Search

CN-115510877-B - Priority-based translation enhancement method and system

CN115510877BCN 115510877 BCN115510877 BCN 115510877BCN-115510877-B

Abstract

The invention provides a translation enhancing method and system based on priority, and belongs to the technical field of translation enhancing. The method comprises the steps of S1, receiving a plurality of translation sequences, determining at least one translation priority contained in each translation sequence, S2, determining a corresponding translation pointer for each translation priority, S3, obtaining resource parameters of a translation engine array, S4, determining a schedulable translation pointer of each translation engine model based on the resource parameters, S5, determining at least one translation sequence as a sequence to be translated by each translation engine model based on the schedulable translation pointer, and S6, taking all the arrays to be translated as inputs of the translation engine array, wherein the translation engine array outputs a translation enhancement result. The system includes a translation engine array including a plurality of translation engine models for implementing the method. The invention can realize the output of translation enhancement results with different priorities.

Inventors

  • HE ZHENGYU

Assignees

  • 语联网（武汉）信息技术有限公司

Dates

Publication Date
20260512
Application Date
20221011

Claims (10)

  1. 1. A priority-based translation enhancement method, the method comprising the steps of: The method comprises the following steps of S1, receiving a plurality of translation sequences, and determining at least one translation priority contained in each translation sequence based on a sequence tag or a sequence keyword of each translation sequence, wherein the translation priorities of the translation sequences comprise a plain text translation priority, a plain audio translation priority, an image-text translation priority and a video translation priority; S2, determining a corresponding translation pointer for each translation priority, wherein the translation pointer corresponding to each translation priority is used for pointing to a storage address of a translation sequence with the translation priority; s3, acquiring resource parameters of a translation engine array; s4, determining a schedulable translation pointer of each translation engine model based on the resource parameters; S5, each translation engine model determines at least one translation sequence as a sequence to be translated based on a schedulable translation pointer; S6, taking all arrays to be translated as inputs of a translation engine array, and outputting a translation enhancement result by the translation engine array.
  2. 2. A priority based translation enhancing method as claimed in claim 1, In the step S1, a text input device is used to input a text sequence to be translated as the translation sequence, and/or a voice input device is used to receive a voice sequence to be translated as the translation sequence.
  3. 3. A priority-based translation enhancing method as claimed in claim 2, wherein: In the step S2, if the translation sequence includes a plain text translation priority and a plain audio translation priority, the pointer P0 of the plain text translation priority and the plain audio translation priority P1 are defined as addresses of the storage unit where the translation sequence is located.
  4. 4. A priority-based translation enhancing method as claimed in claim 2, wherein: The step S3 of obtaining the resource parameters of the translation engine array includes: Acquiring real-time resource parameters of the translation engine array, and predicting available resource parameters of the translation engine array in a future preset time period based on the real-time resource parameters; The translation engine array comprises a plurality of translation engine models, and each translation engine model can realize translation enhancement results corresponding to different translation priorities.
  5. 5. A priority-based translation enhancing method as claimed in claim 2, wherein: in the step S4, a schedulable translation pointer of each translation engine model is determined based on the resource parameter, which specifically includes: determining real-time resource parameters of each translation engine model and available resource parameters in a future preset time period; And when the real-time resource parameters of the first translation engine model and the available resource parameters in a preset time period in the future support the translation enhancement of the first translation priority, taking a first translation pointer corresponding to the first translation priority as a schedulable translation pointer of the first translation engine model.
  6. 6. The priority-based translation enhancing method as recited in claim 5, wherein: The step S5 comprises the steps of taking a translation sequence comprising the first translation priority as a sequence to be translated; The step S6 comprises inputting the sequence to be translated into a first translation engine model, and outputting a translation enhancing result corresponding to a first translation priority by the first translation engine model.
  7. 7. A priority-based translation enhancement system, the translation enhancement system comprising a translation engine array comprising a plurality of translation engine models; Characterized in that the system further comprises: the sequence receiving unit is used for receiving and storing the translation sequence; the priority identification unit is used for identifying the translation priority contained in each translation sequence, wherein the translation priority comprises a pure text translation priority, a pure audio translation priority, an image-text translation priority and a video translation priority; A pointer determination unit that determines a corresponding translation pointer based on each of the translation priorities, the translation pointer corresponding to each of the translation priorities being used to point to a storage address of a translation sequence having the translation priority; The real-time resource parameter acquisition unit is used for acquiring real-time resource parameters of the translation engine array; The available resource parameter prediction unit predicts available resource parameters in a preset time period in the future of the translation engine array based on the real-time resource parameters; A schedulable pointer determination unit for determining a schedulable translation pointer for each translation engine model based on the real-time resource parameter and the available resource parameter; A sequence to be translated determining unit that determines at least one translation sequence based on a schedulable translation pointer as a sequence to be translated based on each translation engine model; And the translation enhancement output unit takes all arrays to be translated as inputs of a translation engine array, and the translation engine array outputs a translation enhancement result.
  8. 8. A priority-based translation enhancing system as recited in claim 7 wherein: the priority identification unit determines at least one translation priority included in each translation sequence based on a sequence tag or a sequence keyword of each translation sequence.
  9. 9. A priority-based translation enhancing system as recited in claim 7 wherein: each translation engine model may implement translation enhancement results corresponding to different translation priorities.
  10. 10. A priority-based translation enhancing system as claimed in claim 7 or 9, wherein: The schedulable pointer determining unit is configured to determine a schedulable translation pointer of each translation engine model based on the real-time resource parameter and the available resource parameter, and specifically includes: And when the real-time resource parameters of the first translation engine model and the available resource parameters in a preset time period in the future support the translation enhancement of the first translation priority, taking a first translation pointer corresponding to the first translation priority as a schedulable translation pointer of the first translation engine model.

Description

Priority-based translation enhancement method and system Technical Field The invention belongs to the technical field of translation enhancement, and particularly relates to a priority-based translation enhancement method and system. Background The translation enhancement is also called enhanced translation (EnhanceTranslation), which means that the corpus to be translated is output in the form of multimedia enhanced translation under the condition that the equipment resource meets the enhancement output. For example, a generic translation may implement text (transliteration) or audio output (interpretation) of the translated sequence, but an enhanced translation may further implement a teletext output, video output, or other form of multimedia output of the translated sequence, such as implementing an augmented reality scene play on a wearable augmented reality device upon receiving a teaching sequence to be translated, obviously, the enhanced translation also includes a generic translation. While enhanced translation may enable multimedia output, not all devices may schedule enough resources to support enhanced translation at a time on the one hand, and not every sequence to be translated may be necessary to schedule resources to enable enhanced translation on the other hand. The former causes a problem of how to dispatch a translation sequence with enhanced translation requirements to a resource-matched device to achieve enhanced translation, and the latter faces a problem of how to achieve only ordinary translation without the translation sequence requiring enhanced translation output, so as not to waste system resources. In addition, in a translation scene which is focused on real-time or near real-time, a user expects a minimum common translation output for a certain sequence to be translated even if the current translation resource does not support enhanced translation, and meanwhile, when the enhanced translation is supported, the user needs to obtain an enhanced translation result within a tolerable waiting period instead of frequent clamping. Therefore, there are a number of technical problems that need to be solved in the current enhanced translation. Disclosure of Invention In order to solve some or all of the above technical problems, the present invention provides a priority-based translation enhancement method and system. Specifically, in a first aspect of the present invention, a priority-based translation enhancement method is provided, which includes the following steps: s1, receiving a plurality of translation sequences, and determining at least one translation priority contained in each translation sequence; specifically, in the step S1, the translation priority of the translation sequence includes a plain text translation priority, a pure audio translation priority, an image-text translation priority, and a video translation priority; the step S1 determines at least one translation priority contained in each translation sequence based on the sequence tag or the sequence keyword of each translation sequence. S2, determining a corresponding translation pointer for each translation priority; Specifically, in step S2, the translation pointer corresponding to each translation priority is used to point to the storage address of the translation sequence with the translation priority. S3, acquiring resource parameters of a translation engine array; s4, determining a schedulable translation pointer of each translation engine model based on the resource parameters; S5, each translation engine model determines at least one translation sequence as a sequence to be translated based on a schedulable translation pointer; S6, taking all arrays to be translated as inputs of a translation engine array, and outputting a translation enhancement result by the translation engine array. As a further improvement, the step S3 of obtaining the resource parameters of the translation engine array includes: Acquiring real-time resource parameters of the translation engine array, and predicting available resource parameters of the translation engine array in a future preset time period based on the real-time resource parameters; The translation engine array comprises a plurality of translation engine models, and each translation engine model can realize translation enhancement results corresponding to different translation priorities. In the step S4, a schedulable translation pointer of each translation engine model is determined based on the resource parameter, which specifically includes: determining real-time resource parameters of each translation engine model and available resource parameters in a future preset time period; And when the real-time resource parameters of the first translation engine model and the available resource parameters in a preset time period in the future support the translation enhancement of the first translation priority, taking a first translation pointer corresponding to the first tr