Training system and method for intelligently training experimental animal administrator based on AI technology
The AI-based intelligent training system enables personalized and comprehensive training for laboratory animal handlers, solving the problems of low efficiency and insufficient evaluation in existing training methods, and improving training quality and animal welfare.
Patent Information
- Application Number
- CN202511238175.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-01
- Publication Date
- 2025-11-21
AI Technical Summary
Existing training methods for laboratory animal handlers are inefficient, lack personalized and comprehensive assessment, are not updated in a timely manner, lack real-time evaluation, and have limited skills assessments, which affects the accuracy of experimental results and animal welfare.
The AI-based intelligent training system includes an interaction module, a data acquisition module, a data transmission module, a data processing module, and an AI module. Through multi-dimensional assessment and real-time data analysis, it provides personalized training content and comprehensive skills evaluation.
It improves learning flexibility and training comprehensiveness, ensures correct skill mastery, reduces animal suffering, improves training quality and efficiency, and meets personalized learning needs.
Smart Images
Figure CN120997009A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of laboratory animal management and training technology, specifically to a training system and method for intelligent training of laboratory animal administrators based on AI technology. Background Technology
[0002] Animal handlers play a crucial role in many fields, including biomedical research, pharmacology experiments, and preclinical studies. They are responsible for key aspects such as animal husbandry, care, and experimental procedures. However, current training methods for animal handlers have several limitations. Traditional training relies heavily on lectures and on-site observation, resulting in low efficiency and difficulty in tailoring instruction to individual differences. Training content is not updated promptly, failing to adapt to new experimental techniques and animal species. Furthermore, there is a lack of real-time and comprehensive evaluation of training effectiveness. These shortcomings can lead to inconsistent professional skills among animal handlers, impacting animal welfare and the accuracy and reliability of experimental results. In addition, animal handlers need to master a wide range of knowledge, including the "Regulations on the Management of Laboratory Animals," AAALAC certification standards, Standard Operating Procedures (SOPs), animal ethics, animal welfare regulations, animal behavior, operational skills, pharmacology, pharmacodynamics, pharmacokinetics, the "Guidelines for Nonclinical Safety Evaluation," data collection, experimental design, and data interpretation. Existing training systems also have deficiencies in the systematic nature, comprehensiveness, and assessment of these aspects. Meanwhile, self-safety protection, laboratory safety, and facility management are also important aspects that laboratory animal administrators must master, but they are often overlooked or poorly implemented in existing training programs.
[0003] Existing training methods have the following shortcomings: First, fixed training time: Traditional training is usually scheduled at a specific time, requiring users to attend according to a fixed schedule. This limits users' self-study time and makes it difficult to meet the learning needs of different users. Second, limited technical skills assessment: Technical skills assessments are usually conducted only once at a specific time, and the assessment methods are relatively simple, failing to comprehensively evaluate users' learning progress and skill improvement throughout the training process. Summary of the Invention
[0004] To address the problems existing in the prior art, this invention provides a training system for intelligent training of laboratory animal handlers based on AI technology. The system includes: an interaction module for displaying the training system interface via a display device and receiving user login and learning operation instructions; a data acquisition module for sending training tasks to users through multi-dimensional evaluation methods and collecting user-input learning operation instructions in real time; a data transmission module for sending learning operation instructions to a data processing module; a data processing module for analyzing user learning operation instructions, generating response text, and generating a recommended learning list based on the response text; and an AI module for feeding back the response text generated by the data processing module and the corresponding recommended learning list to the user through the interaction module. This invention supports learning at any time and technical skills assessment, improving the flexibility of learning and the comprehensiveness of training.
[0005] The present invention adopts the following technical solution: a training system for intelligent training of laboratory animal administrators based on AI technology, comprising: an interaction module, a data acquisition module, a data transmission module, a data processing module, and an AI module; The interactive module is used to display the training system interface through a display device and to receive user login instructions and learning operation instructions. The data acquisition module is used to send training tasks to users through multi-dimensional evaluation methods and to collect the learning operation instructions input by users according to the training tasks in real time. The data transmission module is used to send the learning operation instructions collected by the data acquisition module to the data processing module; The data processing module is used to analyze and generate answer text based on the user's learning operation instructions, and generate a recommended learning list based on the answer text. The AI module is used to feed back the answer text generated by the data processing module and the corresponding recommended learning list to the user through the interaction module.
[0006] Furthermore, the system also includes a manual verification module, which is used to display the current user's training results through the interactive module and receive the user's signature confirmation instruction for the training results.
[0007] Furthermore, the data processing module includes: a preprocessing unit, a base large model unit, and a fine-tuning unit; The preprocessing unit is used to generate factor information based on the learning operation instructions collected by the data acquisition module; The base large model unit is used to process the factor information using a reasoning large model to generate multiple word sequence sequences; The fine-tuning unit is used to adjust the knowledge base of the reasoning model based on the multiple word sequence using LoRA technology, and generate the answer text.
[0008] Furthermore, the knowledge base includes: the Regulations on the Management of Laboratory Animals, AAALAC certification standards, SOPs, animal ethics, animal welfare regulations, animal behavior, operational skills, pharmacology, pharmacodynamics, pharmacokinetics, the Guidelines for Non-Clinical Safety Evaluation, data collection, experimental design, and interpretation of experimental data.
[0009] Furthermore, the multi-dimensional assessment methods include: AI questionnaires, simulated operations, and knowledge tests.
[0010] Furthermore, the learning operation instructions include voice operation instructions and image operation instructions.
[0011] Furthermore, the preprocessing unit is used to generate factor information based on the learning operation instructions collected by the data acquisition module, specifically including: The preprocessing unit performs noise reduction on the voice operation commands and converts the noise-reduced voice operation commands into spectral feature vectors through the voice spectrogram; it calculates the probability of each spectral feature vector on the acoustic feature according to the acoustic model to obtain the factor information corresponding to the voice operation commands. The preprocessing unit performs noise reduction on the image operation instructions and extracts features from the noise-reduced image operation instructions through a convolutional neural network to obtain the user operation posture in the image operation instructions. The user operation posture is then used as the factor information corresponding to the image operation instructions.
[0012] This invention further proposes a training method for intelligent training of laboratory animal handlers based on AI technology, applicable to any of the aforementioned training systems for intelligent training of laboratory animal handlers based on AI technology. The method includes: The system receives user login and learning operation instructions through the training system interface. Training tasks are sent to users based on a multi-dimensional evaluation method, and the learning operation instructions input by users according to the training tasks are collected in real time. The acquired learning operation instructions are transmitted to the data processing unit; The data processing unit analyzes the learning operation instructions, generates response text for the training task, and generates a recommended learning list based on the response text. The AI module feeds back the answer text and recommended learning list to the training system interface.
[0013] The beneficial effects of this invention are as follows: The training system proposed in this invention can ensure that laboratory animal administrators master the correct feeding, care, and experimental operation skills, reduce animal suffering and stress caused by improper operation, and improve the welfare of laboratory animals; through the AI module, it can quickly analyze user data, accurately locate the user's learning needs, and provide users with personalized training content, avoiding the "one-size-fits-all" teaching model in traditional training, enabling users to learn more efficiently and saving training time and energy; through multi-dimensional analysis methods such as image recognition and voice recognition, it can comprehensively and objectively evaluate the user's operational skills and knowledge mastery, promptly identify user problems and provide targeted guidance, helping users better master the various professional skills required for laboratory animal administrators and improve the quality of training. Attached Figure Description
[0014] To more clearly illustrate the technical solutions in the embodiments of the present 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 only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0015] Figure 1 This is a schematic diagram of a training system for intelligent training of laboratory animal handlers based on AI technology, according to an embodiment of the present invention. Figure 2 This is a schematic diagram of a training method for intelligent training of laboratory animal administrators based on AI technology, according to an embodiment of the present invention. Detailed Implementation
[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0017] A schematic diagram of a training system for intelligent training of laboratory animal handlers based on AI technology according to an embodiment of the present invention is shown below. Figure 1 As shown, it includes: an interaction module, a data acquisition module, a data transmission module, a data processing module, and an AI module; In this embodiment of the invention, the interaction module is used to display the training system interface through a display device and receive user login instructions and learning operation instructions; the data acquisition module is used to send training tasks to users through multi-dimensional evaluation methods and collect learning operation instructions input by users according to the training tasks in real time; the data transmission module is used to send the learning operation instructions collected by the data acquisition module to the data processing module; the data processing module is used to analyze and generate answer text according to the user's learning operation instructions and generate a recommended learning list according to the answer text; the AI module is used to feed back the answer text generated by the data processing module and the corresponding recommended learning list to the user through the interaction module. In another embodiment of the present invention, the training system further includes a manual verification module, which is used to display the current user's training results through the interactive module and receive the user's signature confirmation instruction for the training results.
[0018] In one specific embodiment of the present invention, the display device used in the interaction module can be a touch-enabled display screen, and is also compatible with external input devices such as a mouse and keyboard; the system training interface first displays the login interface, so that users can access the training interface after registering and logging in. At the same time, users can select their own intelligent animation image to bind through the AI animation image setting function in the system, and the AI module can then interact with the user through the intelligent animation image.
[0019] In one specific embodiment of the present invention, the data acquisition module employs multi-dimensional evaluation methods such as AI questionnaires, simulated operations, and knowledge tests. AI questionnaires refer to surveys designed using AI technology and incorporating issues from key aspects of animal husbandry, care, and experimental procedures to intelligently collect and analyze relevant user ability information. Simulated operations involve collecting data on actual operations performed by users in simulated scenarios or environments to assess their operational and coping abilities. Knowledge tests involve setting a series of questions related to key aspects of animal husbandry, care, and experimental procedures to determine the user's level of knowledge mastery. By testing users through these multi-dimensional evaluation methods, the user's actions in response to the tests are obtained as learning instructions for further analysis.
[0020] In one specific embodiment of the present invention, the learning operation instructions include voice operation instructions and image operation instructions. In the system's preset key aspects of animal husbandry, care, and experimental operations, the voice operation instructions can be explanations of animal anesthesia principles, including key knowledge points such as the mechanism of action of anesthetic drugs and dosage calculations. In this embodiment, the similarity is measured by calculating the cosine similarity between the user's explanation and the preset knowledge points. The similarity value ranges from 0 to 1, with a value closer to 1 indicating a higher similarity. A similarity threshold of 0.8 is set. If the similarity between the user's explanation and the knowledge point is greater than or equal to 0.8, the explanation is considered accurate; if it is lower than this threshold, the explanation is considered to have deviations or omissions and needs improvement. The image operation instructions can be the posture of holding the needle during injection, the gesture of grasping the animal, the scissor operation gesture, etc., extracting specific steps of animal experimental operations, animal organ separation photos, animal wound size, and key areas of bleeding.
[0021] In a specific embodiment of the present invention, the data processing module includes: a preprocessing unit, a base large model unit, and a fine-tuning unit; wherein, the preprocessing unit is used to generate factor information based on the learning operation instructions collected by the data acquisition module; the base large model unit is used to process the factor information using a reasoning large model to generate multiple word sequence sequences. In this embodiment of the present invention, the reasoning large model can be the DeepSeek R1-0528 model; the fine-tuning unit is used to adjust the knowledge base of the reasoning large model using LoRA technology based on the multiple word sequence sequences to generate response text. Taking injection operation training as an example, the response text generated by the fine-tuning unit can be the injection operation steps, including: preparing the syringe and drug; correctly holding the syringe to ensure the needle is stable; locating the injection site of the animal and slowly inserting the needle at a 45-degree angle, slowly injecting the drug, and observing the animal's reaction at the same time; after the injection is completed, quickly withdrawing the needle; recording the injection dosage and animal reaction for subsequent observation.
[0022] In one specific embodiment of the present invention, the LoRA technology is used to adjust the knowledge base of the large inference model by inserting a low-rank matrix into the key weight matrix of the model. This allows for parameter-efficient adjustment of the model. This fine-tuning method not only enables the model to more accurately understand and generate the language content of laboratory animal administrators, thereby significantly improving the accuracy of the system in identifying laboratory animal administrator-related issues, but also ensures the efficiency of the fine-tuning process and the stability of the model performance. The adjusted knowledge base includes at least the following: the "Regulations on the Management of Laboratory Animals", AAALAC certification standards, SOPs, animal ethics, animal welfare regulations, animal behavior, operational skills, pharmacology, pharmacodynamics, pharmacokinetics, "Guidelines for Non-Clinical Safety Evaluation", collected data, experimental design, and interpretation of experimental data.
[0023] In one specific embodiment of the present invention, the preprocessing unit is used to generate factor information based on the learning operation instructions collected by the data acquisition module. Specifically, the preprocessing unit performs noise reduction processing on the voice operation instructions and converts the noise-reduced voice operation instructions into spectral feature vectors through a voice spectrogram; calculates the probability of each spectral feature vector on the acoustic features according to the acoustic model to obtain the factor information corresponding to the voice operation instructions; the preprocessing unit performs noise reduction processing on the image operation instructions and extracts features from the noise-reduced image operation instructions through a convolutional neural network to obtain the user operation posture in the image operation instructions, and uses the user operation posture as the factor information corresponding to the image operation instructions.
[0024] In another embodiment of the present invention, the present invention also proposes a training method for laboratory animal handlers based on AI technology applied to the above-mentioned training system, the flowchart of which is shown below. Figure 2 As shown, the method includes: The system receives user login and learning operation instructions through the training system interface; sends training tasks to users based on multi-dimensional evaluation methods, and collects the learning operation instructions input by users according to the training tasks in real time; transmits the collected learning operation instructions to the data processing unit; generates answer text for the training tasks through the data processing unit, and generates a recommended learning list based on the answer text; and feeds back the answer text and recommended learning list to the training system interface through the AI module.
[0025] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A training system for intelligent training of laboratory animal handlers based on AI technology, characterized in that, include: The module includes an interaction module, a data acquisition module, a data transmission module, a data processing module, and an AI module. The interactive module is used to display the training system interface through a display device and to receive user login instructions and learning operation instructions. The data acquisition module is used to send training tasks to users through multi-dimensional evaluation methods and to collect the learning operation instructions input by users according to the training tasks in real time. The data transmission module is used to send the learning operation instructions collected by the data acquisition module to the data processing module; The data processing module is used to analyze and generate answer text based on the user's learning operation instructions, and generate a recommended learning list based on the answer text. The AI module is used to feed back the answer text generated by the data processing module and the corresponding recommended learning list to the user through the interaction module.
2. The training system for intelligent training of laboratory animal handlers based on AI technology according to claim 1, characterized in that: The system also includes a manual verification module, which is used to display the current user's training results through the interactive module and receive the user's signature confirmation instruction for the training results.
3. The training system for intelligent training of laboratory animal handlers based on AI technology according to claim 1, characterized in that: The data processing module includes: a preprocessing unit, a base large model unit, and a fine-tuning unit; The preprocessing unit is used to generate factor information based on the learning operation instructions collected by the data acquisition module; The base large model unit is used to process the factor information using a reasoning large model to generate multiple word sequence sequences; The fine-tuning unit is used to adjust the knowledge base of the reasoning model based on the multiple word sequence using LoRA technology, and generate the answer text.
4. The training system for intelligent training of laboratory animal handlers based on AI technology according to claim 3, characterized in that: The knowledge base includes: the Regulations on the Management of Laboratory Animals, AAALAC certification standards, SOPs, animal ethics, animal welfare regulations, animal behavior, operational skills, pharmacology, pharmacodynamics, pharmacokinetics, the Guidelines for Non-Clinical Safety Evaluation, data collection, experimental design, and interpretation of experimental data.
5. The training system for intelligent training of laboratory animal handlers based on AI technology according to claim 1, characterized in that: The multi-dimensional assessment methods include: AI questionnaires, simulation operations, and knowledge tests.
6. The training system for intelligent training of laboratory animal handlers based on AI technology according to claim 1, characterized in that: The learning operation instructions include voice operation instructions and image operation instructions.
7. The training system for intelligent training of laboratory animal handlers based on AI technology according to claim 6, characterized in that: The preprocessing unit is used to generate factor information based on the learning operation instructions collected by the data acquisition module, specifically including: The preprocessing unit performs noise reduction on the voice operation commands and converts the noise-reduced voice operation commands into spectral feature vectors through the voice spectrogram; it calculates the probability of each spectral feature vector on the acoustic feature according to the acoustic model to obtain the factor information corresponding to the voice operation commands. The preprocessing unit performs noise reduction on the image operation instructions and extracts features from the noise-reduced image operation instructions through a convolutional neural network to obtain the user operation posture in the image operation instructions. The user operation posture is then used as the factor information corresponding to the image operation instructions.
8. A training method for intelligent training of laboratory animal handlers based on AI technology, applied to the training system for intelligent training of laboratory animal handlers based on AI technology as described in any one of claims 1-7, characterized in that, include: The system receives user login and learning operation instructions through the training system interface. Training tasks are sent to users based on a multi-dimensional evaluation method, and the learning operation instructions input by users according to the training tasks are collected in real time. The acquired learning operation instructions are transmitted to the data processing unit; The data processing unit analyzes the learning operation instructions, generates response text for the training task, and generates a recommended learning list based on the response text. The AI module feeds back the answer text and recommended learning list to the training system interface.