Project auxiliary support system and method based on artificial intelligence

Abstract

The present disclosure relates to an artificial intelligence-based project auxiliary support system and method thereof. The method includes: performing one-hot encoding on the acquired first project feature type; performing sentence-granular semantic encoding on the acquired project text description segment; calculating the subject-content semantic association between the sequence of semantic encoding vectors of the project text description sentence and the one-hot encoding vector of the first project feature type; then, after normalizing the sequence of subject-content semantic associations, using the sequence of normalized subject-content semantic associations as weighted weights to perform important content enhancement optimization to obtain a sequence of optimized project text description sentence semantic encoding vectors; then, passing it through a decoder-based project feature generator to obtain a project feature text description. In this way, the project feature text description can be automatically generated according to the actual required project feature type, providing auxiliary support for the project team to make decisions and plans.

Description

Technical Field

[0001] The present disclosure relates to the technical field of project documents, and in particular, to an artificial intelligence-based project auxiliary support system and method thereof. Background Art

[0002] Reading the contents of project documents is a key step in the project management process. It involves in-depth analysis and understanding of documents such as project plans, schedules, risk assessment reports, cost analysis, etc., to ensure that the project team can accurately grasp the core elements and potential problems of the project, and provide support for the project team to make decisions and plans.

[0003] However, the traditional process of reading and understanding project document content usually relies on manual reading and analysis. The manual reading method requires personnel to spend a lot of time and energy to understand the content of project documents word by word, which is inefficient when dealing with large amounts of documents or information. At the same time, the manual reading method is easily affected by personal interpretation differences and experience. Different people may have different understandings and key points of the same document, resulting in inconsistent information interpretation. In addition, the traditional project document content reading method has a low level of intelligence and cannot automatically extract and summarize key information and project features from the text, which limits the in-depth understanding and utilization of project documents.

[0004] Therefore, an artificial intelligence-based project auxiliary support system is expected, which can use artificial intelligence technology to automatically conduct in-depth understanding of project documents, thereby extracting important project features and providing auxiliary support for project teams in making decisions and planning. Summary of the invention

[0005] This summary is provided to introduce concepts in a brief form that will be described in detail in the detailed description below. This summary is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0006] In a first aspect, the present disclosure provides a project auxiliary support method based on artificial intelligence, the method comprising:

[0007] extracting a first project feature type from a project feature type database;

[0008] One-hot encoding the first project feature type to obtain a first project feature type one-hot encoding vector;

[0009] Get the project text description segment;

[0010] Performing sentence-granularity-based semantic coding on the project text description segment to obtain a sequence of project text description sentence semantic coding vectors;

[0011] Calculating the topic-content semantic association between each item text description sentence semantic encoding vector in the sequence of item text description sentence semantic encoding vectors and the first item feature type one-hot encoding vector to obtain a sequence of topic-content semantic associations;

[0012] After normalizing the sequence of subject-content semantic associations, the sequence of the normalized subject-content semantic associations is used as a weighted weight to perform important content enhancement optimization on the sequence of semantic coding vectors of the project text description sentences to obtain a sequence of optimized semantic coding vectors of the project text description sentences;

[0013] The sequence of the optimized semantic encoding vectors of the project text description sentences is passed through a decoder-based project feature generator to obtain a project feature text description.

[0014] Optionally, the project text description segments are semantically encoded based on sentence granularity to obtain a sequence of project text description sentence semantic encoding vectors, including: sentence processing of the project text description segments to obtain a sequence of project text description sentences; passing the sequence of project text description sentences through a sentence encoder to obtain a sequence of project text description sentence semantic encoding vectors.

[0015] Optionally, the topic-content semantic association between each project text description sentence semantic encoding vector in the sequence of the project text description sentence semantic encoding vectors and the first project feature type one-hot encoding vector is calculated to obtain a sequence of topic-content semantic associations, including: calculating the product between the project text description sentence semantic encoding vector and a first trainable weight vector to obtain a project text description sentence semantic global representation coefficient; calculating the product between the first project feature type one-hot encoding vector and a second trainable weight vector to obtain a first project feature type understanding coefficient; adding the project text description sentence semantic global representation coefficient and the project feature type understanding coefficient and performing activation processing to obtain the topic-content semantic association.

[0016] Optionally, the semantic global representation coefficient of the project text description sentence and the one-project feature type understanding coefficient are added and then activated to obtain the topic-content semantic association, including: adding the semantic global representation coefficient of the project text description sentence and the one-project feature type understanding coefficient and then inputting An activation process is performed in the activation function to obtain the topic-content semantic association.

[0017] Optionally, after normalizing the sequence of topic-content semantic associations, the sequence of normalized topic-content semantic associations is used as weighted weights to perform important content enhancement optimization on the sequence of project text description sentence semantic coding vectors to obtain a sequence of optimized project text description sentence semantic coding vectors, including: performing maximum value-based normalization on the sequence of topic-content semantic associations to obtain a sequence of topic-content semantic association weighting coefficients; using the sequence of topic-content semantic association weighting coefficients as weights, weighting the sequence of project text description sentence semantic coding vectors to obtain the sequence of optimized project text description sentence semantic coding vectors.

[0018] Optionally, the sequence of semantic coding vectors of the optimized project text description sentences is passed through a decoder-based project feature generator to obtain a project feature text description, including: cascading the sequence of semantic coding vectors of the optimized project text description sentences into an optimized project text description sentence semantic joint coding vector; performing feature correction on the optimized project text description sentence semantic joint coding vector to obtain a corrected optimized project text description sentence semantic joint coding vector; and passing the corrected semantic joint coding vector of the optimized project text description sentence through the decoder-based project feature generator to obtain the project feature text description.

[0019] Optionally, passing the corrected optimized project text description sentence semantic joint encoding vector through the decoder-based project feature generator to obtain the project feature text description, comprising: using the decoder-based project feature generator to perform decoding regression on the corrected optimized project text description sentence semantic joint encoding vector using the following decoding formula to obtain the project feature text description; wherein the decoding formula is: ,in, The semantic joint encoding vector representing the corrected optimized item text description sentence, Represents a textual description of the project characteristics, represents the weight matrix, represents the bias vector, Represents matrix multiplication.

[0020] In a second aspect, the present disclosure provides a project auxiliary support system based on artificial intelligence, the system comprising:

[0021] A project feature type extraction module, used to extract a first project feature type from a project feature type database;

[0022] A one-hot encoding module, configured to perform one-hot encoding on the first project feature type to obtain a one-hot encoding vector of the first project feature type;

[0023] A text description segment acquisition module is used to acquire a project text description segment;

[0024] A semantic encoding module, used for performing sentence-granular semantic encoding on the project text description segment to obtain a sequence of semantic encoding vectors of project text description sentences;

[0025] A topic-content semantic association calculation module, used to calculate the topic-content semantic association between each project text description sentence semantic encoding vector in the sequence of project text description sentence semantic encoding vectors and the first project feature type one-hot encoding vector to obtain a sequence of topic-content semantic associations;

[0026] An important content enhancement optimization module is used to normalize the sequence of subject-content semantic associations, and then use the normalized sequence of subject-content semantic associations as weighted weights to perform important content enhancement optimization on the sequence of semantic coding vectors of the project text description sentences to obtain a sequence of optimized semantic coding vectors of the project text description sentences;

[0027] A decoding module is used to pass the sequence of the optimized project text description sentence semantic encoding vectors through a decoder-based project feature generator to obtain a project feature text description.

[0028] Optionally, the semantic encoding module includes: a sentence processing unit, used to perform sentence processing on the project text description segment to obtain a sequence of project text description sentences; and a sentence encoding unit, used to pass the sequence of project text description sentences through a sentence encoder to obtain a sequence of semantic encoding vectors of the project text description sentences.

[0029] Optionally, the topic-content semantic association calculation module includes: a global representation coefficient calculation unit, used to calculate the product between the semantic encoding vector of the project text description sentence and the first trainable weight vector to obtain the project text description sentence semantic global representation coefficient; an understanding coefficient calculation unit, used to calculate the product between the first project feature type one-hot encoding vector and the second trainable weight vector to obtain the first project feature type understanding coefficient; an activation processing unit, used to add the project text description sentence semantic global representation coefficient and the one project feature type understanding coefficient and then perform activation processing to obtain the topic-content semantic association.

[0030] The above technical solution is adopted to perform one-hot encoding on the acquired first project feature type; perform sentence-granular semantic encoding on the acquired project text description segment; calculate the subject-content semantic association between the sequence of semantic encoding vectors of project text description sentences and the one-hot encoding vector of the first project feature type; then, after normalizing the sequence of subject-content semantic association, use the sequence of normalized subject-content semantic association as weighted weight to perform important content enhancement optimization to obtain a sequence of optimized project text description sentence semantic encoding vectors; then, pass it through a decoder-based project feature generator to obtain a project feature text description. In this way, the project feature text description can be automatically generated according to the actual required project feature type, providing auxiliary support for the project team to make decisions and plans.

[0031] Other features and advantages of the present disclosure will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale. In the drawings:

[0033] Figure 1 The present invention is a flowchart of a project auxiliary support method based on artificial intelligence according to an exemplary embodiment.

[0034] Figure 2 It is a block diagram of a project auxiliary support system based on artificial intelligence according to an exemplary embodiment.

[0035] Figure 3 It is a block diagram of an electronic device according to an exemplary embodiment.

[0036] Figure 4 It is an application scenario diagram of an artificial intelligence-based project auxiliary support method according to an exemplary embodiment. DETAILED DESCRIPTION

[0037] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0038] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.

[0039] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0040] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0041] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0042] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0043] In order to solve the above problems, the present disclosure provides a project auxiliary support system and method based on artificial intelligence, which performs one-hot encoding on the acquired first project feature type; performs sentence-granular semantic encoding on the acquired project text description segment; calculates the subject-content semantic association between the sequence of semantic encoding vectors of the project text description sentence and the one-hot encoding vector of the first project feature type; then, after normalizing the sequence of subject-content semantic association, the sequence of normalized subject-content semantic association is used as a weighted weight to perform important content enhancement optimization to obtain a sequence of semantic encoding vectors of optimized project text description sentences; then, it is passed through a decoder-based project feature generator to obtain a project feature text description. In this way, the project feature text description can be automatically generated according to the actual required project feature type, providing auxiliary support for the project team to make decisions and plans.

[0044] The specific embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

[0045] Figure 1 is a flowchart of a project auxiliary support method based on artificial intelligence according to an exemplary embodiment. Figure 1As shown, the method includes:

[0046] Step 101: extracting a first project feature type from a project feature type database;

[0047] Step 102: One-hot encode the first project feature type to obtain a first project feature type one-hot encoding vector;

[0048] Step 103: Obtain the project text description segment;

[0049] Step 104: performing sentence-granular semantic coding on the project text description segment to obtain a sequence of project text description sentence semantic coding vectors;

[0050] Step 105, calculating the topic-content semantic association between each item text description sentence semantic encoding vector in the sequence of item text description sentence semantic encoding vectors and the first item feature type one-hot encoding vector to obtain a sequence of topic-content semantic associations;

[0051] Step 106: After normalizing the sequence of subject-content semantic associations, the sequence of subject-content semantic associations after normalization is used as a weighted weight to perform important content enhancement optimization on the sequence of semantic coding vectors of the project text description sentences to obtain a sequence of optimized semantic coding vectors of the project text description sentences;

[0052] Step 107: Pass the sequence of the optimized semantic coding vectors of the project text description sentences through a decoder-based project feature generator to obtain a project feature text description.

[0053] In response to the above technical problems, in the technical solution of this application, an artificial intelligence-based project auxiliary support system is proposed, which can automatically encode project text descriptions by using semantic encoders based on artificial intelligence and deep learning to quickly understand the content of project documents and extract important project features, such as risks, costs, timelines, etc. In this way, the project feature text description can be automatically generated according to the actual required project feature types, providing auxiliary support for the project team to make decisions and plans.

[0054] Specifically, in the technical solution of the present application, first, the first project feature type is extracted from the project feature type database. It should be understood that, considering that in the actual process of semantic understanding of the project text, in order to capture the required project feature types, it is necessary to extract the project feature types from the project feature type database in advance. In addition, it is also considered that the project feature types extracted from the project feature type database have different semantic information, and this semantic information is about different types of semantics of project features. Therefore, in the technical solution of the present application, the first project feature type is one-hot encoded to obtain a first project feature type one-hot encoded vector. Through the one-hot encoding method, the first project feature type can be encoded into a vector form in which the type element response represented by a specific position is 1 and the rest are 0, avoiding the problem of size or order relationship between different feature types, ensuring the independence between features, and providing a generation basis and foundation for subsequent project feature text generation.

[0055] When semantic understanding of the project text is performed to automatically generate the required text description of project features, it is necessary to perform semantic understanding and feature extraction on the project document content. Based on this, in the technical solution of the present application, first, a project text description segment is obtained. It should be understood that project documents usually contain a large amount of text information, which describes various aspects of the project, such as goals, plans, risks, etc. First, it is necessary to extract relevant text description segments from the project document as the object of processing. Then, the project text description segment is sentence-processed to obtain a sequence of project text description sentences. In this way, long project text description segments can be split into independent sentences, each sentence representing a complete semantic unit. This helps to perform more fine-grained processing and analysis of project texts and improve the accuracy of semantic understanding.

[0056] In one embodiment of the present disclosure, the project text description segment is semantically encoded based on sentence granularity to obtain a sequence of project text description sentence semantic encoding vectors, including: sentence processing is performed on the project text description segment to obtain a sequence of project text description sentences; the sequence of project text description sentences is passed through a sentence encoder to obtain a sequence of project text description sentence semantic encoding vectors.

[0057] Next, in order to understand each project text description sentence in the sequence of project text description sentences, it is necessary to pass the sequence of project text description sentences through a sentence encoder for semantic understanding and semantic feature extraction to obtain a sequence of semantic encoding vectors of project text description sentences. By inputting each project text description sentence into a sentence encoder, each project text description sentence can be semantically encoded to extract the semantic feature information expressed in the project text description sentence, thereby achieving semantic understanding and representation of the project text description sentence information, providing a basis for subsequent project feature extraction and project feature text description generation, and providing more intelligent support and assistance for project management.

[0058] Further, in order to be able to judge the semantic relevance between the semantic features of each sentence in the project text description segment and the first project feature type, so as to capture the semantics of sentences in the project text description segment that are more relevant to the first project feature type, thereby generating a more comprehensive and accurate project feature text description, in the technical solution of the present application, the subject-content semantic relevance between the semantic encoding vectors of each project text description sentence in the sequence of the project text description sentence semantic encoding vectors and the first project feature type unique hot encoding vector is further calculated to obtain a sequence of subject-content semantic relevance. By calculating the subject-content semantic relevance between the semantic encoding vector of each sentence and the unique hot encoding vector of a specific project feature type, the semantic relevance between each sentence and the feature type can be evaluated. If the semantic information of the sentence is highly correlated with a specific feature type, then its relevance will be higher, otherwise it will be lower. This step helps to determine the degree of relevance between each sentence and the first project feature type and its importance in the entire document, so as to more accurately extract sentences related to a specific project feature type and generate a more accurate project feature text description.

[0059] In one embodiment of the present disclosure, the topic-content semantic association between each project text description sentence semantic encoding vector in the sequence of the project text description sentence semantic encoding vector and the first project feature type one-hot encoding vector is calculated to obtain a sequence of topic-content semantic associations, including: calculating the product between the project text description sentence semantic encoding vector and the first trainable weight vector to obtain the project text description sentence semantic global representation coefficient; calculating the product between the first project feature type one-hot encoding vector and the second trainable weight vector to obtain the first project feature type understanding coefficient; adding the project text description sentence semantic global representation coefficient and the one project feature type understanding coefficient and performing activation processing to obtain the topic-content semantic association.

[0060] Further, the semantic global representation coefficient of the project text description sentence and the one-project feature type understanding coefficient are added and then activated to obtain the topic-content semantic association, including: adding the semantic global representation coefficient of the project text description sentence and the one-project feature type understanding coefficient and then inputting An activation process is performed in the activation function to obtain the topic-content semantic association.

[0061] Specifically, the following topic-content semantic association formula is used to calculate the topic-content semantic association between each item text description sentence semantic encoding vector in the sequence of item text description sentence semantic encoding vectors and the first item feature type one-hot encoding vector to obtain the sequence of the topic-content semantic association;

[0062] The topic-content semantic association formula is: ;in, is the first in the sequence of semantic encoding vectors of the project text description sentence The semantic encoding vector of the item text description sentence, is the one-hot encoding vector of the first item feature type, and are all trainable weight vectors, for Activation function, is the first in the sequence of the subject-content semantic association The topic-content semantic relevance.

[0063] Then, after calculating the subject-content semantic association between the semantic features of each sentence in the project text description segment and the first project feature type, it is necessary to weight and optimize the semantics of each sentence based on this subject-content semantic association, so as to highlight the sentence semantic information related to the specific project feature in the project text description segment, which is conducive to the subsequent generation of a more accurate project feature text description that conforms to the first project feature type. Based on this, in the technical solution of the present application, the sequence of the subject-content semantic association is further normalized based on the maximum value to obtain a sequence of subject-content semantic association weight coefficients, and the sequence of subject-content semantic association weight coefficients is used as a weight to weight the sequence of the project text description sentence semantic encoding vectors to obtain a sequence of optimized project text description sentence semantic encoding vectors. It should be understood that through the normalization based on the maximum value, each subject-content semantic association in the sequence of subject-content semantic association can be mapped to a range between 0 and 1, eliminating the absolute value difference of different subject-content semantic associations, which is conducive to the subsequent weighted optimization operation. By using the normalized topic-content semantic association weight coefficient as the weight for weighted processing, the importance of the sentence in the entire document can be adjusted according to the association between the sentence and the specific feature type, highlighting the sentences that are highly related to the specific feature type and reducing the semantic influence of the sentences that are less related to the specific feature type. In this way, it can ensure that the text semantic features in the extracted project text description segment are more matched with the first project feature type, which is conducive to generating a more accurate project feature text description that conforms to the first project feature type.

[0064] In one embodiment of the present disclosure, after the sequence of subject-content semantic associations is normalized, the sequence of project text description sentence semantic coding vectors is optimized for important content enhancement using the sequence of normalized subject-content semantic associations as weighted weights to obtain a sequence of optimized project text description sentence semantic coding vectors, including: performing maximum value-based normalization on the sequence of subject-content semantic associations to obtain a sequence of subject-content semantic association weighting coefficients; using the sequence of subject-content semantic association weighting coefficients as weights, weighting the sequence of project text description sentence semantic coding vectors to obtain the sequence of optimized project text description sentence semantic coding vectors.

[0065] Then, the sequence of the semantic coding vectors of the optimized project text description sentences is passed through a decoder-based project feature generator to obtain a project feature text description. That is to say, the semantic features of the project text description sentences after weighted optimization are used for decoding regression to generate a project feature text description that conforms to the first project feature type. In this way, artificial intelligence technology can be used to quickly understand the content of the project document and extract important project features therein, such as risk, cost, timeline and other information, so as to intelligently generate a project feature text description according to the actual required project feature type, and provide auxiliary support for the project team to make decisions and plans.

[0066] In one embodiment of the present disclosure, the sequence of semantic coding vectors of the optimized project text description sentences is passed through a decoder-based project feature generator to obtain a project feature text description, including: cascading the sequence of semantic coding vectors of the optimized project text description sentences into an optimized project text description sentence semantic joint coding vector; performing feature correction on the optimized project text description sentence semantic joint coding vector to obtain a corrected optimized project text description sentence semantic joint coding vector; and passing the corrected optimized project text description sentence semantic joint coding vector through the decoder-based project feature generator to obtain the project feature text description.

[0067] Further, the corrected optimized project text description sentence semantic joint encoding vector is passed through the decoder-based project feature generator to obtain the project feature text description, including: using the decoder-based project feature generator to perform decoding regression on the corrected optimized project text description sentence semantic joint encoding vector using the following decoding formula to obtain the project feature text description; wherein the decoding formula is: ,in, The semantic joint encoding vector representing the corrected optimized item text description sentence, Represents a textual description of the project characteristics, represents the weight matrix, represents the bias vector, Represents matrix multiplication.

[0068] Preferably, the process of passing the sequence of the optimized semantic encoding vectors of the project text description sentences through a decoder-based project feature generator to obtain a project feature text description comprises the following steps:

[0069] Cascading the sequence of semantic coding vectors of the optimized project text description sentences into a semantic joint coding vector of the optimized project text description sentences;

[0070] Calculate the self-mean representation matrix and the self-standard deviation representation matrix of the semantic joint encoding vector of the text description sentence of the optimization project, wherein the self-mean representation matrix is The value of position is the semantic joint encoding vector of the text description sentence of the optimization project. Position eigenvalue and The mean of the position eigenvalues, and the standard deviation represents the matrix The value of position is the semantic joint encoding vector of the text description sentence of the optimization project. Position eigenvalue and The standard deviation of the location eigenvalues;

[0071] Performing matrix multiplication of the transposed vector of the optimized project text description sentence semantic joint coding vector with the self-mean representation matrix to obtain a first intermediate vector and performing matrix multiplication of the self-standard deviation representation matrix with the optimized project text description sentence semantic joint coding vector to obtain a second intermediate vector, wherein the optimized project text description sentence semantic joint coding vector is a column vector;

[0072] calculating a point sum of the first intermediate vector and the transposed vector of the second intermediate vector to obtain a third intermediate vector;

[0073] Calculating the matrix product of the self-mean representation matrix and the self-standard deviation representation matrix, and performing matrix multiplication on the transposed vector of the semantic joint encoding vector of the optimized project text description sentence and the matrix product to obtain a fourth intermediate vector;

[0074] Calculating the point sum of the transposed vectors of the third intermediate vector and the fourth intermediate vector to obtain a corrected optimized item text description sentence semantic joint encoding vector; and

[0075] The corrected optimized project text description sentence semantic joint encoding vector is passed through a decoder-based project feature generator to obtain a project feature text description.

[0076] Here, taking into account the differences in the representation of the semantic features of the encoded text caused by the differences in the source texts of the project text description segments of the semantic coding vectors of each project text description sentence in the sequence of semantic coding vectors, after weighting with the sequence of topic-content semantic association weight coefficients as weights, the overall sequence of optimized project text description sentence semantic coding vectors will cause local overflow features due to the semantic-weight differentiated distribution of each optimized project text description sentence semantic coding vector, thereby affecting the accuracy of the sequence of optimized project text description sentence semantic coding vectors when decoded by a decoder-based decoder.

[0077] Based on this, the self-mean representation matrix and the self-standard deviation representation matrix of the group aggregation statistical evaluation of the local distribution of the feature value granularity of the sequence of the semantic coding vector of the optimized project text description sentence are used as the retrieval and response distribution enhancement of the sequence of the semantic coding vector of the optimized project text description sentence, and a reference-free distribution retrieval response framework based on the group aggregation feature distribution open domain of the sequence of the semantic coding vector of the optimized project text description sentence is constructed to avoid the distribution response redundancy caused by the local overflow characteristics of the sequence of the semantic coding vector of the optimized project text description sentence through response superposition, so as to realize the faithfulness constraint of the response self-aggregation statistics related to the corrected semantic joint coding vector of the optimized project text description sentence to the decoding target domain, and improve the semantic accuracy of the project feature text description obtained by the decoder-based project feature generator of the corrected semantic joint coding vector of the optimized project text description sentence. In this way, the project feature text description can be generated more accurately according to the actual required project feature type, providing auxiliary support for the project team to make decisions and plans.

[0078] In summary, the above solution uses a semantic encoder based on artificial intelligence and deep learning to automatically encode the project text description, so as to quickly understand the content of the project document and extract important project features, such as risk, cost, timeline and other information. In this way, the project feature text description can be automatically generated according to the actual required project feature type, providing auxiliary support for the project team to make decisions and plans.

[0079] Figure 2 FIG. 1 is a block diagram of a project auxiliary support system based on artificial intelligence according to an exemplary embodiment. Figure 2 As shown, the system 200 includes:

[0080] A project feature type extraction module 201 is used to extract a first project feature type from a project feature type database;

[0081] A one-hot encoding module 202, configured to perform one-hot encoding on the first project feature type to obtain a first project feature type one-hot encoding vector;

[0082] A text description segment acquisition module 203 is used to acquire a project text description segment;

[0083] A semantic coding module 204, configured to perform sentence-granular semantic coding on the project text description segment to obtain a sequence of semantic coding vectors of project text description sentences;

[0084] The topic-content semantic association calculation module 205 is used to calculate the topic-content semantic association between each item text description sentence semantic encoding vector in the sequence of item text description sentence semantic encoding vectors and the first item feature type one-hot encoding vector to obtain a sequence of topic-content semantic associations;

[0085] The important content enhancement optimization module 206 is used to normalize the sequence of subject-content semantic associations, and then use the normalized sequence of subject-content semantic associations as weighted weights to perform important content enhancement optimization on the sequence of semantic coding vectors of the project text description sentences to obtain a sequence of optimized semantic coding vectors of the project text description sentences;

[0086] The decoding module 207 is used to pass the sequence of the optimized project text description sentence semantic encoding vectors through a decoder-based project feature generator to obtain a project feature text description.

[0087] In one embodiment of the present disclosure, the semantic encoding module includes: a sentence processing unit, used to perform sentence processing on the project text description segment to obtain a sequence of project text description sentences; a sentence encoding unit, used to pass the sequence of project text description sentences through a sentence encoder to obtain a sequence of semantic encoding vectors of the project text description sentences.

[0088] In one embodiment of the present disclosure, the topic-content semantic association calculation module includes: a global representation coefficient calculation unit, which is used to calculate the product between the semantic encoding vector of the project text description sentence and the first trainable weight vector to obtain the project text description sentence semantic global representation coefficient; an understanding coefficient calculation unit, which is used to calculate the product between the first project feature type one-hot encoding vector and the second trainable weight vector to obtain the first project feature type understanding coefficient; an activation processing unit, which is used to add the project text description sentence semantic global representation coefficient and the one project feature type understanding coefficient and then perform activation processing to obtain the topic-content semantic association.

[0089] The automatic project feature annotation system based on semantic encoder is a system that uses natural language processing (NLP) technology to automatically identify and annotate key features in project documents. This system can help project managers and team members quickly understand the content of project documents and extract important project features such as risks, costs, timelines, etc. The following are the general steps to build such a system:

[0090] 1. Requirements analysis: Determine the target users of the system, usage scenarios, and the types of project features that need to be automatically annotated.

[0091] 2. Data collection: Collect a large amount of project documents, such as project plans, risk assessment reports, progress updates, etc., as training data.

[0092] 3. Data preprocessing: Clean the document and remove irrelevant content, such as headers, footers, pictures, etc. Split the document into sentences or paragraphs for subsequent processing.

[0093] 4. Feature extraction: Use NLP technology to extract project features from documents. This may include: Entity recognition (NER): Identify entities such as project names, people, dates, etc. in documents. Relationship extraction: Identify the relationship between entities, such as "Project A is under the responsibility of Zhang San".

[0094] 5. Semantic encoder training: Use pre-trained semantic encoders (such as BERT, RoBERTa, etc.) as the base model. Fine-tune it on the project document data to adapt to the specific project feature labeling task.

[0095] 6. Annotation model training: Use the extracted features as training data to train an annotation model, such as Conditional Random Fields (CRF) or a neural network. The task of the annotation model is to match the extracted features with predefined project feature types (such as risk, cost, timeline, etc.).

[0096] 7. System Integration: Integrate the semantic encoder and annotation model into a system that can receive new project documents and automatically perform feature annotation.

[0097] 8. User Interface Design: Design a user-friendly interface to display the annotation results and allow users to make manual corrections.

[0098] 9. System testing and evaluation: Test the system and evaluate its annotation accuracy and recall. Adjust and optimize the model based on the test results.

[0099] 10. Deployment and maintenance: Deploy the system to a production environment for user use. Regularly update the training data to adapt to new project features and language patterns.

[0100] 11. User feedback collection: Collect user feedback to understand the strengths and weaknesses of the system. Continuously improve the system based on the feedback.

[0101] Reference below Figure 3, which shows a schematic diagram of the structure of an electronic device 600 suitable for implementing the embodiment of the present disclosure. The terminal device in the embodiment of the present disclosure may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle terminals (such as vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 3 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0102] like Figure 3 As shown, the electronic device 600 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the electronic device 600 are also stored. The processing device 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0103] Typically, the following devices may be connected to the I / O interface 605: an input device 606 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 607 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 608 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 609. The communication device 609 may allow the electronic device 600 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 3 The electronic device 600 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0104] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 609, or installed from the storage device 608, or installed from the ROM 602. When the computer program is executed by the processing device 601, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0105] It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0106] In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0107] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0108] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including, but not limited to, object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0109] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0110] The modules involved in the embodiments described in the present disclosure may be implemented by software or hardware. The name of a module does not limit the module itself in some cases. For example, a test parameter acquisition module may also be described as a "module for acquiring device test parameters corresponding to a target device".

[0111] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0112] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0113] Figure 4 FIG. 1 is an application scenario diagram of an artificial intelligence-based project auxiliary support method according to an exemplary embodiment. Figure 4 As shown, in this application scenario, first, a first project feature type (for example, Figure 4 C1 as shown in ); get the item text description segment (for example, Figure 4 Then, the acquired first project feature type and project text description segment are input to a server deployed with an artificial intelligence-based project auxiliary support algorithm (for example, Figure 4 In S) as shown in , the server is capable of processing the first project feature type and the project text description segment based on an artificial intelligence project auxiliary support algorithm to obtain a project feature text description.

[0114] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

[0115] In addition, although each operation is described in a specific order, this should not be understood as requiring these operations to be performed in the specific order shown or in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.

[0116] Although the subject matter has been described in language specific to structural features and / or method logic actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims. Regarding the device in the above embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.

Claims

1. A project auxiliary support method based on artificial intelligence, characterized in that: include: extracting a first project feature type from a project feature type database; One-hot encoding the first project feature type to obtain a first project feature type one-hot encoding vector; Get the project text description segment; Performing sentence-granularity-based semantic coding on the project text description segment to obtain a sequence of project text description sentence semantic coding vectors; Calculating the topic-content semantic association between each item text description sentence semantic encoding vector in the sequence of item text description sentence semantic encoding vectors and the first item feature type one-hot encoding vector to obtain a sequence of topic-content semantic associations; After normalizing the sequence of subject-content semantic associations, the sequence of the normalized subject-content semantic associations is used as a weighted weight to perform important content enhancement optimization on the sequence of semantic coding vectors of the project text description sentences to obtain a sequence of optimized semantic coding vectors of the project text description sentences; Passing the sequence of the optimized project text description sentence semantic encoding vectors through a decoder-based project feature generator to obtain a project feature text description; The step of calculating the topic-content semantic association between each item text description sentence semantic encoding vector in the sequence of item text description sentence semantic encoding vectors and the first item feature type one-hot encoding vector to obtain a sequence of topic-content semantic associations includes: Calculating the product between the semantic encoding vector of the project text description sentence and the first trainable weight vector to obtain a semantic global representation coefficient of the project text description sentence; Calculate the product of the first project feature type one-hot encoding vector and the second trainable weight vector to obtain a first project feature type understanding coefficient; The semantic global representation coefficient of the item text description sentence and the item feature type understanding coefficient are added together and then activated to obtain the topic-content semantic association.

2. The project auxiliary support method based on artificial intelligence according to claim 1 is characterized in that: The project text description segment is semantically encoded based on sentence granularity to obtain a sequence of project text description sentence semantic encoding vectors, including: Sentence processing is performed on the project text description segment to obtain a sequence of project text description sentences; The sequence of the item text description sentences is passed through a sentence encoder to obtain a sequence of semantic encoding vectors of the item text description sentences.

3. The project auxiliary support method based on artificial intelligence according to claim 2 is characterized in that: The semantic global representation coefficient of the project text description sentence and the understanding coefficient of the project feature type are added and then activated to obtain the topic-content semantic association, including: adding the semantic global representation coefficient of the project text description sentence and the understanding coefficient of the project feature type and then inputting An activation process is performed in the activation function to obtain the topic-content semantic association.

4. The project auxiliary support method based on artificial intelligence according to claim 3 is characterized in that: After normalizing the sequence of subject-content semantic associations, the sequence of the normalized subject-content semantic associations is used as a weighted weight to perform important content enhancement optimization on the sequence of semantic coding vectors of the project text description sentences to obtain a sequence of optimized semantic coding vectors of the project text description sentences, including: Performing a maximum-based normalization process on the sequence of subject-content semantic associations to obtain a sequence of subject-content semantic association weight coefficients; The sequence of the semantic coding vectors of the item text description sentences is weighted by using the sequence of the topic-content semantic association weight coefficients as weights to obtain the sequence of the semantic coding vectors of the optimized item text description sentences.

5. The project auxiliary support method based on artificial intelligence according to claim 4 is characterized in that: The sequence of the optimized semantic encoding vectors of the project text description sentences is passed through a decoder-based project feature generator to obtain a project feature text description, including: Cascading the sequence of semantic coding vectors of the optimized project text description sentences into a semantic joint coding vector of the optimized project text description sentences; Performing feature correction on the semantic joint coding vector of the optimized project text description sentence to obtain a corrected semantic joint coding vector of the optimized project text description sentence; The corrected optimized project text description sentence semantic joint encoding vector is passed through the decoder-based project feature generator to obtain the project feature text description.

6. The project auxiliary support method based on artificial intelligence according to claim 5 is characterized in that: The method of passing the corrected optimized project text description sentence semantic joint encoding vector through the decoder-based project feature generator to obtain the project feature text description comprises: using the decoder-based project feature generator to perform decoding regression on the corrected optimized project text description sentence semantic joint encoding vector using the following decoding formula to obtain the project feature text description; Wherein, the decoding formula is: ,in, The semantic joint encoding vector representing the corrected optimized item text description sentence, Represents a textual description of the project characteristics, represents the weight matrix, represents the bias vector, Represents matrix multiplication.

7. A project auxiliary support system based on artificial intelligence, characterized in that: include: A project feature type extraction module, used to extract a first project feature type from a project feature type database; A one-hot encoding module, configured to perform one-hot encoding on the first project feature type to obtain a one-hot encoding vector of the first project feature type; A text description segment acquisition module is used to acquire a project text description segment; A semantic encoding module, used for performing sentence-granular semantic encoding on the project text description segment to obtain a sequence of semantic encoding vectors of project text description sentences; A topic-content semantic association calculation module, used to calculate the topic-content semantic association between each project text description sentence semantic encoding vector in the sequence of project text description sentence semantic encoding vectors and the first project feature type one-hot encoding vector to obtain a sequence of topic-content semantic associations; An important content enhancement optimization module is used to normalize the sequence of subject-content semantic associations, and then use the normalized sequence of subject-content semantic associations as weighted weights to perform important content enhancement optimization on the sequence of semantic coding vectors of the project text description sentences to obtain a sequence of optimized semantic coding vectors of the project text description sentences; A decoding module, used for passing the sequence of the semantic encoding vectors of the optimized project text description sentences through a decoder-based project feature generator to obtain a project feature text description; Wherein, the topic-content semantic association calculation module includes: A global representation coefficient calculation unit, used for calculating the product between the semantic encoding vector of the project text description sentence and the first trainable weight vector to obtain the semantic global representation coefficient of the project text description sentence; a comprehension coefficient calculation unit, configured to calculate the product of the first project feature type one-hot encoding vector and the second trainable weight vector to obtain a first project feature type comprehension coefficient; The activation processing unit is used to add the semantic global representation coefficient of the project text description sentence and the understanding coefficient of the project feature type and then perform activation processing to obtain the topic-content semantic association.

8. The artificial intelligence-based project auxiliary support system according to claim 7 is characterized in that: The semantic encoding module comprises: A sentence processing unit, used for performing sentence processing on the project text description segment to obtain a sequence of project text description sentences; The sentence encoding unit is used to pass the sequence of the project text description sentences through a sentence encoder to obtain a sequence of semantic encoding vectors of the project text description sentences.

Images