High-precision geographical name translation method integrating artificial intelligence and multi-language syllable segmentation

Abstract

The invention discloses a high-precision geographical name translation method integrating artificial intelligence and multi-language syllable segmentation, and relates to the technical field of geographical name translations, the method comprises the following steps: firstly splitting an input geographical name into a sub-word Token sequence, and using a context coding model of a word embedding model to realize semantic context associated coding of a to-be-translated geographical name; and then modeling internal structured information and a dependency relationship in context semantics of the token to be translated in a local semantic relevance reconstruction enhancement mode, identifying recessive syllable relevance in compound words and adhesive words, and enhancing local morphological characteristics near syllable boundaries, so as to solve the problems of irregular spelling and cross-language interference, and improve the translation efficiency of the token to be translated. And further decoding and outputting the geographical name character string after syllable segmentation to carry out geographical name translation, generating a corresponding geographical name target language translation text, and realizing end-to-end effective conversion from an original geographical name to a target translation name.

Description

Technical Field

[0001] The present application relates to the technical field of place name translation, and more specifically, to a high-precision place name translation method integrating artificial intelligence and multi-language syllable segmentation. Background Art

[0002] Place name translation is a key task in cross-language information processing, and its accuracy directly affects the application effects in the fields of geographic information systems, multilingual map navigation, cross-border logistics, etc. However, due to the significant differences in the syllable structure, morphological rules and writing systems of the world's languages, traditional translation methods face the following core challenges when facing multilingual place names: the syllable formation rules of different language families vary greatly. For example, Chinese is based on monosyllabic Chinese characters with clear syllable boundaries; English relies on complex vowel-consonant combinations (such as "strengths" contains 7 phonemes) and has a large number of irregular spellings; agglutinative languages ​​such as Finnish and Turkish form words through syllable superposition, with variable length and structure.

[0003] In addition, high-resource languages ​​(such as English and French) already have mature syllable segmentation rule libraries, but medium- and low-resource languages ​​(such as Vietnamese and Tibetan) lack annotated data and rely on statistical models with limited generalization capabilities. Moreover, in the context of globalization, place names often contain mixed language components (such as a mixture of English and Hindi in "New Delhi"), and a single translation strategy is difficult to accurately segment. These translation limitations have brought challenges to the understanding of place name translation and the application of geographic information.

[0004] Therefore, it is necessary to provide a high-precision place name translation method that integrates artificial intelligence and multilingual syllable segmentation to solve the above technical problems. 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, a high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation is provided, which includes: Receive a place name character string to be translated input by a user; Using the language identification model to detect the source language of the place name string to be translated; Determine the syllable segmentation strategy based on the source language of the place name string to be translated; In response to the syllable segmentation strategy being a syllable segmentation method based on a deep neural network model, based on the syllable segmentation strategy, syllable segmentation is performed on the place name character string to be translated to obtain a syllable segmented place name character string; The syllable-segmented place name character string is input into the place name translation model to obtain the place name target language translation text.

[0007] The present application has at least the following technical effects: Compared with the prior art, the present application provides a high-precision place name translation method that integrates artificial intelligence and multilingual syllable segmentation, which first splits the input place name into a subword Token sequence, and uses the context encoding model of the word embedding model to realize the semantic context association encoding of the place name to be translated; then, the internal structured information and dependency relationships in the context semantics of the place name Token to be translated are modeled by means of local semantic association reconstruction and reinforcement, the implicit syllable associations in compound words and agglutinative languages ​​are identified, and the local morphological features near the syllable boundaries (such as prefixes / suffixes, consonant clusters) are enhanced to solve the problems of irregular spelling and cross-language interference, and then the place name character string after output syllable segmentation is decoded to translate the place name, and the corresponding place name target language translation text is generated, thereby realizing end-to-end effective conversion from the original place name to the target translated name. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] 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.

[0009] Figure 1 A flowchart of a high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation according to an embodiment of the present application; Figure 2 A data flow diagram of a high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation according to an embodiment of the present application; Figure 3 This is a flowchart of sub-step S4 of the high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation according to an embodiment of the present application. DETAILED DESCRIPTION

[0010] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0011] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0012] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0013] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0014] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0015] It should be noted in advance that the acquisition and processing of all information or data in this application are carried out in compliance with the relevant national data protection laws and policies and with the authorization of the authority manager.

[0016] In order to solve the problems of large differences in syllable rules, unbalanced data resources, and mixed language interference in multilingual place name translation, a high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation is proposed in the technical solution of this application. This solution adopts the design concept of "hierarchical decision-making-semantic drive", dynamically adapts the syllable segmentation strategy of different language types, and optimizes the translation accuracy by combining contextual semantic understanding.

[0017] Based on this, in the technical solution of the present application, a high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation is proposed. Figure 1 The present invention is a flowchart of a high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation according to an embodiment of the present application. Figure 2 The data flow diagram of the high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation according to the embodiment of the present application is shown in FIG. Figure 1 and Figure 2As shown, according to the embodiment of the present application, the high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation includes the following steps: S1, receiving a place name character string to be translated input by a user; S2, using a language recognition model to detect the source language of the place name character string to be translated; S3, based on the source language of the place name character string to be translated, determining a syllable segmentation strategy; S4, in response to the syllable segmentation strategy being a syllable segmentation method based on a deep neural network model, based on the syllable segmentation strategy, syllable segmenting the place name character string to be translated to obtain a syllable-segmented place name character string; S5, inputting the syllable-segmented place name character string into a place name translation model to obtain a place name target language translation text.

[0018] In particular, S1 receives a place name string to be translated input by a user. The place name string to be translated is usually composed of a series of characters, representing the name of a specific geographical location, and these characters can be letters, numbers, symbols or a combination thereof. The place name string to be translated may contain a single language component or a mixed language component (for example, "New Delhi" combines English and Hindi elements). For the translation system, it is important to accurately identify the source language of the string and determine the subsequent processing steps based on this. Specifically, in this process, a user-friendly interface can be designed to allow users to input the place names they want to translate through various devices such as computers, mobile phones, etc. The interface can be a form on a web page, an input box in a mobile application, etc.

[0019] In particular, S2 uses a language recognition model to detect the source language of the place name string to be translated. The language recognition model is usually built based on deep learning technology and can automatically learn and recognize the characteristic patterns of different languages. When the user enters the place name string to be translated, the string will be sent to the pre-trained language recognition model. The model will comprehensively analyze the input character sequence and extract key information that can represent the characteristics of a specific language. For example, for some languages ​​with unique character sets (such as Chinese using Chinese characters and Russian using Cyrillic letters), the model can quickly judge based on the character type; while for those languages ​​that use the same character set but have significant differences in grammatical structure or vocabulary distribution (such as English and French), they will rely on more complex statistical features and contextual information to distinguish. The language recognition model can use advanced deep learning frameworks such as convolutional neural networks (CNN), recurrent neural networks (RNN) and their variants (such as long short-term memory networks LSTM, gated recurrent units GRU) or Transformer architectures. These models can effectively capture local and global features in the input string and map them to a probability distribution space representing the possibility of different languages ​​through multiple layers of nonlinear transformations. Ultimately, the model outputs a probability score for each known language, and the one with the highest score is considered the most likely source language for the input string.

[0020] In particular, the S3 determines the syllable segmentation strategy based on the source language of the place name string to be translated. That is, the characteristics and rules of the language are evaluated according to the source language of the place name string to be translated. For example, since high-resource languages ​​already have a mature syllable segmentation rule library, the established rules can be relatively directly applied to deal with the problem of syllable boundaries. Therefore, if a high-resource language such as English, French or Spanish is detected, a rule-based syllable segmentation method is usually selected; wherein the rules may include how to deal with special cases such as vowel-consonant combinations and irregular spellings, so as to ensure that the syllable structure within each word can be accurately identified; for those medium and low-resource languages, such as Chinese, Japanese, Korean and Vietnamese, there may be a lack of sufficient annotation data to support rule-based methods. In this case, the system tends to adopt a syllable segmentation strategy based on a statistical model; for example, it can use conditional random fields (CRF), hidden Markov models (HMM) or other models suitable for sequence labeling tasks to automatically discover and apply syllable segmentation rules applicable to specific languages; when encountering place names with mixed language components, such as "New Delhi" that contains both English and Hindi elements, a single language processing strategy is often difficult to meet the needs. At this time, the system will choose a syllable segmentation method based on a deep neural network model to dynamically adapt to the interaction between different languages ​​and capture more complex language phenomena through a deep architecture. Specifically, deep neural networks can learn the commonalities and differences between multiple languages ​​through training, so that they can still maintain high accuracy in the face of cross-language interference. For example, when processing compound words or agglutinative morphemes, deep neural networks can identify implicit syllable associations between roots and suffixes, and enhance local morphological features near syllable boundaries, effectively solving irregular spelling and cross-language interference problems.

[0021] In particular, in S4, in response to the syllable segmentation strategy being a syllable segmentation method based on a deep neural network model, the place name string to be translated is segmented into syllables based on the syllable segmentation strategy to obtain a syllable segmented place name string. In a specific example of the present application, Figure 3 As shown, S4 includes: S41, performing contextual semantic association encoding on the place name character string to be translated based on the place name Token unit to obtain the semantic contextual association encoding representation of the place name to be translated; S42, performing local semantic association reconstruction and enhancement on the semantic contextual association encoding representation of the place name to be translated to obtain the semantic contextual association enhanced encoding representation of the place name to be translated; S43, determining the place name character string after syllable segmentation based on the semantic contextual association enhanced encoding representation of the place name to be translated.

[0022] Specifically, the S41 performs contextual semantic association encoding based on place name Token units on the place name character string to be translated to obtain a semantic contextual association encoding representation of the place name to be translated. That is, in an embodiment of the present application, first, the place name character string to be translated is segmented to obtain a sequence distribution of place name Token units to be translated. It should be understood that when the input place name character stream contains a mixed writing system (such as the coexistence of Latin letters, tone symbols or agglutinative morphemes), the continuity and unstructured characteristics of the original character string will hinder the model's recognition of language attribution and internal rules. For example, when faced with a long character string composed of agglutinative morphemes, which may contain multiple combinations of roots and suffixes, the traditional coarse-grained segmentation based on spaces cannot capture implicit syllable boundaries (such as consonant clusters or vowel linking rules). At this point, the multilingual subword segmentation algorithm deconstructs the character sequence into token units with independent semantic or phonological meanings (such as breaking down agglutinative morphemes into "root + suffix" combinations), which can explicitly reveal the word-forming logic within the language, such as separating core semantic units from grammatical markers in compound words, or correctly handling the visual adhesion problem of conjoined characters in non-Latin scripts. By generating token sequence distributions, the model can transform heterogeneous input character streams into a set of units with discrete semantic boundaries. For example, in languages ​​containing tone marks, tone marks are combined with basic letters into independent tokens to preserve phonological features (such as binding tone marks "́" to vowel letters), or in spellings with dense consonant clusters, subunits that conform to the phoneme rules of the target language are segmented (such as breaking down consonant clusters into legal syllable combinations). This discretization process provides a parsable infrastructure for subsequent semantic encoding, especially when dealing with non-standard spellings (such as archaic variants in historical place names). The token units generated by the word segmentation module through adaptive strategies can effectively distinguish between core morphemes and interference noise in language variants, such as identifying and isolating character blocks of different language families in mixed language components, thereby avoiding segmentation errors caused by cross-language rule conflicts.

[0023] Next, semantic embedding encoding is performed on each to-be-translated place name Token unit in the sequence distribution of to-be-translated place name Token units to obtain a sequence distribution of semantic embedding encoding vectors of to-be-translated place name Token units. That is, in the technical solution of this application, each to-be-translated place name Token unit in the sequence distribution of to-be-translated place name Token units is respectively passed through a word embedding encoder based on Word2Vec to obtain a sequence distribution of semantic embedding encoding vectors of to-be-translated place name Token units. It should be understood that in traditional natural language processing, as discrete symbols (such as sub-words, roots, or character combinations), Token units themselves cannot directly express semantic relevance and internal language rules. For example, in agglutinative languages, a suffix Token representing the locative case and another suffix Token representing the possessive relationship, if only existing in the form of discrete symbols, it is difficult for the model to automatically infer the similarity of their grammatical functions; in cross-language scenarios, Token symbols of synonymous roots in different language families (such as "mountain" and "mountain") lack a computable relationship expression even more. Through unsupervised training, the word embedding encoder based on Word2Vec can map each discrete Token to a low-dimensional continuous vector space by using the statistical law of Token co-occurrence within the context window, so that Tokens with similar semantics or functions have geometric proximity in the vector space (for example, suffix vectors representing the locative case form a clustering distribution in the vector space). This continuous vector representation provides a quantifiable semantic basis for subsequent models.

[0024] Furthermore, the sequence distribution of the semantic embedding coding vector of the place name token unit to be translated is subjected to the place name token context semantic association coding to obtain the place name semantic context association coding vector to be translated, and is used as the place name semantic context association coding representation to be translated. That is, in the technical solution of the present application, the sequence distribution of the place name token unit semantic embedding coding vector to be translated is passed through a context semantic association encoder based on BiLSTM to obtain the place name semantic context association coding vector to be translated. It should be understood that the place name token unit semantic embedding coding vector to be translated only expresses the static semantics of the token (such as the independent meaning of the agglutinative language root), and cannot model the interaction rules between adjacent tokens (such as the constraining effect of the suffix on the stem syllable structure). For example, a suffix token vector representing a locative case, its semantic function needs to be combined with the vector of the preceding stem token to fully express the grammatical meaning of "in the position of...", while the unidirectional LSTM can only capture forward dependencies, and it is difficult to capture the reverse influence of the suffix on the preceding stem (such as the change in the stress position of the stem syllable by the German separable verb prefix). BiLSTM uses a bidirectional gating mechanism to model the temporal relationship of the token sequence in the forward and reverse time dimensions, so that the hidden state of each position simultaneously integrates historical and future contextual information (such as the grammatical function of the prefix affects the subsequent syllable segmentation, and the morphological characteristics of the suffix correct the pronunciation rules of the previous stem), thereby constructing a global semantic association field. Specifically, the gating structure of BiLSTM (input gate, forget gate, output gate) controls the transmission and forgetting of information flow in a parameterized manner. For example, when processing long token sequences of agglutinative languages, the forget gate can dynamically determine the long-term memory that needs to be retained (such as the core semantics of the root), while the input gate filters the local features of the current token (such as the grammatical attributes of the suffix). The superposition of the bidirectional mechanism enables the model to gradually accumulate the modification effect of the prefix on the stem (such as "mega-" means "huge") when encoding the "prefix + root + suffix" structure, and the reverse LSTM captures the morphological constraints of the suffix on the root (such as "-polis" means "city") in reverse. Finally, a context representation containing bidirectional dependencies is formed through the splicing of hidden states to obtain the semantic context-related encoding vector of the place name to be translated.It is worth mentioning that the bidirectional context encoding breaks through the limitation of local windows, enabling the model to handle long-distance dependencies (such as consonant assimilation across multiple tokens), such as identifying core syllable boundaries separated by multiple modifiers in compound words; in addition, the dynamic gating mechanism enhances the model's adaptability to irregular language phenomena, such as suppressing interference noise in spelling variants (such as redundant characters in historical place names) through the forget gate, while using the input gate to strengthen key morphological features (such as the indicative role of glottal stop symbols on syllable segmentation); finally, the hidden state of the bidirectional fusion provides an anti-ambiguous semantic representation for subsequent modules, such as distinguishing homonymous suffixes in agglutinative languages, and its context-related vector activates specific features in different dimensions.

[0025] Specifically, the S42 reconstructs and strengthens the local semantic relevance of the semantic context-related coding representation of the place name to be translated to obtain the semantic context-related enhanced coding representation of the place name to be translated. It should be understood that after the place name to be translated has been word embedded and contextual semantically encoded, although the global semantic dependency has been partially modeled, the long compound words formed by the superposition of syllables in agglutinative languages, the spelling pattern of alternating consonant clusters and vowels in inflectional languages, and the cross-language interference components of mixed language place names often manifest as nonlinear coupling of local semantic structures. Therefore, in order to achieve feature distillation and dynamic enhancement of the semantic context-related features of the place name to be translated, in the technical solution of the present application, the semantic context-related coding vector of the place name to be translated is reconstructed and strengthened in terms of local semantic relevance to obtain the semantic context-related enhanced coding vector of the place name to be translated. In this process, first, the feature phase reconstruction based on one-dimensional convolutional coding extracts multi-scale local structural patterns on the coding vector sequence through a sliding window mechanism (such as a convolution kernel of length 3 capturing the morphological interaction features of three adjacent tokens), and explicitly models the local phase information implicit in the continuous vector space (i.e., the relative relationship between adjacent dimensions); then, since the initial phase coding set may contain redundant information (such as smooth transition features in non-boundary areas) and noise interference (such as irregular spelling fluctuations caused by cross-language mixing), the information extraction stage uses a learnable attention mechanism or sparse constraints to screen the feature dimensions of the local phase coding vector. For example, in areas with dense consonant clusters, the model will strengthen the dimension that represents the steepness of consonant transitions, while suppressing the dimension that represents vowel stability; while in agglutinative segments with dense suffixes, the dimension that represents grammatical functions is retained; then, by counting the feature effective components, the information density index (i.e., the effective component statistics) of each local phase coding vector is calculated, and a dynamic gain operator is generated. For example, when it is detected that the statistics of a certain local phase encoding are significantly higher than the threshold, the gain operator will nonlinearly amplify the features of the corresponding dimension (such as generating an enhancement coefficient between 0 and 1 through the Sigmoid function), so as to highlight the mutation signals at the syllable boundary (such as the consonant cluster break point or the vowel weakening turning point) in the feature reshaping stage; finally, the local structure and the global context are coordinated and optimized through phase saliency reshaping. For example, when processing mixed language place names, the global context encoding may be difficult to accurately locate the syllable boundary due to cross-language rule conflicts, but the local phase enhancement module amplifies the local features that conform to the phonological rules of the target language (such as the open syllable preference in the Latin language family) and suppresses the interference signals of the conflicting language family (such as the long consonant sequence of the agglutinative language), so that the enhanced semantic context association enhancement encoding vector of the place name to be translated can still maintain stable discrimination in the cross-language interference scenario. This adaptive enhancement mechanism essentially constructs a multi-level feature interaction system, providing the deep decoder with an input representation that has both global robustness and local sensitivity.

[0026] Specifically, first, the semantic context-related coding vector of the place name to be translated is reconstructed by semantic units based on one-dimensional convolutional coding to obtain a set of local phase coding vectors of the semantic features of the place name to be translated. It should be understood that the semantic context-related coding vector of the place name to be translated output by BiLSTM contains global semantic dependencies (such as the cross-position grammatical association between roots and suffixes in agglutinative languages), which may weaken the micro-interaction rules between adjacent feature dimensions (such as the pronunciation transition characteristics of the consonant cluster area or the acoustic continuity of the vowel weakening area). For example, in semantic segments with dense consonant clusters, although the global coding vector can express the overall meaning of the morpheme, it is difficult to capture the steep changes within the consonant cluster (such as the alternating pattern of stops and fricatives), and this local mutation is precisely the key signal of the syllable boundary. Therefore, in the technical solution of the present application, the semantic context-related coding vector of the place name to be translated is reconstructed by semantic units based on one-dimensional convolutional coding to obtain a set of local phase coding vectors of the semantic features of the place name to be translated. Here, by using multiple sets of convolution kernels of different sizes (such as window lengths of 3 / 5 / 7), the model can capture local phase patterns from different receptive fields: short windows focus on microscopic interactions in adjacent dimensions (such as the break characteristics of double consonant clusters), while long windows model gradual changes across dimensions (such as tone continuity in vowel harmony). For example, in the dense suffix area of ​​agglutinative languages, multi-scale convolution kernels can simultaneously capture the consonant-vowel combination rules within the suffix (short window) and the morphological superposition trend of the suffix sequence (long window), forming a set of local phase encoding vectors of the semantic features of the place names to be translated that cover multiple possible forms of syllable boundaries. This multi-angle feature extraction provides a structured intermediate representation for subsequent modules, enabling the model to distinguish between real syllable boundaries (such as the break of consonant clusters) and pseudo boundaries (such as stable consonant combinations within roots). In addition, feature phase reconstruction significantly enhances the model's sensitivity to implicit local rules. For example, when dealing with mixed language interference, global encoding may confuse syllable boundaries due to cross-language rule conflicts, but local phase encoding effectively suppresses irrelevant features of the interfering language (such as long consonant sequences in agglutinative languages) by strengthening the phonological patterns unique to the target language (such as the open syllable preference of the Latin language). This fine-grained structured decoding lays an interpretable physical foundation for subsequent feature distillation and reshaping. In a specific example of the present application, the semantic context-related encoding vector of the place name to be translated is reconstructed based on one-dimensional convolutional coding using the following semantic unit reconstruction formula to obtain a set of local phase encoding vectors of the semantic features of the place name to be translated; wherein the semantic unit reconstruction formula is: ; in, is the semantic context-associated encoding vector of the place name to be translated, It is a one-dimensional convolutional coding process. is the characteristic phase reconstruction step size, are the first, second, and third local phase encoding vectors of the semantic features of the place names to be translated. and The local phase encoding vector of the semantic features of the place names to be translated.

[0027] Next, semantic purification is performed on each local phase encoding vector of the semantic features of the place names to be translated in the set of local phase encoding vectors of the semantic features of the place names to be translated to obtain a set of local phase encoding vectors of the semantic features of the place names to be translated. It should be understood that although the set of local phase encoding vectors of the semantic features of the place names to be translated generated by one-dimensional convolutional coding contains multi-scale structural information (such as the steep change pattern of consonant clusters or the gradual change trend of vowel harmony), there may be dimensional redundancy (such as similar edge features extracted by adjacent convolution kernels) and information mixing (such as the superposition of non-target phonological features caused by cross-language interference) inside it. For example, in the local phase encoding of dense consonant clusters, multiple convolution kernels may simultaneously activate responses to consonant transition features, resulting in collinearity between feature dimensions; in mixed language scenarios, some convolution kernels may capture the phonological rules of non-target languages ​​(such as the interference of long consonant sequences of agglutinative languages ​​on the syllable division of Latin languages), forming semantically irrelevant noise dimensions. Semantic refinement uses a learnable attention mechanism or sparse constraints to perform subspace projection on the high-dimensional space of the local phase encoding vectors of the semantic features of the place names to be translated, screen out key dimensions that are strongly related to syllable boundary judgment (such as mutation signals at consonant breakpoints or transition features of vowel weakening), and suppress low-information dimensions (such as gentle fluctuations in the stable vowel area) and cross-language interference noise. In this process, by introducing task-driven feature importance evaluation, the model can quantify the contribution of each local phase encoding vector of the semantic features of the place names to be translated to the final syllable segmentation goal, so as to strengthen the high-value features of cross-language commonality (such as the statistical laws of consonant cluster breaks) and weaken language-specific noise, thereby improving generalization ability. It is worth mentioning that semantic refinement is not a simple dimensionality reduction, but a reconstruction of the feature space through nonlinear transformation, such as using a gating mechanism to dynamically adjust the activation strength of each dimension, so that the set of extracted local phase encoding vectors of the semantic features of the place names to be translated not only retains the diversity of multi-scale structural information, but also has the discriminability of task adaptation. The purification and reconstruction of this feature space essentially builds a noise-resistant and highly discriminative local feature base for syllable boundary detection, providing an underlying guarantee for the reliability of the end-to-end translation system. In a specific example of the present application, the following semantic purification formula is used to semantically purify each local phase coding vector of the semantic features of the place names to be translated in the set of local phase coding vectors of the semantic features of the place names to be translated to obtain a set of local phase coding vectors of the semantic features of the place names to be translated; wherein the semantic purification formula is: ; in, is the one-norm of the vector, Extract the first local phase encoding vector from the set of semantic features of the place name to be translated The semantic features of the place names to be translated are extracted from the local phase encoding vector.

[0028] Then, the statistics of the effective components of the semantic features of the place names to be translated in the set of local phase encoding vectors for extracting the semantic features of the place names to be translated are calculated. It should be understood that although the set of local phase encoding vectors after semantic purification has initially filtered out redundant dimensions (such as low-variance fluctuation characteristics in the consonant cluster area) and noise interference (such as non-target phonological patterns caused by cross-language mixing), there is still heterogeneity in the contribution of feature dimensions within it - some dimensions may carry strong indications of syllable boundaries (such as steep gradient changes in consonant breakpoints), while other dimensions may only carry weakly correlated or poorly generalized local information (such as rare spelling variants unique to a specific language). For example, when dealing with agglutinative morphemes, the dimension that represents the suffix morphological rules may be globally universal for syllable segmentation, while the dimension that represents the consonant clusters within the root may only be valid in a specific language. In the technical solution of the present application, the statistics of the effective components of the semantic features of the place names to be translated of each local phase encoding vector of the semantic features of the place names to be translated in the set of local phase encoding vectors of the semantic features of the place names to be translated are calculated, wherein the statistics of the effective components identify the feature dimensions with high discriminability (such as cross-language stable consonant transition patterns) and low-value dimensions (such as language-specific decorative character combinations) through quantitative analysis (such as calculating the variance, entropy or mutual information of the dimension activation values ​​with the task labels), thereby providing an interpretable regulatory basis for the subsequent gain operator. By calculating the statistics of the effective components of the semantic features of the place names to be translated, the model can convert the abstract local phase encoding vector of the semantic features of the place names to be translated into an operational numerical indicator. For example, in the consonant assimilation area, the dimension with a higher statistical number may correspond to the physical characteristics of the change of the consonant pronunciation position (such as the transition slope from alveolar to velar), while the dimension with a lower statistical number may reflect irrelevant environmental noise (such as redundant symbols in spelling variants). This quantitative control mechanism based on statistics establishes a task-oriented feature importance ranking for the local phase coding vector extracted from the semantic features of the place names to be translated, so that the subsequent significance reshaping can be adaptively enhanced in a data-driven manner, providing a highly robust intermediate representation for the end-to-end translation system. In a specific example of the present application, the statistics of the effective components of the semantic features of the place names to be translated of the local phase coding vector extracted from the semantic features of the place names to be translated are calculated by the following statistical formula; wherein the statistical formula is: ; in, For the The semantic features of the place names to be translated are extracted from the local phase encoding vector The eigenvalues ​​at the positions, Indicates the active ingredient count, Preset threshold for trainability, for The corresponding statistics of effective components of semantic features of place names to be translated.

[0029] Then, based on the semantic features of each place name to be translated, the statistics of the effective components of the semantic features of the place name to be translated of the local phase encoding vector are extracted, and the phase reshaping gain operator of the semantic features of the place name to be translated of the local phase encoding vector of each semantic feature of the place name to be translated is calculated. It should be understood that after word embedding encoding and contextual semantic modeling, there are still noise interferences in the local semantic representation of the place name to be translated (such as irregularly spelled consonant clusters, morphological conflicts of cross-language mixed components), which will cover up the deep feature patterns that truly determine the syllable boundaries. By calculating the statistics of the effective components of the features, the system can quantify the effective information density that is strongly related to the syllable segmentation task in each local phase encoding vector. Furthermore, in order to construct a dynamic and adaptive feature enhancement mechanism, in the technical solution of the present application, the initial phase reshaping gain operator of the semantic features of the place name to be translated of the local phase encoding vector of the semantic features of each place name to be translated is calculated. Different from the traditional fixed weight feature enhancement method, the phase reshaping gain operator integrates linguistic rules (such as vowel-consonant collocation rules) with data-driven features to form an interpretable feature regulation tool. Specifically, the operator dynamically adjusts the activation thresholds of different local phase vectors according to the statistics of the effective components of the features: for encoding vectors containing high-frequency effective features (such as syllable boundary signals of iconic suffixes in agglutinative languages), the gain operator will amplify the dimensions related to syllable segmentation in its representation space; while for inefficient feature areas that are severely affected by cross-language interference (such as spelling conflict areas in mixed place names), the propagation of noise signals is suppressed through nonlinear scaling. In this way, the model has the ability to distinguish between "effective syllable boundary features" and "cross-language interference noise". For example, when processing mixed place names containing German compound words and Finnish suffixes, it can effectively strengthen the tense association features of consonant conversion nodes, while weakening the morphological distortion caused by differences in writing systems; in addition, this adaptive feature reshaping significantly improves the robustness of syllable segmentation, so that when the translation system faces languages ​​with blurred syllable boundaries such as Tibetan, it can still use the gain operator to phase align the initial and final combination patterns and accurately identify potential segmentation points affected by vowel length and consonant clusters. In a specific example of the present application, the phase reshaping gain operator of the semantic features of place names to be translated of each local phase coding vector of the semantic features of place names to be translated can be calculated by the following steps: based on the statistics of the effective components of the semantic features of place names to be translated that extract the local phase coding vector of each semantic feature of place names to be translated, the inhibition factor corresponding to the local phase coding vector extracted by each semantic feature of place names to be translated is determined; based on the inhibition factor corresponding to the local phase coding vector extracted by each semantic feature of place names to be translated, the initial phase reshaping gain operator of the semantic features of place names to be translated of each local phase coding vector of the semantic features of place names to be translated is calculated.In a specific example of the present application, the phase reshaping gain operator of the semantic features of the place names to be translated of each local phase coding vector of the semantic features of the place names to be translated can be calculated by the following steps: based on the statistics of the effective components of the semantic features of the place names to be translated that extract the local phase coding vector of each semantic feature of the place names to be translated, the inhibition factor corresponding to the local phase coding vector extracted by each semantic feature of the place names to be translated is determined; based on the inhibition factor corresponding to the local phase coding vector extracted by each semantic feature of the place names to be translated, the initial phase reshaping gain operator of the semantic features of the place names to be translated of each local phase coding vector of the semantic features of the place names to be translated is calculated; the initial phase reshaping gain operator of the semantic features of the place names to be translated is corrected for feature phase dispersion loss to obtain the phase reshaping gain operator of the semantic features of the place names to be translated.

[0030] In particular, when the initial phase reshaping gain operator of the semantic feature of the place name to be translated performs nonlinear amplification of the local phase encoding based on statistics (such as enhancing the dimension representing the steep gradient of the consonant break point by 1.5 times), excessive focus on the significance of the local feature may destroy the overall distribution structure of the feature space. For example, in the processing of long suffix sequences of agglutinative morphemes, if the gain operators of multiple suffix tokens independently enhance their grammatical function dimensions, it may cause the distribution of feature vectors in space to show non-uniform scattering (such as excessive separation of some vector clusters), thereby weakening the global context relevance. This phenomenon is mathematically manifested as the lack of symmetry in the phase direction, that is, the geometric direction distribution of the enhanced feature vector in space deviates from the inherent consistency of linguistic laws. Therefore, in the preferred example of the present application, the initial phase reshaping gain operator of the semantic feature of the place name to be translated is corrected for the lack of feature phase dispersion to obtain the phase reshaping gain operator of the semantic feature of the place name to be translated. That is, by introducing the flatness decomposition metric, the effect of the gain operator is constrained to maintain the holomorphic flatness of the space. For example, when dealing with consonant clusters of mixed language place names, the modified gain operator ensures that the consonant cluster features of the Latin language family and the long consonant sequences of the agglutinative morphemes maintain directional compatibility in the vector space, avoiding spatial distortion caused by conflicts in language family rules. This mathematical mechanism transforms the single-mode coupled initial semantic feature phase reshaping gain operator of the place name to be translated into a standardized operation that conforms to spatial geometric constraints through the construction of a canonical field, so that the feature enhancement process maintains the overall stability of the feature distribution while improving local significance. The modified initial semantic feature phase reshaping gain operator of the place name to be translated enables the model to stably generalize the phonological rules learned from scarce annotated data (such as the separation pattern of Tibetan ligatures) to unseen language variants by maintaining the translation invariance of the feature vector. This mathematical correction based on the holomorphic structure essentially constructs a deep mapping of linguistic rules and feature space geometry, providing an end-to-end translation system with a feature enhancement paradigm that is both locally sensitive and globally consistent.

[0031] In this example, based on the statistics of the effective components of the semantic features of the place names to be translated of the local phase coding vectors, the phase reshaping gain operator of the semantic features of the place names to be translated of the local phase coding vectors of the semantic features of the place names to be translated is calculated by the following calculation formula; wherein, the calculation formula is: ; ; ; ; in, is the number of vectors in the set of local phase encoding vectors of semantic features of place names to be translated, for The corresponding polar angle, for The corresponding inhibitory factor, represents pi, represents the inverse tangent function, The phase reshape gain operator for the semantic features of the place name to be translated, for The corresponding initial semantic feature phase reshaping gain operator of the place name to be translated, is the flatness factor of the semantic space of the place name to be translated, is the factor representing the semantic flatness decomposition metric of the place name to be translated, Reshape the gain operator for the phase of semantic features of place names to be translated.

[0032] Then, based on the phase reshaping gain operator of the semantic features of the place names to be translated of the local phase encoding vectors of the semantic features of the place names to be translated, the set of the local phase encoding vectors of the semantic features of the place names to be translated is reshaped in terms of feature phase significance to obtain the semantic context-related enhanced encoding vector of the place names to be translated. Here, after the phase reshaping gain operator of the semantic features of the place names to be translated of the local phase encoding vectors of the semantic features of the place names to be translated is dynamically regulated by statistical driving (such as assigning a 1.5 times gain coefficient to the consonant break point feature), its essence is to convert the phonological rules of linguistics (such as the morphological constraints of agglutinative suffixes) into geometric operations in feature space. For example, when processing compound words, the local phase encoding vector corresponding to the consonant assimilation phenomenon at the junction of the root and the suffix will form a direction-specific enhancement in the vector space after nonlinear scaling by the gain operator (such as the specific dimension of the feature vector is stretched to amplify the steepness of the consonant transition), thereby explicitly marking the potential position of the syllable boundary. This feature enhancement mechanism based on geometric space reconstruction essentially establishes an interpretable mapping between linguistic rules and machine-understandable feature distribution, providing a core technical support for end-to-end translation systems that combines domain knowledge embedding and data-driven adaptability. In a specific example of this application, the following feature phase reshaping formula is used to reshape the set of local phase encoding vectors of semantic features of place names to be translated to obtain semantic context-related enhanced encoding vectors of place names to be translated; wherein the feature phase reshaping formula is: ; ; in, For The value of the natural exponential function with base , Reshape the gain weights for the semantic feature phase of the place name to be translated, Enhance the encoding vector by associating the semantic context of the place name to be translated.

[0033] Specifically, the S43 determines the place name string after syllable segmentation based on the semantic contextual association enhanced coding representation of the place name to be translated. That is, in the technical solution of the present application, the semantic contextual association enhanced coding vector of the place name to be translated is input into the syllable segmentation decoder based on the deep neural network model to obtain the place name string after syllable segmentation. It should be understood that the semantic contextual association enhanced coding vector of the place name to be translated contains dual information of global contextual association and local morphological features, which essentially constructs a cross-language abstract phonological space, which contains both the historical trajectory of root evolution and the dynamic distribution pattern of syllable boundaries. Traditional decoders based on finite state machines are difficult to capture the complex syllable superposition logic inside; and the cross-linguistic components in mixed language place names interact, so the decoder has the flexible ability to dynamically adjust the phonological parsing strategy. The deep neural network decoder can establish a reverse mapping from the abstract feature space to the specific syllable sequence. Specifically, it first analyzes the burst energy distribution of consonant clusters (such as the plosive combination at the beginning of "pf-" in German) through multi-layer nonlinear transformation, and then combines the attention mechanism to dynamically focus on the characteristic mutation area at the splicing of agglutinative morphemes (such as " "The connection boundary between the suffix and the stem). In this way, the system can achieve a dynamic balance between phonological rule reasoning and morphological feature analysis, and achieve a breakthrough in the accuracy of syllable segmentation across language barriers. This decoding paradigm based on holomorphic structural constraints essentially builds a mathematical bridge between linguistic prior knowledge and data-driven models, providing an end-to-end solution for place name translation in a global scenario that is both rule-rigorous and adaptive.

[0034] In particular, in S5, the place name string after syllable segmentation is input into the place name translation model to obtain the place name target language translation text. Among them, the place name translation model is a deep learning model that can understand and convert the complex relationship between different languages. Specifically, the model contains multiple levels of neural network layers, each of which is responsible for capturing different aspects of the input data, from basic language symbol representation to more advanced semantic understanding and cross-language mapping. In this process, first, the model will perform a preliminary analysis of the input place name string to identify the meaning of each character or Token (sub-word unit) and its role in the sentence; then, using a deep neural network architecture, such as a bidirectional long short-term memory network (BiLSTM), Transformer, etc., the model begins to comprehensively analyze the input string; it must not only understand the meaning of each individual word, but also grasp the overall meaning of the entire place name string and the relationship between the various parts. For example, when dealing with compound words or place names containing agglutinative morphemes, the model needs to identify the implicit association between the root and the suffix and adjust the translation strategy accordingly; then, the place name translation model will reorganize and optimize the parsed information according to the characteristics of the target language. This includes but is not limited to adjusting the grammatical structure, selecting appropriate synonyms or near-synonyms, considering cultural background differences and other factors, in order to generate translation results that are both faithful to the original text and in line with the habits of the target language. It is worth noting that this stage may also involve the processing of some special morphological features, such as prefixes / suffixes, consonant clusters, etc., to ensure that the final output of the place name translation text is both natural and accurate. In this way, the system achieves end-to-end effective conversion from the original place name to the target translation name, greatly improving the efficiency and accuracy of cross-language information processing, especially in the fields of geographic information systems, multilingual map navigation, cross-border logistics, etc., showing important value.

[0035] It is worth mentioning that ordinary people in this field should know that in addition to the above-mentioned technical solution of "inputting the place name string after syllable segmentation into the place name translation model to obtain the place name target language translation text", other existing technologies can also be used to implement the technical process. For example, in another specific implementation scheme, the place name translation scheme in the paper "Neural Machine Translation for Low-Resource Named Entities" can be used to translate the place name string after syllable segmentation to obtain the place name target language translation text. It should be understood that "Neural Machine Translation for Low-Resource Named Entities" is a prior art, and in order to avoid redundancy, the content will not be expanded here.

[0036] In summary, according to the embodiment of the present application, a high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation is explained, which first splits the input place name into a sub-word Token sequence, and uses the context encoding model of the word embedding model to realize the semantic context association encoding of the place name to be translated; then, the internal structured information and dependency relationships in the context semantics of the place name Token to be translated are modeled by means of local semantic association reconstruction and reinforcement, the implicit syllable associations in compound words and agglutinative languages ​​are identified, and the local morphological features near the syllable boundaries (such as prefixes / suffixes, consonant clusters) are enhanced to solve the problems of irregular spelling and cross-language interference, and then the place name character string after output syllable segmentation is decoded to perform place name translation, and the corresponding place name target language translation text is generated, thereby realizing end-to-end effective conversion from the original place name to the target translated name.

[0037] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation, characterized in that: include: Receive a place name character string to be translated input by a user; Using the language identification model to detect the source language of the place name string to be translated; Determine the syllable segmentation strategy based on the source language of the place name string to be translated; In response to the syllable segmentation strategy being a syllable segmentation method based on a deep neural network model, based on the syllable segmentation strategy, the place name character string to be translated is syllable segmented to obtain the place name character string after syllable segmentation, including: performing context semantic association encoding based on the place name Token unit on the place name character string to be translated to obtain the semantic context association encoding representation of the place name to be translated; performing local semantic association reconstruction and enhancement on the semantic context association encoding representation of the place name to be translated to obtain the semantic context association enhanced encoding representation of the place name to be translated; based on the semantic context association enhanced encoding representation of the place name to be translated, determining the place name character string after syllable segmentation; The syllable-segmented place name character string is input into the place name translation model to obtain the place name target language translation text.

2. The high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation according to claim 1 is characterized in that: Based on the source language of the place name string to be translated, determine the syllable segmentation strategy, including: If the source language of the place name string to be translated is a high-resource language, choose the rule-based syllable segmentation method; If the source language of the place name string to be translated is a medium- or low-resource language, select the syllable segmentation method based on the statistical model; If the source language of the place name string to be translated is a mixed language, select the syllable segmentation method based on the deep neural network model.

3. The high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation according to claim 2 is characterized in that: High-resource languages ​​include English, French, and Spanish; medium- and low-resource languages ​​are Chinese, Japanese, Korean, and Vietnamese.

4. The high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation according to claim 2 is characterized in that: The place name string to be translated is encoded based on the contextual semantic association of the place name Token unit to obtain the semantic contextual association encoding representation of the place name to be translated, including: Perform word segmentation on the place name string to be translated to obtain the sequence distribution of the place name Token unit to be translated; Perform semantic embedding coding on each place name Token unit to be translated in the sequence distribution of the place name Token unit to be translated to obtain a sequence distribution of the semantic embedding coding vector of the place name Token unit to be translated; The sequence distribution of the semantic embedding coding vector of the place name Token unit to be translated is encoded with the place name token context semantic association to obtain the place name semantic context association coding vector to be translated, and used as the place name semantic context association coding representation to be translated.

5. The high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation according to claim 4 is characterized in that: Semantic embedding coding is performed on each place name Token unit to be translated in the sequence distribution of the place name Token unit to be translated to obtain a sequence distribution of the semantic embedding coding vector of the place name Token unit to be translated, including: Each place name Token unit to be translated in the sequence distribution of the place name Token unit to be translated is passed through a word embedding encoder based on Word2Vec to obtain a sequence distribution of the semantic embedding coding vector of the place name Token unit to be translated.

6. The high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation according to claim 5 is characterized in that: Performing place name token context semantic association coding on the sequence distribution of the semantic embedding coding vector of the place name to be translated to obtain the place name semantic context association coding vector to be translated, and using it as the place name semantic context association coding representation to be translated, including: The sequence distribution of the semantic embedding coding vector of the place name Token unit to be translated is passed through a BiLSTM-based contextual semantic association encoder to obtain the semantic contextual association coding vector of the place name to be translated.

7. The high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation according to claim 6 is characterized in that: The semantic context-related coding representation of the place name to be translated is reconstructed and enhanced in terms of local semantic relevance to obtain the semantic context-related enhanced coding representation of the place name to be translated, including: The semantic unit reconstruction and semantic purification processing are performed on the semantic context associated coding vector of the place name to be translated to obtain a set of local phase coding vectors for extracting the semantic features of the place name to be translated; Calculate the statistics of effective components of semantic features of place names to be translated of each semantic feature extraction local phase encoding vector of place names to be translated in the set of semantic feature extraction local phase encoding vectors of place names to be translated; Based on the semantic features of each place name to be translated, the statistics of the effective components of the semantic features of the place name to be translated of the local phase encoding vector are extracted, and the phase reshaping gain operator of the semantic features of the place name to be translated of the local phase encoding vector of the semantic features of each place name to be translated is calculated; Based on the phase reshaping gain operator of the semantic features of each place name to be translated, the set of local phase coding vectors of the semantic features of the place names to be translated is subjected to feature phase saliency reshaping to obtain the semantic context-related enhanced coding vector of the place names to be translated, which is used as the semantic context-related enhanced coding representation of the place names to be translated.

8. The high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation according to claim 7 is characterized in that: The semantic unit reconstruction and semantic purification processing are performed on the semantic context associated coding vector of the place name to be translated to obtain a set of local phase coding vectors for extracting the semantic features of the place name to be translated, including: The semantic context associated coding vector of the place name to be translated is reconstructed into a semantic unit based on one-dimensional convolutional coding to obtain a set of local phase coding vectors of the semantic features of the place name to be translated; Semantic purification is performed on each local phase coding vector of semantic features of place names to be translated in the set of local phase coding vectors of semantic features of place names to be translated to obtain a set of local phase coding vectors of semantic features of place names to be translated.

9. The high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation according to claim 8 is characterized in that: Based on the semantic features of each place name to be translated, the effective component statistics of the semantic features of the place name to be translated of the local phase encoding vector are extracted, and the phase reshaping gain operator of the semantic features of the place name to be translated of the local phase encoding vector of the semantic features of each place name to be translated is calculated, including: Based on the statistics of effective components of semantic features of place names to be translated extracted from local phase coding vectors of semantic features of place names to be translated, the inhibition factors corresponding to the local phase coding vectors extracted from the semantic features of place names to be translated are determined; Based on the semantic features of each place name to be translated, the suppression factor corresponding to the local phase encoding vector is extracted, and the initial semantic feature phase reshaping gain operator of the semantic feature of each place name to be translated is calculated; The feature phase dispersion missing correction is performed on the initial semantic feature phase reshaping gain operator of the place name to be translated to obtain the semantic feature phase reshaping gain operator of the place name to be translated.

10. The high-precision place name translation method integrating artificial intelligence and multilingual syllable segmentation according to claim 9 is characterized in that: Based on the semantic context-related enhanced coding representation of the place name to be translated, the place name string after syllable segmentation is determined, including: The semantic context-related enhanced coding vector of the place name to be translated is input into the syllable segmentation decoder based on the deep neural network model to obtain a syllable-segmented place name character string.

Images