Construction and application of AI-based antibiotic sensor chip design model
Optimizing the design of metamaterial antibiotic sensing chips through AI-based deep learning models, solving the problems of complexity and experience dependence on traditional design processes, and achieving fast and efficient antibiotic sensing effects in the terahertz band, especially for chloromycin detection.
Patent Information
- Application Number
- CN202411307850.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-19
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-09-19
AI Technical Summary
The design process of traditional metamaterial sensors is complex and dependent on experience, making it difficult to fully explore the design space, resulting in limited optimization possibilities and difficult to quickly find the best structure under the limitation of computing resources.
Using AI-based antibiotic sensing chip design model, the improved deep learning model is used to explore the implicit relationship between the physical structural parameters and electromagnetic response of terahertz metamaterial antibiotic sensing chip, and the design framework is optimized through the transformer model to quickly design it Efficient and accurate structural parameters.
The design speed and accuracy of terahertz metamaterial antibiotic sensing chips are improved, and the function of detecting antibiotics in the terahertz frequency band is realized, especially ceromycin sensing detection, and the design process relies more on system methods than empirical intuition.
Smart Images

Figure CN118821357B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of deep learning, and in particular to a method for constructing and applying an AI-based antibiotic sensor chip design model. Background Art
[0002] Metamaterial sensors are a new type of sensor that utilizes the properties of metamaterials. Electromagnetic induced transparency (EIT) is a quantum optical phenomenon that allows certain frequencies of light to propagate through a medium without being absorbed, thus forming a "transparent window". In metamaterial sensors, the EIT phenomenon is used to enhance the sensitivity and resolution of the sensor. By designing artificial structures with specific geometric shapes and arrangements, the EIT phenomenon can be simulated at specific frequencies. Such metamaterial sensors can be used to detect very weak signal changes because the EIT phenomenon can greatly increase the sensitivity of the medium to certain specific frequencies of light. If a metamaterial sensor is designed to simulate the EIT phenomenon in the terahertz (THz) frequency band, a terahertz metamaterial sensor can be designed.
[0003] The structural design of metamaterial sensors often requires a rather complex design process and design cycle. The traditional calculation method starts with certain initial conditions and boundary conditions, and uses computational electromagnetic simulation to solve the discrete Maxwell equations in space and time. By setting sufficient grids and iterative steps, the optical properties of a given structure can be accurately calculated. It is often necessary to fine-tune the geometry and iteratively perform simulations to gradually approach the target response. This process relies heavily on past experience with the design template, and due to the limitations of simulation capabilities and time, only limited design parameters are adjusted when searching for the optimal structure. The above defects lead to the need for designers to repeatedly simulate and experimentally verify when designing metamaterial sensors that respond in a specific frequency band. Each iteration requires recalculating the electromagnetic field distribution, which is particularly time-consuming when computing resources are limited, and due to the limitations of simulation time and resources, only limited design parameters can be adjusted. This means that it may not be possible to fully explore all possible design options, thereby limiting the possibility of optimization; in addition, the design process is highly dependent on the designer's experience and intuition, especially when selecting the initial design template. There is a lack of systematic methods to guide designers on how to effectively explore the design space, which may lead to missing the optimal solution. Summary of the invention
[0004] The embodiment of the present application provides a method for constructing and applying an AI-based antibiotic sensor chip design model, which uses an improved deep learning model to explore the implicit relationship between the physical structure parameters of the metamaterial antibiotic sensor chip and its electromagnetic response, so as to assist designers in the on-demand rapid design of antibiotic sensor chips.
[0005] In a first aspect, an embodiment of the present application provides a method for constructing an antibiotic sensor chip design model based on AI, comprising:
[0006] Used to design a terahertz metamaterial antibiotic sensor chip, wherein the upper structure of the terahertz metamaterial antibiotic sensor chip includes a circular open resonant ring and a rectangular closed loop arranged at intervals, and the opening of the circular open resonant ring is arranged opposite to the rectangular closed loop;
[0007] A transmission curve sampling point set corresponding to a terahertz metamaterial antibiotic sensor chip with different known structural parameter groups is taken as training data, wherein the known structural parameter group includes an opening width of a circular open resonant ring, a radius of the circular open resonant ring, a ring width of the circular open resonant ring, a width of a rectangular closed loop, a length of the rectangular closed loop, and a ring width of the rectangular closed loop;
[0008] Input the training data into the antibiotic sensor chip design framework for training until the training conditions are met, wherein the antibiotic sensor chip design framework includes an embedding module, an encoding module and a decoding module connected in sequence, the embedding module includes a block embedding layer, a linear injection layer and a position encoding layer, the encoding module includes a plurality of encoding heads connected in sequence, each encoding head includes a multi-head self-attention layer and a feedforward network, a normalization layer is added before each multi-head self-attention layer and each feedforward network, and a residual connection is applied after each multi-head self-attention layer and each feedforward network, and the decoding module includes three linear fully connected layers connected in sequence;
[0009] The transmission curve sampling point set is input into the block embedding unit and processed into blocks to obtain multiple blocks. Each block contains multiple transmission curve sampling points. After each block passes through the linear injection layer, the position code in the position coding layer is added to obtain the input feature; the input feature is input into the encoding module for encoding processing to obtain the encoding feature, and the encoding feature is input into the three-layer linear fully connected layer of the decoding module to output the structure parameter group.
[0010] In a second aspect, an embodiment of the present application provides an application method of an AI-based antibiotic sensor chip design model, inputting a set of transmission curve sampling points into an AI-based antibiotic sensor chip design model obtained by a method for constructing an AI-based antibiotic sensor chip design model, and outputting corresponding structural parameters.
[0011] In a third aspect, an embodiment of the present application provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute a method for constructing an AI-based antibiotic sensor chip design model.
[0012] In a fourth aspect, an embodiment of the present application provides a readable storage medium, in which a computer program is stored. The computer program includes a program code for controlling a process to execute a process, and the process includes the method for constructing the AI-based antibiotic sensor chip design model.
[0013] The main contributions and innovations of the present invention are as follows:
[0014] The embodiment of the present application provides an AI-based antibiotic sensor chip design model, which improves the basic framework of the transformer model, optimizes the design of embedded modules, encoding modules and decoding models specifically for terahertz metamaterial antibiotic sensor chips, and can assist designers to quickly design terahertz metamaterial antibiotic sensor chips as needed, thereby improving the design speed and accuracy. In addition, the present solution designs a terahertz metamaterial antibiotic sensor chip that can detect antibiotics in the terahertz frequency band, and can realize the function of chlortetracycline sensing detection.
[0015] Details of one or more embodiments of the present application are set forth in the following drawings and description to make other features, objects, and advantages of the present application more readily apparent. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:
[0017] Figure 1 It is a schematic diagram of the structural framework of the terahertz metamaterial antibiotic sensor chip according to an embodiment of the present application.
[0018] Figure 2 It is a schematic diagram of the relationship between the refractive index of the medium above the sensor and the change of the transparent window.
[0019] Figure 3 It is a framework diagram of an artificial intelligence-based antibiotic sensor chip design framework according to an embodiment of the present application.
[0020] Figure 4 This is a graph of loss data during the training and testing phases of the AI-based antibiotic sensor chip design model according to an embodiment of the present application.
[0021] Figure 5 It is a schematic diagram comparing the transmission curve obtained by simulation with the transmission curve input by the test set.
[0022] Figure 6 It is a schematic diagram of the structure of an electronic device according to an embodiment of the present application.
[0023] In the figure: 10-upper structure, 11-circular open resonant ring, 111-opening, 12-rectangular closed loop, 20-lower substrate dielectric layer. DETAILED DESCRIPTION
[0024] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with one or more embodiments of this specification. Instead, they are merely examples of devices and methods consistent with some aspects of one or more embodiments of this specification as detailed in the appended claims.
[0025] It should be noted that: in other embodiments, the steps of the corresponding method are not necessarily performed in the order shown and described in this specification. In some other embodiments, the steps included in the method may be more or less than those described in this specification. In addition, a single step described in this specification may be decomposed into multiple steps for description in other embodiments; and multiple steps described in this specification may be combined into a single step for description in other embodiments.
[0026] Embodiment 1
[0027] This scheme provides a method for constructing an AI-based antibiotic sensor chip design model, including:
[0028] Used to design a terahertz metamaterial antibiotic sensor chip, wherein the upper structure of the terahertz metamaterial antibiotic sensor chip includes a circular open resonant ring and a rectangular closed loop arranged at intervals, and the opening of the circular open resonant ring is arranged opposite to the rectangular closed loop;
[0029] A transmission curve sampling point set corresponding to a terahertz metamaterial antibiotic sensor chip with different known structural parameter groups is taken as training data, wherein the known structural parameter group includes an opening width of a circular open resonant ring, a radius of the circular open resonant ring, a ring width of the circular open resonant ring, a width of a rectangular closed loop, a length of the rectangular closed loop, and a ring width of the rectangular closed loop;
[0030] Input the training data into an artificial intelligence-based antibiotic sensor chip design framework for training until the training conditions are met, wherein the artificial intelligence-based antibiotic sensor chip design framework includes an embedding module, an encoding module, and a decoding module connected in sequence, the embedding module includes a block embedding layer, a linear injection layer, and a position encoding layer, the encoding module includes a plurality of encoding heads connected in sequence, each encoding head includes a multi-head self-attention layer and a feedforward network, a normalization layer is set before each multi-head self-attention layer, and a residual connection is applied after each multi-head self-attention layer and each feedforward network, and the decoding module includes three linear fully connected layers connected in sequence;
[0031] The transmission curve sampling point set is input into the block embedding unit and processed into blocks to obtain multiple blocks. Each block contains multiple transmission curve sampling points. After each block passes through the linear injection layer, the position code in the position coding layer is added to obtain the input feature; the input feature is input into the encoding module for encoding processing to obtain the encoding feature, and the encoding feature is input into the three-layer linear fully connected layer of the decoding module to output the structure parameter group.
[0032] This scheme selects the transformer network as the basic framework of the artificial intelligence-based antibiotic sensor chip design framework, and combines the characteristics of the terahertz metamaterial antibiotic sensor chip to optimize the embedded module, encoding module and decoding module of the artificial intelligence-based antibiotic sensor chip design framework to obtain an antibiotic sensor chip design model specifically for the terahertz metamaterial antibiotic sensor chip, aiming to design terahertz metamaterial antibiotic sensor chips on demand efficiently and with high quality.
[0033] It is well known that the transparent window of the EIT effect will change significantly with a slight change in the refractive index of the medium above the sensor. The terahertz metamaterial antibiotic sensor chip designed in this scheme is specifically used to detect antibiotics. Therefore, when the antibiotic solution to be tested is used as the upper medium, if the concentration of the antibiotic solution to be tested changes, the refractive index of the upper medium will change, and the corresponding transparent window of the sensor will also change. The relationship between the refractive index of the upper medium and the change of the transparent window is as follows: Figure 2 As shown by Figure 2It can also be seen that the refractive index of the upper medium varies from 1 to 1.4 with an interval of 0.1. With the linear change of the refractive index, the peak point of the transmittance of the transparent window and the frequency points corresponding to the left and right valleys are also changing regularly. As the refractive index changes by 0.4 from 1 to 1.4, the frequency value of the resonance peak of the transparent window also changes from 0.767 THz to 0.663 THz, red-shifting by 0.1 THz. This means that the resonance peak position under different antibiotic concentrations will change. Corresponding to different resonance peak positions, terahertz metamaterial antibiotics with different structural parameter groups need to be designed. Therefore, the AI-based antibiotic sensor chip design model proposed in this scheme can quickly design different structural parameter groups according to the required resonance peak. This scheme uses the transmission curve sampling point set on the transmission curve to characterize the required resonance peak.
[0034] The structure of the terahertz metamaterial antibiotic sensor chip used in this scheme is as follows Figure 1 As shown, the terahertz metamaterial antibiotic sensor chip proposed in this scheme includes an upper structure 10 and a lower substrate dielectric layer 20 arranged from top to bottom, wherein a circular open resonant ring 11 and a rectangular closed loop 12 are arranged at intervals on the upper structure 10, and the opening 111 of the circular open resonant ring 11 is arranged opposite to the rectangular closed loop 12, and the lower substrate dielectric layer is a substrate material with a low dielectric constant, which has the advantage of enhancing the sensing performance of the metamaterial. Specifically, the circular open resonant ring 11 provided in this scheme has a uniform ring width and a uniform radius; the ring width of the rectangular closed loop 12 is also uniformly set, and the opening 111 of the circular open resonant ring 11 is arranged opposite to the central axis position of the rectangular closed loop 12. The circular open resonant ring designed in this scheme can provide rich electromagnetic responses. When resonance occurs, a large amount of charge gathers at the opening of the circular open resonant ring to form an enhanced electric field, which is easier to couple with the rectangular closed loop to produce a transparent window, and interference destructive interaction occurs between the circular open resonant ring and the rectangular closed loop to produce a transparent window resonance peak, which meets the requirements of sensing detection.
[0035] In some embodiments, the circular open resonant ring 11 and the rectangular closed ring 12 are both made of copper metal, and the electrical conductivity is 5.96×10 7 S / m, with a thickness of 0.2µm; the lower substrate dielectric layer is a low-refractive-index quartz (SiO2) material with a relative dielectric constant of 1.9 and a thickness of 40µm.
[0036] In some embodiments, the widths of the superstructure in the x and y directions are denoted by Px and Py, respectively, Px=180µm, Py=180µm.
[0037] In some embodiments, the present solution specifically provides a terahertz metamaterial antibiotic sensor chip for chlortetracycline antibiotic detection, so the present solution specifically designs the structural dimensions of the structural parameter group of the terahertz metamaterial antibiotic sensor chip so that the transmission curve resonance peak of the terahertz metamaterial antibiotic sensor chip is close to the absorption peak position of chlortetracycline in the terahertz band. It should be noted that the structural parameter group of the terahertz metamaterial antibiotic sensor chip can have a slight change within a reasonable range to ensure that the resonance peak frequency generated by it can match the absorption peak of chlortetracycline.
[0038] The structural parameter group referred to and required to be designed in this scheme includes the opening width of the circular open resonant ring, the radius of the circular open resonant ring, the ring width of the circular open resonant ring, the width of the rectangular closed loop, the length of the rectangular closed loop and the ring width of the rectangular closed loop. Figure 1 As shown in the figure, the opening width of the circular open resonant ring, the radius of the circular open resonant ring, the ring width of the circular open resonant ring, the width of the rectangular closed loop, the length of the rectangular closed loop and the ring width of the rectangular closed loop are respectively expressed as g, r, m, w, l and s. The AI-based antibiotic sensor chip design model proposed in this scheme can be reversely designed according to the transmission curve sampling point set to obtain a reasonable structural parameter group of the terahertz metamaterial antibiotic sensor chip.
[0039] Regarding the construction of the design model, this scheme takes the transmission curve sampling point set corresponding to the terahertz metamaterial antibiotic sensor chip with different known structural parameter groups as training data. Specifically, the terahertz metamaterial antibiotic sensor chip with different known structural parameter groups is used for bionics and data collection to obtain the transmission curve, and 201 points are evenly selected on the transmission curve as the transmission curve sampling point set. In other words, the transmission curve sampling point set corresponding to the terahertz metamaterial antibiotic sensor chip with each known structural parameter group is 201 points evenly selected on the transmission curve.
[0040] In some embodiments, the present solution obtains training data through electromagnetic simulation software, and performs simulation using a structural parameter group (g, r, m, w, l, s) as a variable to complete the collection of training data.
[0041] It should be noted that since the data on the transmission curve has a lot of redundant information, if all the redundant information is analyzed and processed, the calculation amount of the AI-based antibiotic sensor chip design model will be too large and the results will be subject to more interference. Therefore, this scheme specifically adopts the transformer network as the basic framework of the artificial intelligence-based antibiotic sensor chip design framework, and optimizes and improves the self-attention mechanism of the multi-head self-attention layer in the encoding module in the artificial intelligence-based antibiotic sensor chip design framework to pay more attention to the vicinity of the transmission window, and the position encoding unit in the embedded module can also merge the position information of the transmission curve with the curve, so that the optimized design model can be designed on demand for the transmission curve sampling point set of the transmission curve to obtain a terahertz metamaterial antibiotic sensor chip that meets the needs.
[0042] Figure 3 It is a structural schematic diagram of the artificial intelligence-based antibiotic sensor chip design framework provided by the present scheme. The artificial intelligence-based antibiotic sensor chip design framework includes an embedding module, an encoding module and a decoding module connected in sequence. Specifically, the embedding module of the present scheme includes a block embedding layer, a linear injection layer and a position encoding layer. The block embedding layer divides 201 transmission curve sampling points in the input transmission curve sampling point set into 67 blocks, each block contains 3 transmission curve sampling points, and each block is added with the position encoding of the position encoding layer after passing through the linear injection layer to obtain an input feature, which can be input into the encoding module for further encoding processing.
[0043] In some embodiments, the position coding in the position coding layer of the present solution includes the spatial position information of the transmission curve sampling points, that is, the order of each transmission curve sampling point. If there is no position coding, it means that the 201 transmission curve sampling points are input out of order and the frequency information is lost. The position coding layer of the present solution initially generates position coding randomly, and continuously learns and updates the position coding during the training iteration process of the design framework, so that the block can obtain input features that focus more on the transmission window after adding the position coding in the linear injection layer.
[0044] The encoding module of the present solution includes a plurality of encoding heads connected in sequence, each encoding head includes a multi-head self-attention layer and a feedforward network, a normalization layer is provided before each multi-head self-attention layer, and a residual connection is applied after each multi-head self-attention layer and each feedforward network. Specifically, the features output by the multi-head self-attention layer in each encoding head are residually connected with the features input to the multi-head self-attention layer, the features of the residual connection are input to the feedforward network after being activated by an activation function, the features output by the feedforward network in each encoding head are residually connected with the features input to the feedforward network, and the features of the residual connection are output after being activated by an activation function.
[0045] Specifically, the feature input into the encoding head is processed by the normalization layer to obtain the first feature, the first feature is input into the multi-head self-attention layer to output the second feature, the second feature is residually connected with the first feature to obtain the third feature, the third feature is activated by the activation function to obtain the fourth feature, the fourth feature is input into the feedforward network to output the fifth feature, the fifth feature is residually connected with the fourth feature and activated by the activation function to obtain the output feature of the current encoding head.
[0046] In some embodiments, the activation function uses a Gaussian error linear unit.
[0047] In some embodiments, the feedforward network employs a multilayer perceptron.
[0048] The purpose of setting the normalization layer in the encoding head of this scheme is mainly to adjust the scale of the input data to reduce the problem of gradient vanishing or gradient exploding, thereby improving the stability and training efficiency of the AI-based antibiotic sensor chip design model. The residual connection is used to solve the problem of gradient vanishing and weight matrix degradation.
[0049] In view of the data characteristics of the transmission curve sampling points, this scheme specifically places the normalization layer before the multi-head self-attention layer. The advantage of this design is that it can better adapt to the data characteristics of the transmission curve sampling points. In view of the fact that the transmission curve sampling point data usually has a large scale variation, this scheme places the normalization layer between the multi-head self-attention layers to effectively adjust the scale of the input features. The advantage of this design is that the AI-based antibiotic sensor chip design model can better adapt to and process the transmission curve sampling point data, thereby showing better performance in the prediction task of the design model.
[0050] In addition, since the AI-based antibiotic sensor chip design model of this scheme only needs to output 6 structural parameters instead of complex data, the decoding module of this scheme only needs to include three linear fully connected layers connected in sequence, and the features input into the decoding module are decoded and mapped in the decoding module to obtain output features.
[0051] Since the AI-based antibiotic sensor chip design model of this scheme only needs to output numerical structural parameters, the training of the design framework of this scheme uses mean square error as the loss function, and the corresponding loss function is optimized by using AdamOptimizer.
[0052] In a specific embodiment, 4096 sets of training data are obtained by simulation, and 80% are taken as training sets and 20% are taken as test sets. Each set of training data includes a known structural parameter set and a transmission curve sampling point set consisting of 201 transmission curve sampling points on the corresponding transmission curve. This scheme trains the design framework with training data in batches, with 32 data in each batch. The training and test losses of the design model are as follows: Figure 4 As shown, Figure 4 (a) in the figure represents the training set loss data, and (b) represents the test set loss. It can be seen that the training set loss converges to 0.039 after 1200 cycles, and the validation set loss converges to 0.051 after 1600 cycles. A stable and low loss has been achieved at 600 cycles. Compared with fixed position encoding, learnable position encoding converges more slowly, but has lower training and validation losses, which means better design capabilities for metamaterial sensors.
[0053] In some embodiments, the present solution sets the loss function to converge to 0.039 to complete the training of the design model.
[0054] Embodiment 2
[0055] Based on the same concept, this application also proposes an application method of an AI-based antibiotic sensor chip design model, including:
[0056] The transmission curve sampling point set is input into the AI-based antibiotic sensor chip design model constructed in Example 1, and the corresponding structural parameters are output.
[0057] In order to verify the effect of the AI-based antibiotic sensor chip design model in designing terahertz metamaterial antibiotic sensor chips, this scheme randomly selects 4 transmission curve sampling points corresponding to the test set from the test set and inputs them into the design model. The output of 4 sets of simulated structural parameter group data tables are shown in Table 1. The 4 designed structural parameter groups are: [20.056, 38.182, 159.865, 49.921, 9.978, 8.921], [40.102, 30.079,164.872, 50.143, 8.081, 11.109], [24.262, 40.171, 169.973, 50.163, 10.212,9.993], [30.137, 33.406, 154.974, 44.860, 10.187, 9.072 ], the transmission curves obtained after simulation of these four sets of structural parameters are plotted together with the transmission curves of the test set input in the same coordinate system, and the four sets of transmission curves are obtained as follows Figure 5 As shown, Figure 5(a), (b), (c) and (d) represent four groups of transmission curves respectively. It can be seen that the simulation results of the designed structural parameters are in good agreement with the target curve, with the highest accuracy reaching 97.1%, which can well complete the design task of the target curve.
[0058] Table 1 Structural parameter group data table output by simulation
[0059]
[0060] Embodiment 3
[0061] This embodiment also provides an electronic device, referring to Figure 6 , including a memory 404 and a processor 402, wherein the memory 404 stores a computer program, and the processor 402 is configured to run the computer program to execute the steps in any of the above-mentioned embodiments of the method for constructing an AI-based antibiotic sensor chip design model.
[0062] Specifically, the processor 402 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.
[0063] Among them, the memory 404 may include a large capacity memory 404 for data or instructions. For example, but not limitation, the memory 404 may include a hard disk drive (HDD), a floppy disk drive, a solid state drive (SSD), a flash memory, an optical disk, a magneto-optical disk, a tape, or a universal serial bus (USB) drive, or a combination of two or more of these. In appropriate cases, the memory 404 may include a removable or non-removable (or fixed) medium. In appropriate cases, the memory 404 may be inside or outside the data processing device. In a specific embodiment, the memory 404 is a non-volatile memory. In a specific embodiment, the memory 404 includes a read-only memory (ROM) and a random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (Programmable Read-Only Memory, PROM for short), an erasable PROM (Erasable Programmable Read-Only Memory, EPROM for short), an electrically erasable PROM (Electrically Erasable Programmable Read-Only Memory, EEPROM for short), an electrically alterable ROM (Electrically Alterable Read-Only Memory, EAROM for short) or a flash memory (FLASH) or a combination of two or more of these. In appropriate circumstances, the RAM may be a static random access memory (SRAM) or a dynamic random access memory (DRAM), wherein the DRAM may be a fast page mode dynamic random access memory 404 (FPMDRAM), an extended data output dynamic random access memory (EDODRAM), a synchronous dynamic random access memory (SDRAM), etc.
[0064] The memory 404 may be used to store or cache various data files required for processing and / or communication, as well as possible computer program instructions executed by the processor 402 .
[0065] The processor 402 reads and executes the computer program instructions stored in the memory 404 to implement any one of the AI-based methods for constructing an antibiotic sensor chip design model in the above embodiments.
[0066] Optionally, the electronic device may further include a transmission device 406 and an input / output device 408 , wherein the transmission device 406 is connected to the processor 402 , and the input / output device 408 is connected to the processor 402 .
[0067] The transmission device 406 can be used to receive or send data via a network. Specific examples of the above-mentioned network may include a wired or wireless network provided by a communication provider of the electronic device. In one example, the transmission device includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 406 can be a radio frequency (Radio Frequency, referred to as RF) module, which is used to communicate with the Internet wirelessly.
[0068] The input / output device 408 is used to input or output information. In this embodiment, the input information may be a transmission curve sampling point set, etc., and the output information may be a structure parameter group, etc.
[0069] Optionally, in this embodiment, the processor 402 may be configured to perform the following steps through a computer program:
[0070] A transmission curve sampling point set corresponding to a terahertz metamaterial antibiotic sensor chip with different known structural parameter groups is taken as training data, wherein the known structural parameter group includes an opening width of a circular open resonant ring, a radius of the circular open resonant ring, a ring width of the circular open resonant ring, a width of a rectangular closed loop, a length of the rectangular closed loop, and a ring width of the rectangular closed loop;
[0071] Input the training data into an artificial intelligence-based antibiotic sensor chip design framework for training until the training conditions are met, wherein the artificial intelligence-based antibiotic sensor chip design framework includes an embedding module, an encoding module, and a decoding module connected in sequence, the embedding module includes a block embedding layer, a linear injection layer, and a position encoding layer, the encoding module includes a plurality of encoding heads connected in sequence, each encoding head includes a multi-head self-attention layer and a feedforward network, a normalization layer is set before each multi-head self-attention layer, and a residual connection is applied after each multi-head self-attention layer and each feedforward network, and the decoding module includes three linear fully connected layers connected in sequence;
[0072] The transmission curve sampling point set is input into the block embedding unit and processed into blocks to obtain multiple blocks. Each block contains multiple transmission curve sampling points. After each block passes through the linear injection layer, the position code in the position coding layer is added to obtain the input feature; the input feature is input into the encoding module for encoding processing to obtain the encoding feature, and the encoding feature is input into the three-layer linear fully connected layer of the decoding module to output the structure parameter group.
[0073] It should be noted that the specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementation modes, and this embodiment will not be described in detail here.
[0074] In general, various embodiments may be implemented in hardware or dedicated circuits, software, logic, or any combination thereof. Some aspects of the invention may be implemented in hardware, while other aspects may be implemented in firmware or software that may be executed by a controller, microprocessor, or other computing device, but the invention is not limited thereto. Although various aspects of the invention may be shown and described as block diagrams, flow charts, or using some other graphical representation, it should be understood that, as non-limiting examples, the boxes, devices, systems, techniques, or methods described herein may be implemented in hardware, software, firmware, dedicated circuits or logic, general-purpose hardware or controllers or other computing devices, or some combination thereof.
[0075] Embodiments of the present invention can be implemented by computer software, which is executable by a data processor of a mobile device, such as in a processor entity, or implemented by hardware, or implemented by a combination of software and hardware. Computer software or programs (also referred to as program products) including software routines, applets and / or macros can be stored in any device readable data storage medium, and they include program instructions for performing specific tasks. Computer program products can include one or more computer executable components configured to perform embodiments when the program is running. One or more computer executable components can be at least one software code or a part thereof. In addition, at this point, it should be noted that any box of the logic flow in the figure can represent a program step, or interconnected logic circuits, boxes and functions, or a combination of program steps and logic circuits, boxes and functions. Software can be stored in physical media such as memory chips or storage blocks implemented in processors, magnetic media such as hard disks or floppy disks, and optical media such as, for example, DVDs and data variants thereof, CDs. Physical media are non-transient media.
[0076] Those skilled in the art should understand that the technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0077] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.
Claims
1. A method for constructing an antibiotic sensor chip design model based on AI, characterized in that: include: Used to design a terahertz metamaterial antibiotic sensor chip, wherein the upper structure of the terahertz metamaterial antibiotic sensor chip includes a circular open resonant ring and a rectangular closed loop arranged at intervals, and the opening of the circular open resonant ring is arranged opposite to the rectangular closed loop; A transmission curve sampling point set corresponding to a terahertz metamaterial antibiotic sensor chip with different known structural parameter groups is taken as training data, wherein the known structural parameter group includes an opening width of a circular open resonant ring, a radius of the circular open resonant ring, a ring width of the circular open resonant ring, a width of a rectangular closed loop, a length of the rectangular closed loop, and a ring width of the rectangular closed loop; Input the training data into an artificial intelligence-based antibiotic sensor chip design framework for training until the training conditions are met, wherein the artificial intelligence-based antibiotic sensor chip design framework includes an embedding module, an encoding module, and a decoding module connected in sequence, the embedding module includes a block embedding layer, a linear injection layer, and a position encoding layer, the encoding module includes a plurality of encoding heads connected in sequence, each encoding head includes a multi-head self-attention layer and a feedforward network, a normalization layer is set before each multi-head self-attention layer, and a residual connection is applied after each multi-head self-attention layer and each feedforward network, and the decoding module includes three linear fully connected layers connected in sequence; The transmission curve sampling point set is input into the block embedding unit and processed into multiple blocks. Each block contains multiple transmission curve sampling points. After each block passes through the linear injection layer, the position code in the position coding layer is added to obtain the input feature. The input features are input into the encoding module for encoding processing to obtain the encoded features, and the encoded features are input into the three-layer linear fully connected layer of the decoding module to output the structure parameter group.
2. The method for constructing an AI-based antibiotic sensor chip design model according to claim 1, characterized in that: The transmission curve sampling point set corresponding to the terahertz metamaterial antibiotic sensor chip of each known structural parameter group is 201 points evenly selected on the transmission curve.
3. The method for constructing an AI-based antibiotic sensor chip design model according to claim 2, characterized in that: The block embedding layer divides the 201 transmission curve sampling points in the input transmission curve sampling point set into 67 blocks, each block contains 3 transmission curve sampling points, and each block is added with the position encoding of the position encoding layer after passing through the linear injection layer to obtain the input feature.
4. The method for constructing an AI-based antibiotic sensor chip design model according to claim 1, characterized in that: The position coding in the position coding layer contains the spatial position information of the transmission curve sampling points. The position coding layer initially generates the position coding randomly and continuously learns and updates the position coding during the training iteration process of the design framework.
5. The method for constructing an AI-based antibiotic sensor chip design model according to claim 1, characterized in that: The features output by the multi-head self-attention layer in each encoding head are residually connected with the features input to the multi-head self-attention layer, and the residually connected features are activated by the activation function and input into the feedforward network. The features output by the feedforward network in each encoding head are residually connected with the features input to the feedforward network, and the residually connected features are activated by the activation function and output.
6. The method for constructing an AI-based antibiotic sensor chip design model according to claim 1, characterized in that: The feedforward network uses a multi-layer perceptron.
7. The method for constructing an AI-based antibiotic sensor chip design model according to claim 1, characterized in that: The terahertz metamaterial antibiotic sensor chip comprises an upper structure and a lower substrate dielectric layer arranged from top to bottom, wherein a circular open resonant ring and a rectangular closed loop are arranged at intervals on the upper structure, and the lower substrate dielectric layer is a substrate material with a low dielectric constant.
8. An application method of an AI-based antibiotic sensor chip design model, characterized in that: The transmission curve sampling point set is input into the AI-based antibiotic sensor chip design model obtained by the AI-based antibiotic sensor chip design model construction method described in any one of claims 1 to 7, and the corresponding structural parameters are output.
9. An electronic device comprising a memory and a processor, characterized in that: The memory stores a computer program, and the processor is configured to run the computer program to execute the method for constructing an AI-based antibiotic sensor chip design model according to any one of claims 1 to 7.
10. A readable storage medium, characterized in that: The readable storage medium stores a computer program, which includes a program code for controlling a process to execute a process, and the process includes a method for constructing an AI-based antibiotic sensor chip design model according to any one of claims 1 to 7.
