Question and answer task processing method and device, equipment, storage medium and program product

Abstract

This invention provides a question-answering task processing method, apparatus, device, storage medium, and program product, applied in the field of natural language processing technology. The method includes: acquiring an input question and question-answering instances for a question-answering task; determining a target instance with the highest similarity to the input question from the question-answering instances; determining multiple LORA modules with the highest similarity to the target instance from a pre-trained low-rank adaptation LORA module library; fusing the multiple LORA modules with a first question-answering model using a non-gradient parameter optimization technique to obtain a second question-answering model; and inputting the input question into the second question-answering model to obtain the corresponding answer.

Description

Technical Field

[0001] This invention relates to the field of natural language processing technology, and in particular to a question-and-answer task processing method, apparatus, device, storage medium, and program product. Background Technology

[0002] In the field of natural language processing, large language models are generally not good at handling unseen tasks. Unseen tasks are entirely new tasks that are not part of the tasks that large language models are pre-trained on.

[0003] In existing technologies, the most direct way to solve the cross-task generalization problem of large language models is to pre-train the model. However, directly training the model requires a large amount of training data, and also consumes a lot of computational resources due to the influence of gradient backpropagation. Summary of the Invention

[0004] This invention provides a question-answering task processing method, apparatus, device, storage medium, and program product to solve the problem that direct training of models in the prior art requires a large amount of training data and consumes a lot of computing resources.

[0005] This invention provides a question-answering task processing method, comprising: acquiring an input question and question-answering instances for the question-answering task; determining a target instance with the highest similarity to the input question from the question-answering instances; determining multiple LORA modules with the highest similarity to the target instance from a pre-trained low-rank adaptation LORA module library; fusing the multiple LORA modules with a first question-answering model using a non-gradient parameter optimization technique to obtain a second question-answering model; and inputting the input question into the second question-answering model to obtain a corresponding answer.

[0006] According to a question-answering task processing method provided by the present invention, the step of determining the target instance with the highest similarity to the input question from the question-answering instances includes: extracting features from the input and output of the question-answering instances based on a sentence vectorization model to obtain an instance feature vector of the question-answering instances, and extracting features from the input question to obtain an input feature vector; calculating the cosine similarity between the input feature vector and each instance feature vector, and determining the question-answering instance corresponding to the instance feature vector with the highest similarity as the target instance.

[0007] According to a question-answering task processing method provided by the present invention, the step of determining the multiple LORA modules with the highest similarity to the target instance from a pre-trained low-rank adapted LORA module library includes: vectorizing each training data in the training dataset of the LORA module based on a sentence vectorization model, and determining the average vector of the training dataset as the representative vector of the LORA module; calculating the cosine similarity between the instance feature vector of the target instance and the representative vector of each LORA module; sorting the LORA modules in the LORA module library in descending order of cosine similarity, and determining the multiple LORA modules with the highest similarity.

[0008] According to a question-answering task processing method provided by the present invention, the step of fusing the multiple LORA modules with a first question-answering model using non-gradient parameter optimization technology to obtain a second question-answering model includes: adjusting the fusion parameters using non-gradient parameter optimization technology, weighting and adding the multiple LORA modules according to the fusion parameters to obtain a LORA module group of the target instance; and fusing the LORA module group with the first question-answering model to obtain the second question-answering model.

[0009] The present invention also provides a question-answering task processing device, comprising the following modules: an acquisition module and a processing module; the acquisition module is used to acquire the input question and question-answer instances of the question-answering task; the processing module is used to determine the target instance with the highest similarity to the input question from the question-answer instances, and to determine multiple LORA modules with the highest similarity to the target instance from a pre-trained low-rank adaptation LORA module library; the multiple LORA modules are fused with a first question-answering model using a non-gradient parameter optimization technique to obtain a second question-answering model, and the input question is input into the second question-answering model to obtain the corresponding answer.

[0010] According to a question-answering task processing device provided by the present invention, the processing module is used to extract features from the input and output of the question-answering instance based on a sentence vectorization model to obtain an instance feature vector of the question-answering instance, and to extract features from the input question to obtain an input feature vector; calculate the cosine similarity between the input feature vector and each instance feature vector, and determine the question-answering instance corresponding to the instance feature vector with the highest similarity as the target instance.

[0011] According to a question-answering task processing device provided by the present invention, the processing module is configured to: vectorize each training data in the training dataset of the LORA module based on a sentence vectorization model; determine the average value of the vectors in the training dataset as the representative vector of the LORA module; calculate the cosine similarity between the instance feature vector of the target instance and the representative vector of each LORA module; sort the LORA modules in the LORA module library in descending order of cosine similarity; and determine the plurality of LORA modules with the highest similarity.

[0012] According to a question-answering task processing device provided by the present invention, the processing module is used to adjust the fusion parameters using non-gradient parameter optimization technology, weight and sum the multiple LORA modules according to the fusion parameters to obtain the LORA module group of the target instance; and fuse the LORA module group with the first question-answering model to obtain the second question-answering model.

[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the question-answering task processing method as described above.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the question-answering task processing method as described above.

[0015] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the question-answering task processing method as described above.

[0016] The question-answering task processing method, apparatus, device, storage medium, and program product provided by this invention can acquire the input question and question-answer instances for a question-answering task; determine the target instance with the highest similarity to the input question from the question-answer instances; determine multiple LORA modules with the highest similarity to the target instance from a pre-trained low-rank adapted LORA module library; use non-gradient parameter optimization techniques to fuse the multiple LORA modules with a first question-answering model to obtain a second question-answering model; and input the input question into the second question-answering model to obtain the corresponding answer. Through this scheme, since multiple effective LORA modules can be retrieved and fused based on the target instance, the possibility of the question-answering model capturing the potential specificity of the input using limited instance data can be improved, thereby achieving improved cross-task generalization performance at a lower cost. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 This is one of the flowcharts of the question-and-answer task processing method provided by the present invention;

[0019] Figure 2 This is the second flowchart of the question-and-answer task processing method provided by the present invention;

[0020] Figure 3 This is a schematic diagram of the question-and-answer task processing device provided by the present invention;

[0021] Figure 4 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] It should be noted that in the embodiments of this application, the words "exemplary" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the words "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0024] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0025] To facilitate a clear description of the technical solutions of the embodiments of this application, the terms "first" and "second" are used in the embodiments of this application to distinguish the same or similar items with essentially the same function and effect. Those skilled in the art can understand that the terms "first" and "second" are not intended to limit the quantity or execution order.

[0026] This application describes some exemplary embodiments for illustrative purposes. It should be understood that this application may be implemented in other ways not specifically shown in the accompanying drawings.

[0027] Low-Rank Adaptation (LORA) is a technique in Natural Language Processing (NLP) for efficient parameter fine-tuning, particularly in large-scale pre-trained language models. The core idea of ​​LORA is to introduce a low-rank matrix to approximate the update of model parameters, thereby reducing the number of parameters that need to be fine-tuned.

[0028] LoraHub leverages the modular nature of LORA technology to dynamically combine LORA modules trained on different tasks into a powerful integrated module. This improves the model's performance on diverse tasks without increasing input requirements, achieving a balance between performance and energy consumption.

[0029] However, when selecting LoRa modules from the trained LoRa library for task-adaptive composition, LoraHub's strategy is to randomly select modules from the trained LoRa library. On the one hand, this random selection may result in the selected tasks being completely unrelated, leading to significant performance differences for the same task across multiple random selections, exhibiting poor stability. On the other hand, when training instances, LoraHub does not consider the differences between individual instances, which prevents the possibility of capturing potential input specificity using limited instance data, thus limiting LoraHub's performance.

[0030] To solve the above problems, such as Figure 1 As shown, this application provides a question-and-answer task processing method, which can be applied to a question-and-answer task processing device. The question-and-answer task processing method may include steps S101-S103:

[0031] S101, The question-and-answer task processing device obtains the input question and question-and-answer instance of the question-and-answer task.

[0032] Optionally, the question-answering task processing device can acquire a small number of question-answering instances of the question-answering task. The input question can be an unknown question, that is, a completely new task outside of the pre-training task of the first question-answering model.

[0033] S102, the question-answering task processing device determines the target instance with the highest similarity to the input question from the question-answering instances, and determines multiple LORA modules with the highest similarity to the target instance from the pre-trained low-rank adaptation LORA module library.

[0034] Optionally, the question-answering task processing device determines the target instance with the highest similarity to the input question from the question-answering instances, including: extracting features from the input and output of the question-answering instances based on a sentence vectorization model to obtain an instance feature vector of the question-answering instance, and extracting features from the input question to obtain an input feature vector; calculating the cosine similarity between the input feature vector and each instance feature vector, and determining the question-answering instance corresponding to the instance feature vector with the highest similarity as the target instance.

[0035] Optionally, the question-answering task processing device determines multiple LORA modules with the highest similarity to the target instance from a pre-trained low-rank adapted LORA module library, including: vectorizing each training data in the training dataset of the LORA module based on a sentence vectorization model, and determining the average vector of the training dataset as the representative vector of the LORA module; calculating the cosine similarity between the instance feature vector of the target instance and the representative vector of each LORA module; sorting the LORA modules in the LORA module library in descending order of cosine similarity, and determining the multiple LORA modules with the highest similarity.

[0036] Specifically, such as Figure 2 As shown, the question-answering task processing device can first determine the input feature vector corresponding to the input question, the instance feature vector corresponding to the question-answering instance, and the representative vector corresponding to each LORA module in the LORA module library; then, it uses cosine similarity to calculate the similarity between the instance feature vector and the input feature vector one by one, thereby determining the target instance most similar to the input feature vector; then, it uses cosine similarity to calculate the similarity between the instance feature vector of the target instance and the representative vector of each LORA module; finally, it selects multiple LORA modules that match the target instance according to the order of similarity from largest to smallest.

[0037] S103. The question-answering task processing device uses non-gradient parameter optimization technology to fuse the multiple LORA modules with the first question-answering model to obtain a second question-answering model, and inputs the input question into the second question-answering model to obtain the corresponding answer.

[0038] Optionally, the question-answering task processing device uses non-gradient parameter optimization technology to fuse the multiple LORA modules with the first question-answering model to obtain a second question-answering model, including: adjusting the fusion parameters using non-gradient parameter optimization technology, weighting and adding the multiple LORA modules according to the fusion parameters to obtain the LORA module group of the target instance; and fusing the LORA module group with the first question-answering model to obtain the second question-answering model.

[0039] Specifically, the question-answering task processing device can use the LoraHub method, employing non-gradient parameter optimization techniques to adjust the fusion parameters, and weighting the Lora modules according to the fusion parameters to obtain the Lora module group of the target instance. Then, the question-answering task processing device can fuse the Lora module group, the target instance, and the first question-answering model to obtain the second question-answering model. Finally, inputting the input question into the second question-answering model will yield the corresponding answer.

[0040] It should be noted that when there are many input questions, the question-answering task processing device can perform batch reasoning on the input questions, using question-answering instances as boundaries. The solution proposed in this application can effectively improve the question-answering model's processing capability on unseen tasks, while maintaining a balance between performance and energy consumption compared to full-scale fine-tuning.

[0041] In this embodiment, since multiple valid LORA modules can be retrieved based on the target instance for parameter fusion, the possibility of the question answering model capturing the potential specificity of the input using limited instance data can be improved, thereby achieving improved cross-task generalization performance at a lower cost.

[0042] The foregoing mainly describes the solutions provided by the embodiments of this application from a methodological perspective. To achieve the above functions, it includes corresponding hardware structures and / or software modules for executing each function. Those skilled in the art should readily recognize that, in conjunction with the units and algorithm steps of the various examples described in the embodiments disclosed herein, the embodiments of this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0043] The question-and-answer task processing method provided in this application can be executed by a question-and-answer task processing device or a control module for question-and-answer task processing within that device. This application uses the execution of the question-and-answer task processing method by a question-and-answer task processing device as an example to illustrate the question-and-answer task processing device provided in this application.

[0044] It should be noted that the embodiments of this application can divide the question-answering task processing device into functional modules according to the above method examples. For example, each function can be divided into its own functional modules, or two or more functions can be integrated into one processing module. The integrated modules can be implemented in hardware or as software functional modules. Optionally, the module division in the embodiments of this application is illustrative and is only a logical functional division; other division methods may be used in actual implementation.

[0045] like Figure 3As shown in the figure, this application embodiment provides a question-answering task processing device 300. The question-answering task processing device 300 includes an acquisition module 301 and a processing module 302. The acquisition module 301 is used to acquire the input question and question-answer instances of the question-answering task; the processing module 302 is used to determine the target instance with the highest similarity to the input question from the question-answer instances, determine multiple LORA modules with the highest similarity to the target instance from a pre-trained low-rank adaptation LORA module library; use non-gradient parameter optimization technology to fuse the multiple LORA modules with a first question-answering model to obtain a second question-answering model, and input the input question into the second question-answering model to obtain the corresponding answer.

[0046] Optionally, the above processing module 502 is used to extract features from the input and output of the question-and-answer instance based on the sentence vectorization model to obtain the instance feature vector of the question-and-answer instance, and extract features from the input question to obtain the input feature vector; calculate the cosine similarity between the input feature vector and each instance feature vector, and determine the question-and-answer instance corresponding to the instance feature vector with the highest similarity as the target instance.

[0047] Optionally, the processing module 502 is used to vectorize each training data in the training dataset of the LORA module based on the sentence vectorization model, and determine the average value of the vectors in the training dataset as the representative vector of the LORA module; calculate the cosine similarity between the instance feature vector of the target instance and the representative vector of each LORA module; sort the LORA modules in the LORA module library in descending order of cosine similarity, and determine the multiple LORA modules with the highest similarity.

[0048] Optionally, the above-mentioned processing module 502 is used to adjust the fusion parameters using non-gradient parameter optimization techniques, weight the multiple LORA modules according to the fusion parameters to obtain the LORA module group of the target instance, and fuse the LORA module group with the first question-answering model to obtain the second question-answering model.

[0049] In this embodiment, since multiple valid LORA modules can be retrieved based on the target instance for parameter fusion, the possibility of the question answering model capturing the potential specificity of the input using limited instance data can be improved, thereby achieving improved cross-task generalization performance at a lower cost.

[0050] Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4As shown, the electronic device may include a processor 410, a communications interface 420, a memory 430, and a communication bus 440, wherein the processor 410, communications interface 420, and memory 430 communicate with each other via the communication bus 440. The processor 410 can call logical instructions in the memory 430 to execute a question-answering task processing method. This method includes: acquiring an input question and question-answering instances; determining a target instance with the highest similarity to the input question from the question-answering instances; determining multiple LORA modules with the highest similarity to the target instance from a pre-trained low-rank adaptation LORA module library; fusing the multiple LORA modules with a first question-answering model using non-gradient parameter optimization techniques to obtain a second question-answering model; and inputting the input question into the second question-answering model to obtain a corresponding answer.

[0051] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0052] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the question-answering task processing method provided by the above methods. The method includes: obtaining an input question and a question-answering instance for the question-answering task; determining a target instance with the highest similarity to the input question from the question-answering instances; determining multiple LORA modules with the highest similarity to the target instance from a pre-trained low-rank adaptation LORA module library; using non-gradient parameter optimization techniques to fuse the multiple LORA modules with a first question-answering model to obtain a second question-answering model; and inputting the input question into the second question-answering model to obtain a corresponding answer.

[0053] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the question-answering task processing method provided by the above methods. The method includes: obtaining an input question and a question-answering instance for a question-answering task; determining a target instance with the highest similarity to the input question from the question-answering instances; determining multiple LORA modules with the highest similarity to the target instance from a pre-trained low-rank adaptation LORA module library; fusing the multiple LORA modules with a first question-answering model using a non-gradient parameter optimization technique to obtain a second question-answering model; and inputting the input question into the second question-answering model to obtain a corresponding answer.

[0054] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0055] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0056] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A question-answering task processing method, characterized in that, include: Obtain the input questions and question-and-answer instances for the question-and-answer task; From the question-answering instances, determine the target instance with the highest similarity to the input question, and from the pre-trained low-rank adapted LORA module library, determine multiple LORA modules with the highest similarity to the target instance; The multiple LORA modules are fused with the first question-answering model using a non-gradient parameter optimization technique to obtain a second question-answering model. The input question is then input into the second question-answering model to obtain the corresponding answer. The step of determining the target instance with the highest similarity to the input question from the question-answering instances includes: Based on the sentence vectorization model, feature extraction is performed on the input and output of the question-and-answer instance to obtain the instance feature vector of the question-and-answer instance, and feature extraction is performed on the input question to obtain the input feature vector; Calculate the cosine similarity between the input feature vector and each instance feature vector, and determine the question-answer instance corresponding to the instance feature vector with the highest similarity as the target instance; The step of determining the multiple LORA modules with the highest similarity to the target instance from the pre-trained low-rank adapted LORA module library includes: Based on the sentence vectorization model, each training data in the training dataset of the LORA module is vectorized, and the average value of the vectors in the training dataset is determined as the representative vector of the LORA module. Calculate the cosine similarity between the instance feature vector of the target instance and the representative vector of each LORA module; The LORA modules in the LORA module library are sorted in descending order of cosine similarity, and the LORA modules with the highest similarity are determined.

2. The question-answering task processing method according to claim 1, characterized in that, The second question-answering model is obtained by fusing the multiple LORA modules with the first question-answering model using a non-gradient parameter optimization technique, including: The fusion parameters are adjusted using a non-gradient parameter optimization technique, and the multiple LORA modules are weighted and summed according to the fusion parameters to obtain the LORA module group of the target instance; The LORA module group is fused with the first question-answering model to obtain the second question-answering model.

3. A question-and-answer task processing device, characterized in that, include: Acquisition module and processing module; The acquisition module is used to acquire the input questions and question-and-answer instances for the question-and-answer task; The processing module is used to determine the target instance with the highest similarity to the input question from the question-answering instances, determine multiple LORA modules with the highest similarity to the target instance from the pre-trained low-rank adaptation LORA module library, and use non-gradient parameter optimization technology to fuse the multiple LORA modules with the first question-answering model to obtain the second question-answering model, and input the input question into the second question-answering model to obtain the corresponding answer; The step of determining the target instance with the highest similarity to the input question from the question-answering instances includes: Based on the sentence vectorization model, feature extraction is performed on the input and output of the question-and-answer instance to obtain the instance feature vector of the question-and-answer instance, and feature extraction is performed on the input question to obtain the input feature vector; Calculate the cosine similarity between the input feature vector and each instance feature vector, and determine the question-answer instance corresponding to the instance feature vector with the highest similarity as the target instance; The step of determining the multiple LORA modules with the highest similarity to the target instance from the pre-trained low-rank adapted LORA module library includes: Based on the sentence vectorization model, each training data in the training dataset of the LORA module is vectorized, and the average value of the vectors in the training dataset is determined as the representative vector of the LORA module. Calculate the cosine similarity between the instance feature vector of the target instance and the representative vector of each LORA module; The LORA modules in the LORA module library are sorted in descending order of cosine similarity, and the LORA modules with the highest similarity are determined.

4. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the question-answering task processing method as described in any one of claims 1 to 2.

5. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the question-answering task processing method as described in any one of claims 1 to 2.

6. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the question-answering task processing method as described in any one of claims 1 to 2.

Images