Intelligent remote management system and method for tumor patient after immunotherapy

Abstract

The invention relates to the field of intelligent management, and particularly discloses an intelligent remote management system and method for a tumor patient after immunotherapy, and the method comprises the steps: obtaining rehabilitation knowledge questions inputted by a patient object, and extracting alternative rehabilitation knowledge from a database; the rehabilitation knowledge question semantic coding and the word granularity semantic analysis of the alternative rehabilitation knowledge are respectively carried out by adopting a text analysis and processing technology based on artificial intelligence; and intelligently obtaining an analysis result of whether to return the alternative rehabilitation knowledge as a retrieval result or not based on the semantic interaction matching characteristics of the rehabilitation knowledge questions and the alternative rehabilitation knowledge. By means of the mode, the system can more accurately understand the problem of the patient and retrieve alternative rehabilitation knowledge most related to the problem, meanwhile, the patient can obtain information at any time, and the rehabilitation compliance and satisfaction degree of the patient are greatly improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent management, and more specifically, to an intelligent remote management system and method for tumor patients after immunotherapy. Background Art

[0002] In recent years, tumor immunotherapy has made significant progress and has become the fourth major tumor treatment method after surgery, chemotherapy, and radiotherapy. Immunotherapy is different from the response pattern of traditional treatments and requires specific biomarkers and monitoring methods to evaluate the treatment effect. Through effective follow-up management, the effect of immunotherapy can be enhanced and patients can achieve better clinical responses.

[0003] However, traditional post-immunotherapy management of cancer patients relies on regular hospital follow-up or telephone consultations, which is not timely and convenient, so that patients cannot obtain the information they need at any time. In addition, under the traditional management model, patients may have to wait for a long time to get professional answers after asking questions, thus delaying the treatment of changes in the patient's condition.

[0004] Therefore, an intelligent remote management system for cancer patients after immunotherapy is desired. Summary of the invention

[0005] In order to solve the above technical problems, this application is proposed.

[0006] According to one aspect of the present application, an intelligent remote management system for tumor patients after immunotherapy is provided, which includes: A rehabilitation knowledge question acquisition module is used to acquire rehabilitation knowledge questions input by a patient object; A rehabilitation knowledge question semantic coding module, used for semantically coding the rehabilitation knowledge question to obtain a rehabilitation knowledge question semantic coding vector; An alternative rehabilitation knowledge extraction module is used to extract alternative rehabilitation knowledge from a database; A granular semantic coding module for candidate rehabilitation knowledge words, used for semantic coding after word segmentation processing of the candidate rehabilitation knowledge to obtain a sequence of granular semantic coding vectors of the candidate rehabilitation knowledge words; A semantic enhancement module, used for inputting the sequence of candidate rehabilitation knowledge word granular semantic encoding vectors into a semantic enhancer based on semantic relevance and semantic distance to obtain a sequence of enhanced candidate rehabilitation knowledge word granular semantic encoding vectors; A semantic interaction module, used for inputting the sequence of the rehabilitation knowledge question semantic encoding vector and the enhanced candidate rehabilitation knowledge word granularity semantic encoding vector into a one-way transmission semantic interaction module to obtain a rehabilitation knowledge-rehabilitation question one-way interaction matching representation vector as a rehabilitation knowledge-rehabilitation question one-way interaction matching representation feature; The analysis result generating module is used to obtain the analysis result based on the one-way interactive matching representation feature of the rehabilitation knowledge-rehabilitation question.

[0007] In the above-mentioned intelligent remote management system for tumor patients after immunotherapy, the alternative rehabilitation knowledge word granularity semantic encoding module is used to: perform word segmentation processing on the alternative rehabilitation knowledge and then pass it through a semantic encoder including a word embedding layer to obtain a sequence of the alternative rehabilitation knowledge word granularity semantic encoding vectors.

[0008] In the above-mentioned intelligent remote management system for tumor patients after immunotherapy, the semantic enhancement module includes: a predetermined feature vector extraction unit, which is used to extract a predetermined alternative rehabilitation knowledge word granular semantic coding vector from the sequence of the alternative rehabilitation knowledge word granular semantic coding vector; a semantic association calculation unit, which is used to calculate the semantic association between the predetermined alternative rehabilitation knowledge word granular semantic coding vector and all other alternative rehabilitation knowledge word granular semantic coding vectors to obtain a sequence of alternative rehabilitation knowledge word granular semantic associations; a global mean calculation unit, which is used to calculate the global mean of the sequence of alternative rehabilitation knowledge word granular semantic associations to obtain the global average association of alternative rehabilitation knowledge word granular semantics; a feature weighted optimization unit, which is used to use the global average association of the alternative rehabilitation knowledge word granular semantics as a weight to perform weighted optimization on the predetermined alternative rehabilitation knowledge word granular semantic coding vector to obtain an enhanced alternative rehabilitation knowledge word granular semantic coding vector corresponding to the predetermined alternative rehabilitation knowledge word granular semantic coding vector.

[0009] In the above-mentioned intelligent remote management system for tumor patients after immunotherapy, the semantic association calculation unit is used to: calculate the semantic similarity between the predetermined alternative rehabilitation knowledge word granularity semantic coding vector and the semantic coding vectors of all other alternative rehabilitation knowledge word granularity to obtain a sequence of alternative rehabilitation knowledge word granularity semantic similarities; calculate the number of feature vectors between the predetermined alternative rehabilitation knowledge word granularity semantic coding vector and the semantic coding vectors of all other alternative rehabilitation knowledge word granularity to obtain a sequence of alternative rehabilitation knowledge word granularity distance meter numbers; divide the sequence of alternative rehabilitation knowledge word granularity semantic similarities by the sequence of alternative rehabilitation knowledge word granularity distance meter numbers by position to obtain a sequence of alternative rehabilitation knowledge word granularity semantic associations. In the above-mentioned intelligent remote management system for tumor patients after immunotherapy, the semantic interaction module includes: a question-alternative semantic association calculation unit, which is used to calculate the semantic association between the rehabilitation knowledge question semantic coding vector and each enhanced alternative rehabilitation knowledge word granular semantic coding vector in the sequence of enhanced alternative rehabilitation knowledge word granular semantic coding vectors to obtain a sequence of question-alternative semantic associations; and a one-way interactive matching unit, which is used to use the sequence of question-alternative semantic associations as a weight sequence to calculate the vector weighted sum of the sequence of enhanced alternative rehabilitation knowledge word granular semantic coding vectors to obtain the rehabilitation knowledge-rehabilitation question one-way interactive matching representation vector.

[0010] In the above-mentioned intelligent remote management system for tumor patients after immunotherapy, the question-alternative semantic association calculation unit is used to: calculate the position-by-position difference between the rehabilitation knowledge question semantic coding vector and each of the enhanced alternative rehabilitation knowledge word granular semantic coding vectors to obtain a sequence of question-alternative semantic difference coding vectors; calculate the norm of each question-alternative semantic difference coding vector in the sequence of question-alternative semantic difference coding vectors to obtain a sequence of question-alternative semantic associations. In the above-mentioned intelligent remote management system for tumor patients after immunotherapy, the analysis result generation module is used to: pass the rehabilitation knowledge-rehabilitation question one-way interactive matching representation vector through a classifier-based matching analyzer to obtain an analysis result, and the analysis result is used to indicate whether to return the alternative rehabilitation knowledge as a retrieval result.

[0011] According to another aspect of the present application, a method for intelligent remote management of tumor patients after immunotherapy is provided, comprising: Obtaining rehabilitation knowledge questions input by patient subjects; Performing semantic encoding on the rehabilitation knowledge question to obtain a rehabilitation knowledge question semantic encoding vector; Extracting alternative rehabilitation knowledge from the database; Performing word segmentation processing on the candidate rehabilitation knowledge and then performing semantic coding to obtain a sequence of granular semantic coding vectors of the candidate rehabilitation knowledge words; Inputting the sequence of candidate rehabilitation knowledge word granular semantic encoding vectors into a semantic enhancer based on semantic relevance and semantic distance to obtain a sequence of enhanced candidate rehabilitation knowledge word granular semantic encoding vectors; Inputting the sequence of the rehabilitation knowledge question semantic encoding vector and the enhanced candidate rehabilitation knowledge word granular semantic encoding vector into a one-way transmission semantic interaction module to obtain a rehabilitation knowledge-rehabilitation question one-way interaction matching representation vector as a rehabilitation knowledge-rehabilitation question one-way interaction matching representation feature; Based on the one-way interactive matching representation feature of rehabilitation knowledge-rehabilitation question, an analysis result is obtained.

[0012] Compared with the prior art, the present application provides an intelligent remote management system and method for tumor patients after immunotherapy, which obtains rehabilitation knowledge questions input by the patient object and extracts alternative rehabilitation knowledge from the database, and uses artificial intelligence-based text analysis and processing technology to perform semantic encoding of the rehabilitation knowledge questions and word-granular semantic analysis of the alternative rehabilitation knowledge, respectively, so as to intelligently obtain the analysis result of whether to return the alternative rehabilitation knowledge as the retrieval result based on the semantic interactive matching characteristics of the rehabilitation knowledge questions and the alternative rehabilitation knowledge. In this way, the system can understand the patient's problem more accurately and retrieve the alternative rehabilitation knowledge most relevant to the question. At the same time, the patient can obtain information at any time, which greatly improves the patient's rehabilitation compliance and satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0014] Figure 1 This is a block diagram of an intelligent remote management system for tumor patients after immunotherapy according to an embodiment of the present application.

[0015] Figure 2 Schematic diagram of the architecture of an intelligent remote management system for tumor patients after immunotherapy according to an embodiment of the present application.

[0016] Figure 3 It is a block diagram of the semantic enhancement module in the intelligent remote management system for tumor patients after immunotherapy according to an embodiment of the present application.

[0017] Figure 4 It is a block diagram of the semantic interaction module in the intelligent remote management system for tumor patients after immunotherapy according to an embodiment of the present application.

[0018] Figure 5 This is a flow chart of a method for intelligent remote management of tumor patients after immunotherapy according to an embodiment of the present application. DETAILED DESCRIPTION

[0019] The 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, but rather these embodiments are 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.

[0020] 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.

[0021] In the description of the embodiments of the present disclosure, the term "including" and similar terms should be understood as open inclusion, that is, "including but not limited to". The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc. may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0022] 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".

[0023] In recent years, tumor immunotherapy has made significant progress and has become the fourth major tumor treatment method after surgery, chemotherapy, and radiotherapy. Immunotherapy is different from the response pattern of traditional treatments and requires specific biomarkers and monitoring methods to evaluate the treatment effect. Through effective follow-up management, the effect of immunotherapy can be enhanced and patients can achieve better clinical responses.

[0024] However, traditional post-immunotherapy management of cancer patients relies on regular hospital follow-up or telephone consultations, which is not timely and convenient, so that patients cannot obtain the information they need at any time. In addition, under the traditional management model, patients may have to wait for a long time to get professional answers after asking questions, thus delaying the treatment of changes in the patient's condition.

[0025] Therefore, in response to the above technical problems, the technical concept of this application is to obtain rehabilitation knowledge questions input by patient subjects and extract alternative rehabilitation knowledge from the database, and use artificial intelligence-based text analysis and processing technology to perform semantic encoding of rehabilitation knowledge questions and word-granular semantic analysis of alternative rehabilitation knowledge, respectively, so as to intelligently obtain the analysis result of whether to return the alternative rehabilitation knowledge as a retrieval result based on the semantic interactive matching features of the rehabilitation knowledge questions and the alternative rehabilitation knowledge. In this way, the system can understand the patient's questions more accurately and retrieve the alternative rehabilitation knowledge most relevant to the questions. At the same time, the patient can obtain information at any time, which greatly improves the patient's rehabilitation compliance and satisfaction.

[0026] Figure 1 This is a block diagram of an intelligent remote management system for tumor patients after immunotherapy according to an embodiment of the present application. Figure 2 FIG. 1 is a schematic diagram of the architecture of an intelligent remote management system for tumor patients after immunotherapy according to an embodiment of the present application. Figure 1 and Figure 2 As shown, according to the embodiment of the present application, the intelligent remote management system 100 for tumor patients after immunotherapy includes: a rehabilitation knowledge question acquisition module 110, which is used to acquire rehabilitation knowledge questions input by the patient object; a rehabilitation knowledge question semantic encoding module 120, which is used to semantically encode the rehabilitation knowledge questions to obtain a rehabilitation knowledge question semantic encoding vector; an alternative rehabilitation knowledge extraction module 130, which is used to extract alternative rehabilitation knowledge from a database; an alternative rehabilitation knowledge word granularity semantic encoding module 140, which is used to semantically encode the alternative rehabilitation knowledge after word segmentation processing to obtain a sequence of alternative rehabilitation knowledge word granularity semantic encoding vectors; a semantic enhancement module 150 , which is used to input the sequence of the alternative rehabilitation knowledge word granular semantic coding vectors into a semantic enhancer based on semantic relevance and semantic distance to obtain a sequence of reinforced alternative rehabilitation knowledge word granular semantic coding vectors; a semantic interaction module 160, which is used to input the sequence of the rehabilitation knowledge question semantic coding vectors and the reinforced alternative rehabilitation knowledge word granular semantic coding vectors into a one-way transmission semantic interaction module to obtain a rehabilitation knowledge-rehabilitation question one-way interaction matching representation vector as a rehabilitation knowledge-rehabilitation question one-way interaction matching representation feature; an analysis result generation module 170, which is used to obtain an analysis result based on the rehabilitation knowledge-rehabilitation question one-way interaction matching representation feature.

[0027] In the embodiment of the present application, the rehabilitation knowledge question acquisition module 110 is used to acquire rehabilitation knowledge questions input by the patient object. It should be understood that considering that the patient may have questions during the rehabilitation process, they may involve life habits, diet, exercise, drug management, side effect management, etc. after treatment. These questions reflect the patient's concern about his or her own health status and the need for more information. Based on this, in the technical solution of the present application, acquiring the rehabilitation knowledge questions input by the patient object and performing semantic analysis and understanding on them can help improve the patient's self-management ability.

[0028] In an embodiment of the present application, the rehabilitation knowledge question semantic encoding module 120 is used to semantically encode the rehabilitation knowledge question to obtain a rehabilitation knowledge question semantic encoding vector. Accordingly, it is considered that the rehabilitation knowledge question contains semantic information about the patient object's question about rehabilitation knowledge, and there is a contextual association relationship between the various semantic information. Therefore, in order to understand and capture the semantic information implied in the rehabilitation knowledge question, so as to perform more accurate matching and comparison later, in the technical solution of the present application, the rehabilitation knowledge question is semantically encoded to obtain a rehabilitation knowledge question semantic encoding vector, which can identify and extract the key concepts and relationships in the rehabilitation knowledge question, so that the model can understand the intention and information needs of the question.

[0029] In an embodiment of the present application, the alternative rehabilitation knowledge extraction module 130 is used to extract alternative rehabilitation knowledge from a database. It should be understood that, considering that the alternative rehabilitation knowledge is information and suggestions related to patient questions stored in the database, it may include medical guidelines, expert advice, rehabilitation cases, nutritional guidance, etc., and can provide accurate alternative rehabilitation knowledge, which can help patients better understand their conditions and guide them to perform correct rehabilitation activities. Therefore, in the technical solution of the present application, extracting alternative rehabilitation knowledge from the database can help patients learn how to self-manage various problems in the rehabilitation process, thereby improving their quality of life.

[0030] In an embodiment of the present application, the alternative rehabilitation knowledge word granularity semantic encoding module 140 is used to perform semantic encoding on the alternative rehabilitation knowledge after word segmentation to obtain a sequence of alternative rehabilitation knowledge word granularity semantic encoding vectors. Specifically, in an embodiment of the present application, the alternative rehabilitation knowledge word granularity semantic encoding module 140 is used to: perform word segmentation on the alternative rehabilitation knowledge and then pass it through a semantic encoder including a word embedding layer to obtain a sequence of alternative rehabilitation knowledge word granularity semantic encoding vectors. Accordingly, considering that the alternative rehabilitation knowledge contains a large amount of text information about rehabilitation knowledge, such as daily life management and nursing recommendations, treatment plans, drug use, side effect management, etc. for tumor patients, and these text information are composed of multiple words or phrases. Therefore, in order to analyze the semantic information of each word in the alternative rehabilitation knowledge and the correlation influence between them in a more fine-grained manner, in the technical solution of the present application, the alternative rehabilitation knowledge is segmented to obtain a sequence composed of multiple alternative rehabilitation knowledge words. Furthermore, considering the semantic interdependence and influence between the alternative rehabilitation knowledge words, in the technical solution of the present application, the alternative rehabilitation knowledge after word segmentation is passed through a semantic encoder including a word embedding layer to capture and mine the contextual semantic relationship and semantic dependency between each word in the alternative knowledge, thereby obtaining a sequence of granular semantic coding vectors of the alternative rehabilitation knowledge words.

[0031] In an embodiment of the present application, the semantic enhancement module 150 is used to input the sequence of the alternative rehabilitation knowledge word granular semantic coding vectors into a semantic enhancer based on the use of semantic relevance and semantic distance to obtain a sequence of enhanced alternative rehabilitation knowledge word granular semantic coding vectors. It should be understood that, considering that each alternative rehabilitation knowledge word granular semantic coding vector in the sequence of the alternative rehabilitation knowledge word granular semantic coding vectors has different alternative rehabilitation knowledge word granular semantic coding features in the semantics of the entire alternative rehabilitation knowledge, and that different similarities exist between the various alternative rehabilitation knowledge word granular semantic coding vectors at different semantic spans. Based on this, in order to perform feature analysis based on the overall semantics on each of the alternative rehabilitation knowledge word granular semantic coding vectors to more comprehensively consider the global semantic information of the alternative rehabilitation knowledge, in the technical solution of the present application, the sequence of the alternative rehabilitation knowledge word granular semantic coding vectors is input into a semantic enhancer based on the use of semantic relevance and semantic distance to obtain a sequence of enhanced alternative rehabilitation knowledge word granular semantic coding vectors. In particular, the semantic enhancer based on utilizing semantic relevance and semantic distance can utilize the relevance and semantic distance information about the granular semantics of the alternative rehabilitation knowledge words between the granular semantic encoding vectors of the alternative rehabilitation knowledge words to further enrich and enhance the semantic encoding representation of the alternative knowledge, which helps to better capture and extract the semantic connections and semantic hierarchies between the various words in the alternative rehabilitation knowledge, thereby constructing a more comprehensive and rich granular semantic representation of the alternative knowledge words.

[0032] Figure 3 is a block diagram of a semantic enhancement module in an intelligent remote management system for tumor patients after immunotherapy according to an embodiment of the present application. Specifically, in the embodiment of the present application, Figure 3 As shown, the semantic enhancement module 150 includes: a predetermined feature vector extraction unit 151, which is used to extract a predetermined alternative rehabilitation knowledge word granular semantic coding vector from the sequence of the alternative rehabilitation knowledge word granular semantic coding vectors; a semantic association calculation unit 152, which is used to calculate the semantic association between the predetermined alternative rehabilitation knowledge word granular semantic coding vector and all other alternative rehabilitation knowledge word granular semantic coding vectors to obtain a sequence of alternative rehabilitation knowledge word granular semantic associations; a global mean calculation unit 153, which is used to calculate the global mean of the sequence of alternative rehabilitation knowledge word granular semantic associations to obtain the global average association of alternative rehabilitation knowledge word granular semantics; a feature weighted optimization unit 154, which is used to perform weighted optimization on the predetermined alternative rehabilitation knowledge word granular semantic coding vector using the global average association of the alternative rehabilitation knowledge word granular semantics as a weight to obtain an enhanced alternative rehabilitation knowledge word granular semantic coding vector corresponding to the predetermined alternative rehabilitation knowledge word granular semantic coding vector.

[0033] More specifically, in the embodiment of the present application, the semantic association calculation unit 152 is used to: calculate the semantic similarity between the predetermined alternative rehabilitation knowledge word granularity semantic coding vector and all other alternative rehabilitation knowledge word granularity semantic coding vectors to obtain a sequence of alternative rehabilitation knowledge word granularity semantic similarities; calculate the number of feature vectors between the predetermined alternative rehabilitation knowledge word granularity semantic coding vector and all other alternative rehabilitation knowledge word granularity semantic coding vectors to obtain a sequence of alternative rehabilitation knowledge word granularity distance meter numbers; divide the sequence of alternative rehabilitation knowledge word granularity semantic similarities by the sequence of alternative rehabilitation knowledge word granularity distance meter numbers by position to obtain a sequence of alternative rehabilitation knowledge word granularity semantic associations.

[0034] In the embodiment of the present application, specifically, the semantic enhancement module 150 is used to: input the sequence of the candidate rehabilitation knowledge word granular semantic coding vectors into the semantic enhancer based on semantic relevance and semantic distance, and process it with the following enhancement formula to obtain the sequence of enhanced candidate rehabilitation knowledge word granular semantic coding vectors; wherein the enhancement formula is: ; in, and Respectively represent the first and Alternative rehabilitation knowledge word granular semantic encoding vectors, express and The number of eigenvectors between express and The similarity between The total number of feature vectors in the sequence of the candidate rehabilitation knowledge word granular semantic encoding vectors is minus one, and The values ​​of are equal, The first in the sequence of the granular semantic encoding vectors of the enhanced candidate rehabilitation knowledge words A granular semantic encoding vector of reinforced alternative rehabilitation knowledge words.

[0035] In an embodiment of the present application, the semantic interaction module 160 is used to input the sequence of the rehabilitation knowledge question semantic coding vector and the enhanced alternative rehabilitation knowledge word granularity semantic coding vector into the one-way transmission semantic interaction module to obtain the rehabilitation knowledge-rehabilitation question one-way interaction matching representation vector as the rehabilitation knowledge-rehabilitation question one-way interaction matching representation feature. Accordingly, considering that the rehabilitation knowledge question semantic coding vector expresses the intention and demand expressed in the rehabilitation knowledge question, the keywords and conceptual semantic information involved in the question. The sequence of the enhanced alternative rehabilitation knowledge word granularity semantic coding vector expresses the semantic representation of each alternative rehabilitation knowledge word in the alternative rehabilitation knowledge. Therefore, in order to compare the matching degree of the rehabilitation knowledge question semantic coding vector and the enhanced alternative rehabilitation knowledge word granularity semantic coding vector, and facilitate accurate recommendation, in the technical solution of the present application, the sequence of the rehabilitation knowledge question semantic coding vector and the enhanced alternative rehabilitation knowledge word granularity semantic coding vector are input into the one-way transmission semantic interaction module to obtain the rehabilitation knowledge-rehabilitation question one-way interaction matching representation vector. In detail, the one-way transmission semantic interaction module measures the degree of matching between the question semantic representation and the alternative knowledge semantic representation by calculating the semantic association between the rehabilitation knowledge question semantic coding vector and each reinforced alternative rehabilitation knowledge word granular semantic coding vector in the sequence of the reinforced alternative rehabilitation knowledge word granular semantic coding vectors, thereby realizing one-way semantic interaction between the question semantic representation and the alternative knowledge semantic representation, and then performs weighted summation on each reinforced alternative rehabilitation knowledge word granular semantic coding vector based on each semantic association, so as to aggregate the semantic information of each reinforced alternative rehabilitation knowledge word granular semantic coding vector to form a comprehensive semantic representation. This comprehensive semantic representation represents the overall semantic association between the alternative knowledge and the question semantic representation, thereby enhancing the semantic understanding and matching capabilities of rehabilitation knowledge.

[0036] Figure 4 is a block diagram of a semantic interaction module in an intelligent remote management system for tumor patients after immunotherapy according to an embodiment of the present application. Specifically, in the embodiment of the present application, Figure 4 As shown, the semantic interaction module 160 includes: a question-alternative semantic association calculation unit 161, which is used to calculate the semantic association between the rehabilitation knowledge question semantic coding vector and each reinforced alternative rehabilitation knowledge word granular semantic coding vector in the sequence of reinforced alternative rehabilitation knowledge word granular semantic coding vectors to obtain a sequence of question-alternative semantic associations; and a one-way interactive matching unit 162, which is used to use the sequence of question-alternative semantic associations as a weight sequence to calculate the vector weighted sum of the sequence of reinforced alternative rehabilitation knowledge word granular semantic coding vectors to obtain the rehabilitation knowledge-rehabilitation question one-way interactive matching representation vector.

[0037] More specifically, in an embodiment of the present application, the question-alternative semantic association calculation unit 161 is used to: calculate the position-by-position difference between the rehabilitation knowledge question semantic coding vector and each of the reinforced alternative rehabilitation knowledge word granular semantic coding vectors to obtain a sequence of question-alternative semantic difference coding vectors; calculate a norm of each question-alternative semantic difference coding vector in the sequence of question-alternative semantic difference coding vectors to obtain a sequence of question-alternative semantic associations.

[0038] In the embodiment of the present application, specifically, the question-alternative semantic association calculation unit 161 is used to: calculate the semantic association between the rehabilitation knowledge question semantic coding vector and each enhanced alternative rehabilitation knowledge word granular semantic coding vector in the sequence of enhanced alternative rehabilitation knowledge word granular semantic coding vectors using the following association formula to obtain the sequence of question-alternative semantic associations; wherein the association formula is: ; in, and are the rehabilitation knowledge question semantic encoding vector and each of the enhanced candidate rehabilitation knowledge word granular semantic encoding vectors, represents the one-norm of the eigenvector, The question-alternative semantic association degree of each position in the sequence of the question-alternative semantic association degree is represented.

[0039] In an embodiment of the present application, the analysis result generation module 170 is used to obtain an analysis result based on the one-way interactive matching representation feature of the rehabilitation knowledge-rehabilitation question. Specifically, in an embodiment of the present application, the analysis result generation module 170 is used to: pass the one-way interactive matching representation vector of the rehabilitation knowledge-rehabilitation question through a classifier-based matching analyzer to obtain an analysis result, and the analysis result is used to indicate whether to return the alternative rehabilitation knowledge as a retrieval result. That is, the one-way interactive matching representation feature of the rehabilitation knowledge-rehabilitation question obtained by voice interaction using the sequence of the semantic encoding vector of the rehabilitation knowledge question and the semantic encoding vector of the enhanced alternative rehabilitation knowledge word granularity is classified and processed, so as to intelligently obtain the analysis result of whether to return the alternative rehabilitation knowledge as a retrieval result. In this way, the system can understand the patient's problem more accurately and retrieve the alternative rehabilitation knowledge that is most relevant to the question. At the same time, the patient can obtain information at any time, which greatly improves the patient's rehabilitation compliance and satisfaction.

[0040] In a preferred embodiment, taking into account the misalignment of the source semantic distribution of the rehabilitation knowledge questions input by the patient subject and the alternative rehabilitation knowledge extracted from the database, as well as the semantic enhancement of the sequence of the alternative rehabilitation knowledge word granularity semantic encoding vectors using semantic relevance and semantic distance, the sequence of the rehabilitation knowledge question semantic encoding vector and the enhanced alternative rehabilitation knowledge word granularity semantic encoding vector will have text semantic feature correspondence, so that the rehabilitation knowledge-rehabilitation question one-way interaction matching representation vector obtained by the one-way transmission semantic interaction module will also have the text semantic distribution expression complexity relative to the source text semantic distribution, resulting in instance decision-making loss and affecting the accuracy of the classification results.

[0041] Based on this, the present application optimizes the one-way interactive matching representation vector of rehabilitation knowledge-rehabilitation question, including the following steps: Calculate the first and The weighted sum and weighted subtraction between the eigenvalues ​​are used to obtain the first rehabilitation knowledge-rehabilitation question one-way interactive matching linear heterogeneous matrix and the second rehabilitation knowledge-rehabilitation question one-way interactive matching linear heterogeneous matrix, namely: ; ; ; in, , , , They represent different weight hyperparameters. Here, it should be known that the weight hyperparameters can be selected by existing means, such as grid search, random search, Bayesian optimization, and respectively represent the first and Eigenvalues, represents the one-way interactive matching representation vector of rehabilitation knowledge-rehabilitation question, Represents the linear heterogeneous matrix of the first rehabilitation knowledge-rehabilitation question one-way interactive matching No. Eigenvalues, Represents the linear heterogeneous matrix of the one-way interaction matching between the second rehabilitation knowledge and rehabilitation questions No. Eigenvalue; The rehabilitation knowledge-rehabilitation question one-way interactive matching representation vector as a row vector is respectively matrix-multiplied with the first rehabilitation knowledge-rehabilitation question one-way interactive matching linear heterogeneous matrix and the second rehabilitation knowledge-rehabilitation question one-way interactive matching linear heterogeneous matrix to obtain the first rehabilitation knowledge-rehabilitation question one-way interactive matching information compensation vector One-way interactive matching information compensation vector with the second rehabilitation knowledge-rehabilitation question ,in, Represents matrix multiplication; The first rehabilitation knowledge-rehabilitation question one-way interactive matching information compensation vector and the second rehabilitation knowledge-rehabilitation question one-way interactive matching information compensation vector are respectively multiplied by the inverse of the F norm of the first rehabilitation knowledge-rehabilitation question one-way interactive matching linear heterogeneous matrix and the inverse of the F norm of the second rehabilitation knowledge-rehabilitation question one-way interactive matching linear heterogeneous matrix to obtain the first rehabilitation knowledge-rehabilitation question one-way interactive matching discriminant optimization vector One-way interactive matching discriminant optimization vector with the second rehabilitation knowledge-rehabilitation question ,in, represents dot product, represents the F norm of the matrix; The first rehabilitation knowledge-rehabilitation question one-way interactive matching discriminant optimization vector and the second rehabilitation knowledge-rehabilitation question one-way interactive matching discriminant optimization vector are subtracted to obtain an optimized rehabilitation knowledge-rehabilitation question one-way interactive matching representation vector , Indicates point addition.

[0042] The optimized rehabilitation knowledge-rehabilitation question one-way interactive matching representation vector is passed through a classifier-based matching analyzer to obtain an analysis result.

[0043] That is, by constructing a linear transformation optimization of the heterogeneous feature space representation of the rehabilitation knowledge-rehabilitation question one-way interactive matching representation vector, the multi-layer probability distribution architecture iterative optimization in the self-supervised isomorphic feature extraction process of the rehabilitation knowledge-rehabilitation question one-way interactive matching representation vector is realized, thereby introducing dynamic weight compensation to eliminate the influence of the asymmetric information loss term of the rehabilitation knowledge-rehabilitation question one-way interactive matching representation vector, thereby combining the discriminant optimization under low-rank constraints to significantly improve the classification discrimination efficiency of the feature importance weighted evaluation index of the rehabilitation knowledge-rehabilitation question one-way interactive matching representation vector in the heterogeneous feature space, that is, quantifying the contribution of the feature value to the decision boundary through the statistical significance of the feature value to improve the accuracy of the classification result. In this way, the system can understand the patient's problem more accurately and retrieve the alternative rehabilitation knowledge most relevant to the question. At the same time, the patient can obtain information at any time, which greatly improves the patient's rehabilitation compliance and satisfaction.

[0044] In summary, the intelligent remote management system 100 for tumor patients after immunotherapy based on the embodiment of the present application is explained, which obtains the rehabilitation knowledge questions input by the patient object and extracts the alternative rehabilitation knowledge from the database, and adopts the text analysis and processing technology based on artificial intelligence to respectively perform the semantic encoding of the rehabilitation knowledge questions and the word granularity semantic analysis of the alternative rehabilitation knowledge, so as to intelligently obtain the analysis result of whether to return the alternative rehabilitation knowledge as the retrieval result based on the semantic interaction matching characteristics of the rehabilitation knowledge questions and the alternative rehabilitation knowledge. In this way, the system can understand the patient's problem more accurately and retrieve the alternative rehabilitation knowledge that is most relevant to the question. At the same time, the patient can obtain information at any time, which greatly improves the patient's rehabilitation compliance and satisfaction.

[0045] As described above, the intelligent remote management system 100 for tumor patients after immunotherapy according to the embodiment of the present application can be implemented in various wireless terminals, such as a server with an intelligent remote management algorithm for tumor patients after immunotherapy. In one possible implementation, the intelligent remote management system 100 for tumor patients after immunotherapy according to the embodiment of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the intelligent remote management system 100 for tumor patients after immunotherapy can be a software module in the operating system of the wireless terminal, or it can be an application developed for the wireless terminal; of course, the intelligent remote management system 100 for tumor patients after immunotherapy can also be one of the many hardware modules of the wireless terminal.

[0046] Alternatively, in another example, the intelligent remote management system 100 for tumor patients after immunotherapy and the wireless terminal may also be separate devices, and the intelligent remote management system 100 for tumor patients after immunotherapy may be connected to the wireless terminal via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0047] Figure 5 Flow chart of the intelligent remote management method for tumor patients after immunotherapy according to an embodiment of the present application. Figure 5As shown, according to the embodiment of the present application, the intelligent remote management method for tumor patients after immunotherapy includes: S110, obtaining rehabilitation knowledge questions input by the patient object; S120, semantically encoding the rehabilitation knowledge questions to obtain rehabilitation knowledge question semantic encoding vectors; S130, extracting alternative rehabilitation knowledge from the database; S140, semantically encoding the alternative rehabilitation knowledge after word segmentation processing to obtain a sequence of alternative rehabilitation knowledge word granularity semantic encoding vectors; S150, inputting the sequence of alternative rehabilitation knowledge word granularity semantic encoding vectors into a semantic enhancer based on semantic relevance and semantic distance to obtain a sequence of enhanced alternative rehabilitation knowledge word granularity semantic encoding vectors; S160, inputting the rehabilitation knowledge question semantic encoding vector and the sequence of enhanced alternative rehabilitation knowledge word granularity semantic encoding vectors into a one-way transmission semantic interaction module to obtain a rehabilitation knowledge-rehabilitation question one-way interaction matching representation vector as a rehabilitation knowledge-rehabilitation question one-way interaction matching representation feature; S170, obtaining an analysis result based on the rehabilitation knowledge-rehabilitation question one-way interaction matching representation feature.

[0048] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned intelligent remote management method for tumor patients after immunotherapy have been referred to above. Figures 1 to 4 The description of the intelligent remote management system for tumor patients after immunotherapy has been introduced in detail, and therefore, its repeated description will be omitted.

[0049] The above descriptions of various implementations of the present disclosure are exemplary and not exhaustive. They are not limited to the disclosed implementations, and many modifications and changes are obvious to those of ordinary skill in the art without departing from the scope and spirit of the described implementations. The terms used herein are selected to best explain the principles of the implementations, practical applications, or improvements to the technology in the market, or to enable other ordinary skill in the art to understand the various practical methods disclosed herein.

Claims

1. An intelligent remote management system for tumor patients after immunotherapy, characterized in that: include: A rehabilitation knowledge question acquisition module is used to acquire rehabilitation knowledge questions input by a patient object; A rehabilitation knowledge question semantic encoding module, used for semantically encoding the rehabilitation knowledge question to obtain a rehabilitation knowledge question semantic encoding vector; An alternative rehabilitation knowledge extraction module is used to extract alternative rehabilitation knowledge from a database; A granular semantic coding module for candidate rehabilitation knowledge words, used for semantic coding after word segmentation processing of the candidate rehabilitation knowledge to obtain a sequence of granular semantic coding vectors of the candidate rehabilitation knowledge words; A semantic enhancement module, used for inputting the sequence of candidate rehabilitation knowledge word granular semantic encoding vectors into a semantic enhancer based on semantic relevance and semantic distance to obtain a sequence of enhanced candidate rehabilitation knowledge word granular semantic encoding vectors; A semantic interaction module, used for inputting the sequence of the rehabilitation knowledge question semantic encoding vector and the enhanced candidate rehabilitation knowledge word granularity semantic encoding vector into a one-way transmission semantic interaction module to obtain a rehabilitation knowledge-rehabilitation question one-way interaction matching representation vector as a rehabilitation knowledge-rehabilitation question one-way interaction matching representation feature; The analysis result generating module is used to obtain the analysis result based on the one-way interactive matching representation feature of the rehabilitation knowledge-rehabilitation question.

2. The intelligent remote management system for tumor patients after immunotherapy according to claim 1, characterized in that: The candidate rehabilitation knowledge word granularity semantic encoding module is used to: perform word segmentation processing on the candidate rehabilitation knowledge and then pass it through a semantic encoder including a word embedding layer to obtain a sequence of the candidate rehabilitation knowledge word granularity semantic encoding vectors.

3. The intelligent remote management system for tumor patients after immunotherapy according to claim 2, characterized in that: The semantic enhancement module comprises: A predetermined feature vector extraction unit, used for extracting a predetermined candidate rehabilitation knowledge word granular semantic coding vector from the sequence of the candidate rehabilitation knowledge word granular semantic coding vectors; A semantic association calculation unit, used for calculating the semantic association between the predetermined candidate rehabilitation knowledge word granularity semantic coding vector and all other candidate rehabilitation knowledge word granularity semantic coding vectors to obtain a sequence of candidate rehabilitation knowledge word granularity semantic associations; A global mean value calculation unit, used for calculating the global mean value of the sequence of the granular semantic association degree of the candidate rehabilitation knowledge words to obtain the global average granular semantic association degree of the candidate rehabilitation knowledge words; The feature weighted optimization unit is used to perform weighted optimization on the predetermined candidate rehabilitation knowledge word granular semantic coding vector using the global average relevance of the candidate rehabilitation knowledge word granular semantic coding vector as a weight to obtain a reinforced candidate rehabilitation knowledge word granular semantic coding vector corresponding to the predetermined candidate rehabilitation knowledge word granular semantic coding vector.

4. The intelligent remote management system for tumor patients after immunotherapy according to claim 3, characterized in that: The semantic relevance calculation unit is used to: Calculating the semantic similarity between the predetermined candidate rehabilitation knowledge word granularity semantic encoding vector and all other candidate rehabilitation knowledge word granularity semantic encoding vectors to obtain a sequence of candidate rehabilitation knowledge word granularity semantic similarities; Calculating the number of feature vectors between the predetermined candidate rehabilitation knowledge word granularity semantic coding vector and all other candidate rehabilitation knowledge word granularity semantic coding vectors to obtain a sequence of candidate rehabilitation knowledge word granularity distance meter numbers; The sequence of the candidate rehabilitation knowledge word granularity semantic similarities and the sequence of the candidate rehabilitation knowledge word granularity distance meter quantities are divided by position to obtain the sequence of the candidate rehabilitation knowledge word granularity semantic associations.

5. The intelligent remote management system for tumor patients after immunotherapy according to claim 4, characterized in that: The semantic interaction module comprises: A question-alternative semantic association calculation unit, used for calculating the semantic association between the rehabilitation knowledge question semantic coding vector and each enhanced alternative rehabilitation knowledge word granular semantic coding vector in the sequence of enhanced alternative rehabilitation knowledge word granular semantic coding vectors to obtain a sequence of question-alternative semantic associations; The one-way interactive matching unit is used to calculate the vector weighted sum of the sequence of the granular semantic encoding vectors of the enhanced alternative rehabilitation knowledge words using the sequence of the question-alternative semantic association as the weight sequence to obtain the one-way interactive matching representation vector of the rehabilitation knowledge-rehabilitation question.

6. The intelligent remote management system for tumor patients after immunotherapy according to claim 5, characterized in that: The question-alternative semantic association calculation unit is used to: Calculating the position-by-position difference between the rehabilitation knowledge question semantic encoding vector and each of the enhanced alternative rehabilitation knowledge word granular semantic encoding vectors to obtain a sequence of question-alternative semantic difference encoding vectors; A norm of each question-alternative semantic difference encoding vector in the sequence of question-alternative semantic difference encoding vectors is calculated to obtain the sequence of question-alternative semantic associations.

7. The intelligent remote management system for tumor patients after immunotherapy according to claim 6, characterized in that: The analysis result generating module is used to: pass the rehabilitation knowledge-rehabilitation question one-way interactive matching representation vector through a classifier-based matching analyzer to obtain an analysis result, and the analysis result is used to indicate whether to return the candidate rehabilitation knowledge as a retrieval result.

8. An intelligent remote management method for tumor patients after immunotherapy, characterized in that: include: Obtaining rehabilitation knowledge questions input by patient subjects; Performing semantic encoding on the rehabilitation knowledge question to obtain a rehabilitation knowledge question semantic encoding vector; Extracting alternative rehabilitation knowledge from the database; Performing word segmentation processing on the candidate rehabilitation knowledge and then performing semantic coding to obtain a sequence of granular semantic coding vectors of the candidate rehabilitation knowledge words; Inputting the sequence of candidate rehabilitation knowledge word granular semantic encoding vectors into a semantic enhancer based on semantic relevance and semantic distance to obtain a sequence of enhanced candidate rehabilitation knowledge word granular semantic encoding vectors; Inputting the sequence of the rehabilitation knowledge question semantic encoding vector and the enhanced candidate rehabilitation knowledge word granular semantic encoding vector into a one-way transmission semantic interaction module to obtain a rehabilitation knowledge-rehabilitation question one-way interaction matching representation vector as a rehabilitation knowledge-rehabilitation question one-way interaction matching representation feature; Based on the one-way interactive matching representation feature of rehabilitation knowledge-rehabilitation question, an analysis result is obtained.

Images