Method, apparatus, device, storage medium and product for data processing for semiconductor detection

Abstract

Provided are a data processing method, device, equipment, storage medium and product for semiconductor detection. The method provided herein includes: obtaining configuration information configured by a user, the configuration information indicating a set of industrial detection tasks to be performed, the set of industrial detection tasks being used for detecting a semiconductor sample; obtaining to-be-detected data of the semiconductor sample; based on the to-be-detected data, performing the set of industrial detection tasks to obtain a set of confidence degrees, the set of confidence degrees being associated with a set of detection results of the set of industrial detection tasks; in response to the set of confidence degrees satisfying a preset condition, obtaining feedback information input by the user; and providing the feedback information to a set of models corresponding to the set of industrial detection tasks as training samples of the set of models. In this way, embodiments of the present disclosure can determine a set of industrial detection tasks to be performed based on the configuration information of the user, so that the entire detection process can be more flexible and efficient.

Description

Technical Field

[0001] This article relates to the field of semiconductor testing and data processing technology, and in particular to data processing methods, apparatus, equipment and storage media for semiconductor testing. Background Technology

[0002] In the semiconductor manufacturing process, semiconductor devices undergo hundreds of precision processes. Any micron-level particle contamination, scratches, or etching abnormalities can lead to performance failure of semiconductor devices or even the scrapping of the entire batch. Traditional quality inspection relies heavily on manual visual inspection, which is not only inefficient and has a high rate of missed detections, but also struggles to cope with increasingly complex microscopic defects and changing production environments. Summary of the Invention

[0003] In a first aspect, a data processing method for semiconductor inspection is provided. The method includes: acquiring configuration information configured by a user, the configuration information indicating a set of industrial inspection tasks to be executed, the set of industrial inspection tasks being used to inspect semiconductor samples; acquiring inspection data of the semiconductor samples; executing the set of industrial inspection tasks based on the inspection data to obtain a set of confidence scores, the set of confidence scores being associated with a set of inspection results of the set of industrial inspection tasks; in response to the set of confidence scores satisfying preset conditions, acquiring user-inputted feedback information; and providing the feedback information to a set of models corresponding to the set of industrial inspection tasks as training samples for the set of models.

[0004] In a second aspect, an apparatus for data processing in semiconductor inspection is provided. The apparatus includes: a first acquisition module configured to acquire configuration information configured by a user, the configuration information indicating a set of industrial inspection tasks to be executed, the set of industrial inspection tasks being used to inspect semiconductor samples; a second acquisition module configured to acquire inspection data of the semiconductor samples; a first execution module configured to execute the set of industrial inspection tasks based on the inspection data to obtain a set of confidence scores, the set of confidence scores being associated with a set of inspection results of the set of industrial inspection tasks; a third acquisition module configured to acquire user-inputted feedback information in response to a set of confidence scores satisfying preset conditions; and a providing module configured to provide the feedback information to a set of models corresponding to the set of industrial inspection tasks as training samples for the set of models.

[0005] In a third aspect, an electronic device is provided. The device includes at least one processor; and at least one memory coupled to the at least one processor and storing instructions for execution by the at least one processor. When executed by the at least one processor, the instructions cause the device to perform the method of the first aspect.

[0006] In a fourth aspect, a computer-readable storage medium is provided. The computer-readable storage medium stores computer-executable instructions that can be executed by a processor to implement the method of the first aspect.

[0007] In a fifth aspect, a computer program product is provided, which is tangibly stored in a computer storage medium and includes computer-executable instructions that, when executed by a device, cause the device to perform the method of the first aspect.

[0008] It should be understood that the content described in this section is not intended to limit the key or important features of the examples in this article, nor is it intended to restrict the scope of the solution. Other features will become readily apparent from the following description. Attached Figure Description

[0009] The above and other features, advantages, and aspects of the various examples herein will become more apparent when taken in conjunction with the accompanying drawings and the following detailed description. In the accompanying drawings, the same or similar reference numerals denote the same or similar elements, wherein: Figure 1 A schematic diagram of the example environment is shown; Figure 2 Flowcharts of example data processing procedures for semiconductor testing are shown in some scenarios; Figure 3 Schematic block diagrams of example devices for data processing in semiconductor detection are shown in some scenarios; and Figure 4 A block diagram of an electronic device capable of implementing multiple illustrative scenarios is shown. Detailed Implementation

[0010] The examples in this document will now be described in more detail with reference to the accompanying drawings. While some examples are shown in the drawings, it should be understood that solutions can be implemented in various forms and should not be construed as limited to the examples presented herein. Rather, these examples are provided to provide a more thorough and complete understanding of the solutions. It should be understood that the drawings and examples in this document are for illustrative purposes only and are not intended to limit the scope of protection of the solutions.

[0011] It should be noted that the headings of any section / subsection provided herein are not restrictive. Various examples are described throughout this document, and examples of any type may be included under any section / subsection. Furthermore, examples described in any section / subsection may be combined in any way with any other examples described in the same section / subsection and / or different sections / subsections.

[0012] In the description of the examples in this document, the term "including" and similar terms should be understood as open inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "an example" or "the example" should be understood as "at least one example". The term "some examples" should be understood as "at least some examples". Other explicit and implicit definitions may also be included below. The terms "first", "second", etc., may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0013] The examples in this document may involve user data, data acquisition, and / or use. All of these aspects comply with relevant laws, regulations, and rules. In the examples, all data collection, acquisition, processing, manipulation, forwarding, and use are conducted with the user's knowledge and confirmation. Accordingly, when implementing each example, the type, scope of use, and usage scenarios of any data or information that may be involved should be communicated to the user and their authorization obtained through appropriate means, in accordance with relevant laws and regulations. The specific methods of notification and / or authorization can vary depending on the actual situation and application scenario; the scope of the solution is not limited in this regard.

[0014] In this manual and the sample solutions, any processing of personal information will be conducted only under legal grounds (such as obtaining the consent of the data subject or being necessary for the performance of a contract) and will only be carried out within the scope stipulated or agreed upon. A user's refusal to process personal information beyond what is necessary for basic functions will not affect the user's use of basic functions.

[0015] As mentioned above, semiconductor devices undergo hundreds of precision processes during semiconductor manufacturing. Any micron-level particle contamination, scratches, or etching abnormalities can lead to performance failure or even the scrapping of the entire batch. Traditional quality inspection relies heavily on manual visual inspection, which is not only inefficient and has a high rate of missed detections, but also struggles to cope with increasingly complex microscopic defects and changing production environments.

[0016] A data processing scheme for semiconductor inspection is proposed. The scheme includes: acquiring user-configured configuration information indicating a set of industrial inspection tasks to be executed, which are used to inspect semiconductor samples; acquiring the inspection data of the semiconductor samples; executing the set of industrial inspection tasks based on the inspection data to obtain a set of confidence scores, which are correlated with a set of inspection results of the set of industrial inspection tasks; acquiring user-input feedback information in response to the set of confidence scores meeting preset conditions; and providing the feedback information to a set of models corresponding to the set of industrial inspection tasks as training samples for the models.

[0017] In this way, embodiments of this disclosure can determine a set of industrial inspection tasks to be executed based on user configuration information, thereby making the entire inspection process more flexible and efficient. Furthermore, embodiments of this disclosure can obtain user feedback information after the confidence level meets preset conditions, achieving precise human-machine collaboration. Manual review is triggered only when there is uncertainty, ensuring inspection efficiency while continuously optimizing the model through feedback information, thereby improving the accuracy of data processing.

[0018] The following describes various examples of this scheme in further detail with reference to the accompanying drawings.

[0019] Example Environment Figure 1 A schematic diagram of example environment 100 is shown. (e.g.) Figure 1 As shown, example environment 100 may include electronic device 110.

[0020] In this example environment 100, electronic device 110 can communicate with inspection device / inspection system 120. Inspection system 120 can acquire user configuration information and execute corresponding industrial inspection tasks based on this information. The configuration information may include industrial inspection tasks configured by the user. Inspection system 120 can deploy one or more data analysis models for performing industrial inspection tasks, including classification models, rule analysis models, or statistical analysis models based on machine learning or deep learning. Inspection system 120 can be deployed within electronic device 110 or within a server. Inspection device can communicate with electronic device 110. The server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks, and big data and artificial intelligence platforms. The server may include, for example, computing systems / servers such as mainframes, edge computing nodes, computing devices in a cloud environment, etc.

[0021] exist Figure 1 In environment 100, if the detection system 120 is active, electronic device 110 can present interface 150 for supporting data processing for semiconductor detection. User 140 can configure industrial detection tasks and view the detection results of industrial detection tasks through interface 150.

[0022] In some cases, electronic device 110 can be any type of mobile terminal, fixed terminal, or portable terminal, including mobile phones, desktop computers, laptop computers, notebook computers, netbook computers, tablet computers, media computers, multimedia tablets, handheld computers, portable gaming terminals, VR / AR devices, personal communication system (PCS) devices, personal navigation devices, personal digital assistants (PDAs), audio / video players, digital cameras / camcorders, positioning devices, television receivers, radio receivers, e-book devices, gaming devices, or any combination thereof, including accessories and peripherals of these devices or any combination thereof. In some cases, electronic device 110 may also support any type of user-facing interface (such as "wearable" circuitry).

[0023] It should be understood that the structure and function of the various elements in environment 100 are described for illustrative purposes only and do not imply any limitation on the scope of the scheme.

[0024] The following description of the example will continue with reference to the accompanying drawings.

[0025] Example Process Figure 2 A flowchart of an example process 200 for data processing in semiconductor detection, based on certain scenarios, is shown. Process 200 can be implemented at the detection system 120. Reference will be made below. Figure 1 The process 200 is described using the example of a detection system 120 being deployed in an electronic device 110.

[0026] In box 210, electronic device 110 acquires configuration information configured by the user, indicating a set of industrial inspection tasks to be executed for inspecting semiconductor samples. As an example, electronic device 110 may present a configuration interface through which the user can input configuration information. For example, the user can input configuration information by entering codes. Alternatively, the user can input configuration information using appropriate controls in the local configuration interface. The configuration information can indicate a set of industrial inspection tasks to be executed. Such industrial inspection tasks can be used to inspect semiconductor samples, which may include wafers (silicon wafers, compound semiconductor wafers), semiconductor devices (chips, sensors, memory), semiconductor modules (packaged finished modules), semiconductor components (intermediates in the manufacturing process, tooling fixture contacts), etc. The industrial inspection tasks are used to identify, classify, or analyze defects, abnormal states, or process-related characteristics of the semiconductor samples.

[0027] In box 220, electronic device 110 acquires the test data of the semiconductor sample. As an example, electronic device 110 can acquire the test data of the semiconductor sample stored in a data storage unit. For example, a user can store image data (e.g., scanning electron microscope images), spectral data (e.g., energy dispersive spectroscopy data), etc., in the data storage unit, and electronic device 110 can retrieve the data in the data storage unit through an appropriate retrieval method. The test data can be acquired directly by industrial testing equipment, or it can be provided by a data storage system, laboratory analysis system, or production data management system. The industrial testing equipment mentioned here may include, but is not limited to: scanning electron microscope (SEM), optical microscope (OM), transmission electron microscope (TEM), focused ion beam (FIB), secondary ion mass spectrometry (SIMS), dynamic secondary ion mass spectrometry (DSIMS), energy dispersive spectrometer (EDS), etc. The specific acquisition process will be described below, and will not be elaborated upon here.

[0028] In some embodiments, the electronic device 110 can acquire identification information of a semiconductor sample. Further, the electronic device 110 can acquire test data corresponding to a set of industrial testing tasks based on the identification information. The identification information includes at least one of a sample identifier and time information of the semiconductor sample. The test data includes at least one of image data associated with the semiconductor sample, spectral data associated with the semiconductor sample, production parameter data associated with the semiconductor sample, and information about the production equipment associated with the semiconductor.

[0029] As an example, electronic device 110 can acquire identification information of a semiconductor sample, including a sample identifier and time information. The sample identifier indicates a unique number for the semiconductor sample, which can be used to distinguish different semiconductor samples. The time information indicates the production time and inspection time (e.g., scanning electron microscopy inspection) of the semiconductor sample, which is used to associate the sample with the corresponding production stage.

[0030] Furthermore, the electronic device 110 may include a preset data source mapping table. The data source mapping table can be used to record the storage location and association rules of different types of data to be tested (such as the cloud file path pattern of scanning electron microscope images, the database table name of EDS energy dispersive spectroscopy data, and the table path and column names of process parameters). The electronic device 110 can use sample identifiers and time information as key fields to query and obtain the data to be tested according to the rules of the data source mapping table. For example, based on an appropriate path pattern, combined with the year and month in the sample identifier and time information, the corresponding image data (e.g., scanning electron microscope images) can be obtained from the data storage unit. Another example is obtaining the corresponding spectral data (e.g., EDS energy dispersive spectroscopy data) from the data storage unit using the sample identifier. Yet another example is automatically filtering and obtaining the corresponding production parameter data such as temperature, pressure, and process duration, as well as information about the production equipment used to generate the semiconductor sample (e.g., the identification of the production equipment), based on the sample identifier and production time in a preset process parameter storage file (e.g., a table). Furthermore, the electronic device 110 can obtain the production equipment identifier associated with the semiconductor sample based on the sample identifier information, time information, or production batch information.

[0031] In some scenarios, when acquiring data to be detected, the electronic device 110 can also refer to the detection task to be performed. For example, when the detection task to be performed is determined to be the first task, the electronic device 110 can retrieve the data to be detected corresponding to the first task from the data storage unit. For instance, if the first task is a preliminary screening task, used to determine whether the semiconductor sample to be detected is in an abnormal state, the electronic device 110 can retrieve the scanning electron microscope image associated with the semiconductor sample required for the preliminary screening task from the data storage unit based on the identification information of the semiconductor sample. As another example, when the detection task to be performed is a root cause analysis task, the electronic device 110 can retrieve the production parameters of the semiconductor sample and the identification of the production equipment required for the root cause analysis task from the data storage unit based on the identification information of the semiconductor sample. The root cause analysis task is used to determine the production equipment that caused the semiconductor sample to be in an abnormal state.

[0032] In this way, embodiments of the present disclosure can obtain the data to be tested by querying based on the identification information of the semiconductor sample, thereby effectively improving the efficiency of obtaining the data to be tested.

[0033] In box 230, electronic device 110 performs a set of industrial inspection tasks based on the data to be inspected to obtain a set of confidence scores, which are associated with a set of inspection results from the set of industrial inspection tasks. As an example, electronic device 110 can sequentially execute each industrial inspection task according to the task order and dependencies indicated by the configuration information. Each inspection task is executed by a corresponding analysis engine (such as a deep learning model, a traditional machine learning model, or a rule engine) and outputs inspection results and corresponding confidence scores. The confidence scores are used to measure the reliability of the inspection results. In some scenarios, the industrial inspection tasks indicated by the configuration information may include only one task. When the industrial inspection tasks indicated by the configuration information include only one task, electronic device 110 can execute that task based on the data to be inspected to obtain the corresponding inspection results and confidence scores. For example, when the configuration information only indicates a preliminary screening task, electronic device 110 can provide a scanning electron microscope image of a semiconductor sample to the corresponding engine to obtain inspection results. Such inspection results are used to indicate whether the semiconductor sample is abnormal.

[0034] As an example, in a single-task configuration scenario, the user configures only one initial screening task. The initial screening task is suitable for production stages where defect patterns are relatively simple and the detection objectives are clearly defined. For instance, the user might want to quickly screen out wafers with obvious particle contamination on their surface to reduce the workload of manual visual inspection.

[0035] Electronic device 110 can acquire scanning electron microscope (SEM) images of the wafer sample to be inspected as inspection data. Then, electronic device 110 can invoke a preliminary screening model (e.g., a CNN model) specified in the configuration information to process the image. The preliminary screening model can output two results, which may include the detection result (e.g., "pass" or "fail") and the corresponding confidence score. For example, the preliminary screening model might output a confidence score of 0.95 for "fail".

[0036] Electronic device 110 can compare the confidence level with a preset threshold in the configuration information. A first threshold and a second threshold can be preset, where the first threshold represents a lower confidence level and the second threshold represents a higher confidence level. In one example, the first threshold can be 0.1 and the second threshold can be 0.9. When the confidence level corresponding to the detection result is higher than the second threshold, electronic device 110 can determine that the detection result has high reliability and directly output the detection result of the detection task. For example, when the confidence level is 0.95 and higher than the second threshold of 0.9, electronic device 110 can directly output "unqualified" as the detection result of the detection task. When the confidence level corresponding to the detection result is between the first and second thresholds, electronic device 110 can determine that the detection result has a certain degree of uncertainty and trigger a manual review process. For example, when the confidence level is 0.5 (between the first threshold of 0.1 and the second threshold of 0.9), electronic device 110 can mark the sample as "to be reviewed" and store its detection result and corresponding confidence level information.

[0037] In some scenarios, the configuration information can also indicate multiple industrial inspection tasks to be performed. The electronic device 110 can perform these multiple industrial inspection tasks according to the execution mode indicated in the configuration information. The specific process will be described below.

[0038] In some embodiments, configuration information may indicate a first execution mode. The first execution mode indicates that a set of industrial inspection tasks are executed according to a preset execution relationship. When the configuration information indicates the first execution mode, the electronic device 110 may execute the first industrial inspection task in a set of industrial inspection tasks based on the data to be inspected, in order to obtain a first inspection result.

[0039] As an example, the user can also configure multiple industrial inspection tasks and execution modes for these tasks. The electronic device 110 can determine the multiple industrial inspection tasks configured by the user from the configuration information and instruct a first inspection mode to execute the multiple industrial inspection tasks according to a preset execution relationship. This preset execution relationship may, for example, indicate the execution order and data transmission order of the multiple industrial inspection tasks.

[0040] After receiving the data to be tested, the electronic device 110 can determine the first industrial inspection task from multiple industrial inspection tasks according to a preset execution relationship. The first industrial inspection task can be, for example, a preliminary screening task. When executing the preliminary screening task, the electronic device 110 can provide the scanning electron microscope image from the data to be tested to a first model (e.g., a CNN model) corresponding to the preliminary screening task to obtain a first detection result. The first model can determine whether the semiconductor sample is in an abnormal state based on the scanning electron microscope image.

[0041] Furthermore, in response to a first detection result indicating that the semiconductor sample is in an abnormal state, the electronic device 110 can determine a second industrial inspection task from a set of industrial inspection tasks based on a preset execution relationship. As an example, multiple user-configured industrial inspection tasks may include a second industrial inspection task. When it is determined that the semiconductor sample is in an abnormal state, the electronic device 110 can determine the second industrial inspection task, which is the second most important task, from among the multiple industrial inspection tasks based on a preset execution relationship. For example, the configuration information may specify that the failure mode classification task is only triggered when the detection result of the initial screening task indicates that the semiconductor sample is abnormal and the confidence level is greater than a preset maximum confidence level. The second industrial inspection task can be a failure mode classification task. The failure mode classification task is used to determine the type of abnormality of the semiconductor sample.

[0042] Finally, the electronic device 110 can perform a second industrial inspection task from a set of industrial inspection tasks based on the data to be inspected, to obtain a second inspection result. The second inspection result indicates the type of anomaly of the semiconductor sample. The anomaly type can include contamination type and damage type. The contamination type indicates the presence of contaminants on the surface of the semiconductor sample. The damage type indicates that the flatness of the semiconductor sample surface does not meet the first constraint (i.e., the semiconductor sample shows scratches, etc.).

[0043] When performing the second industrial inspection task, the electronic device 110 can use an appropriate model (e.g., a defect classification model) to classify the anomaly type of the semiconductor sample based on scanning electron microscope images. For example, when the defect classification model detects foreign particles, chemical residues, organic contaminants, or metallic impurities adhering to the surface of the semiconductor sample, the anomaly type can be classified as "contamination type." As another example, when the defect classification model detects that the flatness of the semiconductor sample surface does not meet the first constraint, i.e., the semiconductor sample surface has scratches, dents, cracks, chipping, or other mechanical damage, the anomaly type can be classified as "damage type." After the second industrial inspection task is completed, the electronic device 110 can obtain a second inspection result and a corresponding confidence level. For example, "particles" and the corresponding confidence level (e.g., 0.9).

[0044] In some embodiments, electronic device 110 can determine a third industrial inspection task based on the anomaly type and a preset execution relationship. Electronic device 110 can execute the third industrial inspection task based on the data to be inspected to obtain a third inspection result. The third inspection result indicates attribute parameters associated with the anomaly state. The attribute parameters indicate the material and / or damage morphology of the contaminant.

[0045] As an example, after determining the anomaly type of a semiconductor sample, electronic device 110 can determine a third industrial inspection task in the third order from a set of industrial inspection tasks based on the anomaly type and a preset execution relationship. The third industrial inspection task could be, for example, an anomaly subdivision task used to determine attribute parameters associated with the anomaly state. For example, determining the material of the contaminant and the morphology of the damage.

[0046] Taking the third detection task as particulate matter segmentation as an example, since the particulate matter segmentation task requires energy spectrum data as input, the electronic device 110 can obtain the EDS energy spectrum corresponding to the semiconductor sample from the data storage unit based on the sample identifier. For example, the electronic device 110 can obtain the EDS energy spectrum based on the sample identifier and the data source mapping table. Furthermore, the electronic device 110 can call a pre-trained particulate matter segmentation model. This model is configured to determine the material type of the particles based on the EDS energy spectrum. The particulate matter segmentation model can process the input EDS energy spectrum to output the third detection result. The third detection result can indicate the material properties of the particles, such as metal particles, organic particles, or silicon-based particles, and simultaneously output the corresponding confidence score.

[0047] Taking the third task, scratch morphology analysis, as an example, the electronic device 110 can call the corresponding model to further analyze the scanning electron microscope image and output the third detection result. The third detection result indicates the damage morphology parameters, such as scratch depth, scratch width, and scratch direction.

[0048] Taking the third task as the analysis of pollutant components as an example, the electronic device 110 can call the corresponding model to analyze the scanning electron microscope image and other associated spectral data, and output the third detection result. This third detection result can indicate the chemical composition or type of pollutants, such as organic pollutants, metallic pollutants, etc.

[0049] In this way, embodiments of this disclosure can dynamically determine and execute more in-depth third-level detection tasks based on anomaly types and preset execution relationships, thereby obtaining more detailed attribute parameters. This can provide richer information for process improvement and quality control.

[0050] In some embodiments, electronic device 110 may determine a fourth industrial inspection task from a set of industrial inspection tasks based on preset execution relationships and attribute parameters. Further, electronic device 110 may execute the fourth industrial inspection task based on the data to be inspected to determine a fourth inspection result. The fourth inspection result indicates the production equipment that caused the abnormal state.

[0051] As an example, after obtaining the third detection result indicating the attribute parameter through the third industrial detection task, the electronic device 110 can further determine the fourth industrial detection task based on the preset execution relationship and the attribute parameter. The preset execution relationship here can refer to the execution order or data transmission order of a set of industrial detection tasks defined by the user in the configuration information. The configuration information may include: triggering the execution of a root cause analysis task related to metal particles when the third detection result indicates that the particle material is metal; triggering the execution of a root cause analysis task related to organic particles when the third detection result indicates that the particle material is organic; and triggering the execution of a root cause analysis task related to scratches when the third detection result indicates that the damage morphology is scratches. For example, after obtaining the third detection result of metal particles, the electronic device 110 can determine the next industrial detection task to be executed as a root cause analysis task related to "metal," i.e., the fourth industrial detection task, based on the preset execution relationship and the attribute parameter "metal."

[0052] Since root cause analysis requires process parameters as input, electronic device 110 can retrieve the process parameters for the production period of the sample from the data storage unit based on the sample's production timestamp. The process parameter library includes parameters such as equipment identification, temperature, pressure, and duration for each process the semiconductor sample undergoes during production.

[0053] Furthermore, electronic device 110 can invoke a root cause analysis model to perform a fourth industrial inspection task. The root cause analysis model is configured to perform correlation analysis based on attribute parameters (e.g., metal particles) and process parameters to identify production equipment suspected of causing defects. The root cause analysis model can be implemented using various algorithms, such as statistical analysis-based anomaly detection algorithms, rule engine-based logical reasoning, or machine learning-based classification models. The root cause analysis model processes the input attribute parameters and process parameters, outputting a fourth inspection result that indicates one or more production equipment suspected of causing the abnormal state.

[0054] For example, a root cause analysis model might output "suspected production equipment or production process" as the fourth detection result. In another scenario, the model can also output multiple suspected devices and their confidence levels, such as "production equipment A (confidence level 0.8), production equipment B (confidence level 0.6)," for reference during subsequent manual review.

[0055] In some embodiments, the electronic device 110 can determine the target detection result based at least on the first detection result and the second detection result. As an example, after obtaining the first and second detection results, the electronic device 110 can integrate these two results to obtain the target detection result. For example, the first detection result is "abnormal" with a confidence level of 0.95. The second detection result is "particle" with a confidence level of 0.8. The detection management platform combines these two results to form the target detection result, such as "the sample is in an abnormal state, and the abnormality type is particle."

[0056] As another example, the target detection results can also include the confidence level corresponding to each detection result. For example, the target detection results could be "Initial screening result: Abnormal (confidence level 0.95); Defect type: Particle (confidence level 0.8)". Such confidence level information helps engineers assess the reliability of the detection results and also provides a basis for prioritizing subsequent manual review.

[0057] As another example, if the first detection result is abnormal but the confidence level of the second detection result is low (e.g., between a preset low threshold and a high threshold), the electronic device 110 can mark the sample as "awaiting manual review" in the target detection result. For example, the target detection result can be expressed as: "The sample is in an abnormal state, the abnormality type is particles, and due to the low confidence level, it awaits manual review."

[0058] In some scenarios, when a user configures three or four industrial inspection tasks, the electronic device 110 can also determine the target inspection result by comprehensively considering the inspection results of all configured industrial inspection tasks. For example, the user configures four levels of inspection tasks, such as initial screening, defect classification, particle subdivision, and root cause analysis. After the electronic device 110 executes these tasks sequentially, it obtains four inspection results. The electronic device 110 can determine the target inspection result based on all the results. Such a target inspection result could be, for example, "This wafer sample has an anomaly; the defect type is particles, the particle material is metal, and it is suspected to be caused by injection device A."

[0059] In this way, the embodiments of this disclosure can flexibly determine the target detection results suitable for actual needs based on the actual task level and results.

[0060] In some embodiments, the configuration information further indicates a second execution mode, which instructs the parallel execution of a set of industrial inspection tasks. As an example, the user can also use configuration information to instruct the parallel processing of a set of industrial inspection tasks. During parallel processing, the electronic device 110 can acquire the data to be inspected according to the data required by different industrial inspection tasks. After the data to be inspected is acquired, the electronic device 110 can invoke different models to process the data to obtain the corresponding inspection results and the confidence levels corresponding to the inspection results.

[0061] In box 240, electronic device 110 receives user input feedback in response to a set of confidence levels meeting preset conditions. As an example, the preset conditions here could indicate that a set of confidence levels is between a minimum confidence level and a maximum confidence level. When a set of confidence levels is between the minimum and maximum confidence levels, electronic device 110 can receive user input feedback. Such feedback could, for example, indicate whether a set of detection results is correct. For example, electronic device 110 can provide a feedback collection interface for the user, which includes the detection results of the semiconductor sample, the data to be detected, and the confidence levels. Electronic device 110 can receive user input via the feedback collection interface. Such input indicates feedback information, which can indicate whether the detection results are correct.

[0062] In some embodiments, the electronic device 110 may display a prompt message in response to a set of confidence levels meeting preset conditions. The prompt message indicates a request for feedback information. As an example, after determining that a set of confidence levels meets preset conditions (e.g., confidence levels between 0.1 and 0.9), the electronic device 110 may display a feedback collection interface. The feedback collection interface includes a prompt message. The prompt message may, for example, be "Please review the test results."

[0063] In box 250, electronic device 110 provides feedback information to a set of models corresponding to a set of working detection tasks, as training samples for the set of models. As an example, electronic device 110 can structure the collected feedback information to fit the training format of the models corresponding to each detection task. The feedback information may include sample identifiers, task identifiers, original detection results, correction results, and timestamps. Based on the sample identifier, electronic device 110 can retrieve the input data for that sample at each detection level from the data storage unit, such as scanning electron microscope images, energy dispersive spectroscopy (EDS) spectra, and process parameters.

[0064] In some scenarios, users configure four-layer cascaded detection tasks, including initial screening, defect classification, particle size analysis, and root cause analysis. After receiving feedback from the user regarding the semiconductor sample, the electronic device 110 can analyze the feedback to determine the user's confirmation or correction of the detection results at each level.

[0065] For example, in a second-layer defect classification task, the user confirms that the "particle" result output by the model is correct. The detection management platform combines the second-layer input data (scanning electron microscope image) of this sample with the user-confirmed label "particle" to form a training sample, and adds it to the training dataset of the defect classification model.

[0066] For the third-layer particle subdivision task, the electronic device 110 can obtain feedback information from the user who corrects the "metal particles" output by the model to "organic particles". Based on the input data of the third detection task and the user's feedback information, the electronic device 110 constructs training samples for subsequent optimization of the particle subdivision model.

[0067] For the fourth layer of root cause analysis, the user confirmed that the model output "suspected production equipment A" was correct. Electronic device 110 can construct training samples based on the input data (particle material information and process parameter table) and feedback information of the fourth industrial inspection task for retraining of the root cause analysis model.

[0068] Electronic device 110 can process feedback information periodically. For example, it can perform a model update process daily, weekly, or monthly. When the number of feedback samples for a certain task reaches a preset threshold (e.g., 100), electronic device 110 can trigger the training of the model corresponding to that task. After training, electronic device 110 can evaluate the performance of the updated model. If the evaluation result is better than the current model, it will automatically deploy the new model and update the model version in the configuration information so that subsequent detection tasks can use the optimized model.

[0069] In some scenarios, the electronic device 110 can also perform statistical analysis on the feedback information to identify the sample types or patterns in which the model frequently makes mistakes. For example, if it is found that the defect classification model frequently misclassifies particles of a certain morphology as scratches, the electronic device 110 can collect more feedback information on such samples and increase the weight of these samples during training to improve the model's recognition accuracy on such samples.

[0070] In other scenarios, the collection and use of feedback information can be carried out independently for each task. Each industrial inspection task has its own independent set of feedback samples and training and update cycle. The model updates for different industrial inspection tasks can not affect models at other levels. This hierarchical closed-loop mechanism allows the electronic device 110 to specifically optimize specific poorly performing models without retraining the entire system.

[0071] In some embodiments, the electronic device 110 can also perform structured storage and associated mapping of input data, feature extraction results, detection results, confidence scores, risk classification results, manual review access points, and feedback information for an industrial inspection task.

[0072] In this way, embodiments of this disclosure can determine a set of industrial inspection tasks to be executed based on user configuration information, thereby making the entire inspection process more flexible and efficient. Furthermore, embodiments of this disclosure can obtain user feedback information after the confidence level meets preset conditions, achieving precise human-machine collaboration. Manual review is triggered only when there is uncertainty, ensuring inspection efficiency while continuously optimizing the model through feedback information, thereby improving the accuracy of data processing.

[0073] Example Apparatus and Device A corresponding apparatus for implementing the above methods or processes is also provided. Figure 3 A schematic structural block diagram of an example device 300 for data processing according to some scenarios is shown. Device 300 can be implemented as or included in electronic device 110. The various modules / components in device 300 can be implemented by hardware, software, firmware, or any combination thereof.

[0074] like Figure 3 As shown, the device 300 includes: a first acquisition module 310 configured to acquire configuration information configured by a user, the configuration information indicating a set of industrial inspection tasks to be executed, the set of industrial inspection tasks being used to inspect semiconductor samples; a second acquisition module 320 configured to acquire data to be inspected from the semiconductor samples; a first execution module 330 configured to execute a set of industrial inspection tasks based on the data to be inspected to obtain a set of confidence levels, the set of confidence levels being associated with a set of inspection results of the set of industrial inspection tasks; a third acquisition module 340 configured to acquire feedback information input by a user in response to a set of confidence levels satisfying a preset condition; and a providing module 350 configured to provide the feedback information to a set of models corresponding to the set of industrial inspection tasks as training samples for the set of models.

[0075] In some cases, device 300 also includes an information presentation module configured to present a prompt message in response to a set of confidence levels meeting preset conditions, the prompt message indicating a request for feedback information.

[0076] In some cases, the configuration information also indicates a first execution mode, which instructs a set of industrial inspection tasks to be executed according to a preset execution relationship. The first execution module 330 is further configured to: execute a first industrial inspection task in a set of industrial inspection tasks based on the data to be inspected to obtain a first inspection result; in response to the first inspection result indicating that the semiconductor sample is in an abnormal state, determine a second industrial inspection task from a set of industrial inspection tasks based on the preset execution relationship; and execute the second industrial inspection task in a set of industrial inspection tasks based on the data to be inspected to obtain a second inspection result, the second inspection result indicating the abnormality type of the semiconductor sample.

[0077] In some cases, the anomaly type includes at least one of the following: contamination type, which indicates the presence of contaminants on the surface of the semiconductor sample; damage type, which indicates that the flatness of the semiconductor sample surface does not meet the first constraint.

[0078] In some cases, the device 300 also includes a second execution module configured to determine a third industrial detection task based on the anomaly type and a preset execution relationship; and to execute the third industrial detection task based on the data to be detected to obtain a third detection result, wherein the third detection result indicates attribute parameters associated with the anomaly state, and the attribute parameters indicate the material and / or damage morphology of the contaminant.

[0079] In some cases, the device 300 also includes a third execution module configured to determine a fourth industrial inspection task from a set of industrial inspection tasks based on preset execution relationships and attribute parameters; and to execute the fourth industrial inspection task based on the data to be inspected to determine a fourth inspection result, the fourth inspection result indicating the production equipment that caused the abnormal state.

[0080] In some cases, the device 300 also includes a determining module configured to determine a target detection result based at least on a first detection result and a second detection result.

[0081] In some cases, the configuration information also indicates a second execution mode, which instructs a set of industrial inspection tasks to be executed in parallel.

[0082] In some cases, the data to be detected includes at least one of the following: image data associated with the semiconductor sample; spectral data associated with the semiconductor sample; production parameter data associated with the semiconductor sample; and information about the production equipment associated with the semiconductor.

[0083] In some cases, the second acquisition module 320 is also configured to: acquire identification information of semiconductor samples; and, based on the identification information, acquire data to be tested corresponding to a set of industrial testing tasks.

[0084] In some cases, the identification information includes at least one of the following: the sample identifier of the semiconductor sample; the time information of the semiconductor sample.

[0085] The modules included in device 300 can be implemented in various ways, including software, hardware, firmware, or any combination thereof. In some cases, one or more modules can be implemented using software and / or firmware, such as machine-executable instructions stored on a storage medium. In addition to or as an alternative to machine-executable instructions, some or all of the units in device 300 can be implemented at least partially by one or more hardware logic components. By way of example and not limitation, exemplary types of hardware logic components that can be used include field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard parts (ASSPs), systems on a chip (SOCs), complex programmable logic devices (CPLDs), and so on.

[0086] Figure 4 A block diagram of an electronic device 400 in which one or more examples may be implemented is shown. It should be understood that... Figure 4 The electronic device 400 shown is merely exemplary and should not be construed as limiting the functionality and scope of the examples described herein. Figure 4 The illustrated electronic device 400 can be used to implement the electronic device 110 discussed above.

[0087] like Figure 4 As shown, electronic device 400 is in the form of a general-purpose electronic device. Components of electronic device 400 may include, but are not limited to, one or more processing units or processors 410, memory 420, storage device 430, one or more communication units 440, one or more input devices 450, and one or more output devices 460. Processor 410 may be a physical or virtual processor and is capable of performing various processes according to programs stored in memory 420. In a multiprocessor system, multiple processors execute computer-executable instructions in parallel to improve the parallel processing capability of electronic device 400.

[0088] Electronic device 400 typically includes multiple computer storage media. Such media can be any accessible media that is accessible to electronic device 400, including but not limited to volatile and non-volatile media, removable and non-removable media. Memory 420 can be volatile memory (e.g., registers, cache, random access memory (RAM)), non-volatile memory (e.g., read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory), or some combination thereof). Storage device 430 can be removable or non-removable media and can include machine-readable media, such as flash drives, disks, or any other media that can be used to store information and / or data and can be accessed within electronic device 400.

[0089] Electronic device 400 may further include additional removable / non-removable, volatile / non-volatile storage media. Although not explicitly stated... Figure 4 As shown, disk drives for reading from or writing to removable, non-volatile disks (e.g., "floppy disks") and optical disk drives for reading from or writing to removable, non-volatile optical disks can be provided. In these cases, each drive can be connected to a bus (not shown) via one or more data media interfaces. Memory 420 may include computer program product 425 having one or more program modules configured to perform various methods or actions of various examples.

[0090] The communication unit 440 enables communication with other electronic devices via a communication medium. Additionally, the functionality of the components of the electronic device 400 can be implemented using a single computing cluster or multiple computing machines capable of communicating via communication connections. Therefore, the electronic device 400 can operate in a networked environment using logical connections to one or more other servers, networked personal computers, or another network node.

[0091] Input device 450 can be one or more input devices, such as a mouse, keyboard, trackball, etc. Output device 460 can be one or more output devices, such as a monitor, speaker, printer, etc. Electronic device 400 can also communicate with one or more external devices (not shown) via communication unit 440 as needed. These external devices include storage devices, display devices, etc., and can communicate with one or more devices that enable user interaction with electronic device 400, or with any device that enables electronic device 400 to communicate with one or more other electronic devices (e.g., network card, modem, etc.). Such communication can be performed via an input / output (I / O) interface (not shown).

[0092] A computer-readable storage medium is provided that stores computer-executable instructions thereon, wherein the computer-executable instructions are executed by a processor to implement the methods described above. A computer program product is also provided, which is tangibly stored on a non-transitory computer-readable medium and includes computer-executable instructions, which are executed by a processor to implement the methods described above.

[0093] The flowcharts and / or block diagrams of the methods, apparatus, devices, and computer program products referred to herein describe various aspects. It should be understood that each block of the flowcharts and / or block diagrams, as well as combinations of blocks in the flowcharts and / or block diagrams, can be implemented by computer-readable program instructions.

[0094] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0095] Computer-readable program instructions can be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions that execute on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0096] The flowcharts and block diagrams in the accompanying figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products under various scenarios. In this respect, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction, which contains one or more executable instructions for implementing the specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the figures. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0097] Various examples have been described above. The foregoing descriptions are exemplary and not exhaustive, nor are they limited to the disclosed implementations. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described implementations. The terminology used herein is chosen to best explain the principles, practical applications, or improvements to technology in the market, or to enable others skilled in the art to understand the various implementations disclosed herein.

Claims

1. A data processing method for semiconductor detection, comprising: Obtain configuration information configured by the user, the configuration information indicating a set of industrial inspection tasks to be executed, the set of industrial inspection tasks being used to inspect semiconductor samples; Obtain the test data of the semiconductor sample; Based on the data to be tested, the set of industrial testing tasks are executed to obtain a set of confidence levels, which are associated with a set of test results of the set of industrial testing tasks. In response to the set of confidence levels satisfying preset conditions, the user input feedback information is obtained; as well as The feedback information is provided to a set of models corresponding to the set of industrial inspection tasks, as training samples for the set of models.

2. The method according to claim 1, further comprising: In response to the set of confidence levels meeting preset conditions, a prompt message is displayed, indicating a request to obtain the feedback information.

3. The method according to claim 1, wherein the configuration information further indicates a first execution mode, the first execution mode instructing the set of industrial inspection tasks to be executed according to a preset execution relationship, and executing the set of industrial inspection tasks based on the data to be inspected includes: Based on the data to be tested, the first industrial testing task in the set of industrial testing tasks is executed to obtain the first testing result; In response to the first detection result indicating that the semiconductor sample is in an abnormal state, a second industrial detection task is determined from the set of industrial detection tasks based on the preset execution relationship; as well as Based on the data to be tested, the second industrial testing task in the set of industrial testing tasks is executed to obtain a second testing result, which indicates the type of anomaly of the semiconductor sample.

4. The method of claim 3, wherein the exception type includes at least one of the following: Contamination type, which indicates the presence of contaminants on the surface of the semiconductor sample; Damage type, the damage type indicating that the flatness of the semiconductor sample surface does not meet the first constraint.

5. The method according to claim 3, further comprising: Based on the anomaly type and the preset execution relationship, the third industrial inspection task is determined; Based on the data to be detected, a third industrial detection task is performed to obtain a third detection result, which indicates attribute parameters associated with the abnormal state, and the attribute parameters indicate the material and / or damage morphology of the pollutant.

6. The method of claim 5, further comprising: Based on the preset execution relationship and the attribute parameters, a fourth industrial inspection task is determined from the set of industrial inspection tasks; as well as Based on the data to be tested, a fourth industrial testing task is performed to determine a fourth testing result, which indicates the production equipment that caused the abnormal state.

7. The method according to claim 3, further comprising: The target detection result is determined based at least on the first detection result and the second detection result.

8. The method of claim 1, wherein the configuration information further indicates a second execution mode, the second execution mode indicating that the set of industrial inspection tasks are executed in parallel.

9. The method according to claim 1, wherein the data to be detected includes at least one of the following: Image data associated with the semiconductor sample; Spectral data associated with the semiconductor sample; Production parameter data associated with the semiconductor sample; Information about the manufacturing equipment associated with the semiconductor.

10. The method of claim 1, wherein obtaining the test data of the semiconductor sample comprises: Obtain the identification information of the semiconductor sample; as well as Based on the identification information, the data to be tested corresponding to the set of industrial testing tasks is obtained.

11. The method of claim 10, wherein the identification information includes at least one of the following: The sample identifier of the semiconductor sample; The time information of the semiconductor sample.

12. An apparatus for data processing in semiconductor detection, comprising: The first acquisition module is configured to acquire configuration information configured by the user, the configuration information indicating a set of industrial inspection tasks to be executed, the set of industrial inspection tasks being used to inspect semiconductor samples; The second acquisition module is configured to acquire the detection data of the semiconductor sample; The first execution module is configured to execute the set of industrial testing tasks based on the data to be tested, so as to obtain a set of confidence levels, the set of confidence levels being associated with a set of testing results of the set of industrial testing tasks; The third acquisition module is configured to acquire the user input feedback information in response to the set of confidence levels meeting preset conditions. as well as A module is configured to provide the feedback information to a set of models corresponding to the set of industrial inspection tasks, as training samples for the set of models.

13. An electronic device, comprising: At least one processor; as well as At least one memory coupled to the at least one processor and storing instructions for execution by the at least one processor, the instructions causing the electronic device to perform the method according to any one of claims 1 to 11 when executed by the at least one processor.

14. A computer-readable storage medium having stored thereon computer-executable instructions that can be executed by a processor to implement the method according to any one of claims 1 to 11.

15. A computer program product tangibly stored in a computer storage medium and comprising computer-executable instructions that, when executed by a device, cause the device to perform the method according to any one of claims 1 to 11.

Images