and entity movement trajectory information; S22. Use the situation simulation data information to train a large language model to obtain an optimized large language model.

3. The situation simulation method based on a large language model according to claim 2, wherein The step of processing the situation simulation information library to obtain situation simulation data information includes: S211. Process the situation simulation information library to obtain a set of graph structures G of entities; The expression of the set of graph structures G of entities is: G = {V, E, Σ, L} where V is a non-empty set of vertices, each vertex representing an entity, is a set of edges, Σ is a set of vertex labels, L is a label assignment function, for all v ∈ V, L(v) represents the label of vertex v, where L(v) ∈ Σ; S212. Process the set of graph structures G to obtain entity screen coordinate information; S213. Process the communication relationship between entities to obtain entity movement trajectory information.

4. The situation simulation method based on a large language model according to claim 3, wherein The step of processing the communication relationship between entities to obtain entity movement trajectory information includes: S2131. Process the communication relationship between entities to obtain the environmental information, initial position information, and target position information of the entities; S2132. Process the environmental information, initial position information, and target position information of the entities to obtain network training parameter information; The network training parameter information includes the current action a to be taken, the next state s obtained by executing action a t+1 , immediate reward r and end signal; S2133. Use the network training parameter information to train a preset neural network model to obtain an optimized neural network model; S2134. Use the optimized neural network model to process the communication relationship between entities to be processed to obtain entity movement trajectory information.

5. The situation simulation method based on a large language model according to claim 2, wherein The step of using the situation simulation data information to train a large language model to obtain an optimized large language model includes: S221. Divide the situation simulation data information to obtain a labeled sample set and an unlabeled sample set; S222. Use the labeled sample set to train a large language model to obtain a first large language model; S223. Use the first large language model to process the unlabeled sample set to obtain a pseudo-labeled sample set; S224. Integrate the labeled sample set and the pseudo-labeled sample set to obtain a training sample set; S225. Use the training sample set to train a large language model to obtain an optimized large language model.

6. The situation simulation method based on a large language model according to claim 5, wherein, The step of using the first large language model to process the unlabeled sample set to obtain a pseudo-labeled sample set includes: S2231. Extract the unlabeled sample set to obtain a first unlabeled sample set; S2232. Perform confidence processing on the first set of unlabeled samples to obtain a second set of unlabeled samples; S2233. Perform a difference test on the second set of unlabeled samples to obtain a set of pseudo-labeled samples.

7. The situation simulation method based on a large language model according to claim 1, characterized in that: The rendering of the situation simulation result to obtain a situation display result includes: S41. Perform visualization processing on the situation simulation result to obtain a situation simulation diagram; S42. Denoise the situation simulation diagram to obtain a denoised situation simulation diagram; S43. Perform visual enhancement on the denoised situation simulation diagram to obtain a situation display result.

8. A situation simulation device based on a large language model, characterized in that, The device includes: An information library construction module for constructing a situation simulation information library; the simulation information library includes entity names, entity attributes, entity locations, and communication relationships between entities; A model training module for training a large language model using the situation simulation information library to obtain an optimized large language model; A situation simulation module for obtaining the user's situation simulation requirement information and using the optimized large language model to process the user's situation simulation requirement information to obtain a situation simulation result; A situation display module for rendering the situation simulation result to obtain a situation display result.

9. A situation simulation device based on a large language model, characterized in that: The device includes: A memory storing executable program code; A processor coupled to the memory; The processor invokes the executable program code stored in the memory to execute the situation simulation method based on a large language model according to any one of claims 1-7.

10. A computer storable medium, characterized in that The computer-readable storage medium stores computer instructions which, when invoked, are used to execute the situation simulation method based on a large language model according to any one of claims 1-7.

Images