A method, device, storage medium and electronic device for identifying fast radio bursts

Abstract

The specification discloses a method and device for identifying a fast radio burst. After obtaining observation data, first data representing the relationship between the frequency of each observed signal and the time at which each signal is observed can be determined based on the observation data, and second data representing the relationship between the dispersion value of each observed signal and the time at which each signal is observed can be determined based on the first data. Then, the second data can be input into a pre-trained first identification model to obtain an output identification result, and the signal of the fast radio burst can be identified from the observation data based on the identification result. As can be seen from the above method, after obtaining the observation data, the second data for inputting into the first identification model can be obtained based on the observation data, so that the signal of the fast radio burst can be quickly and accurately identified from the observation data through the first identification model, and the search efficiency of the fast radio burst is improved.

Description

TECHNICAL FIELD

[0001] The present specification relates to the technical field of computer technology, and in particular, to a method and device for identifying a fast radio burst, a storage medium, and an electronic device. BACKGROUND

[0002] A fast radio burst (FRB) is a high-energy astrophysical phenomenon occurring in the radio frequency band and lasting for a very short time. It is of great significance to the study of the distribution and evolution of matter in the universe. However, so far, there is no universally accepted theory to explain the physical mechanism behind its ultra-high energy release rate.

[0003] Since the signal will disperse when propagating in the universe, that is, the propagation speed of different frequencies of light is different, so when the signal of the fast radio burst reaches the earth, the signal of the high frequency band arrives before the signal of the low frequency band. This dispersion phenomenon causes the signal of the fast radio burst, which is originally in the order of milliseconds, to be lengthened to several seconds to tens of seconds, and the concentrated energy is dispersed, so that the signal of the fast radio burst is submerged in the noise and cannot be detected.

[0004] Therefore, how to accurately search for the signal of the fast radio burst from the captured many signals is a technical problem to be solved. SUMMARY

[0005] The present specification provides a method and device for identifying a fast radio burst, a storage medium, and an electronic device to partially solve the above problems existing in the prior art.

[0006] The present specification adopts the following technical solutions:

[0007] The present specification provides a method for identifying a fast radio burst, comprising:

[0008] Obtaining observation data;

[0009] According to the observation data, determining first data, the first data being used to represent the relationship between the frequency of each observed signal and the time when each signal is observed;

[0010] According to the first data, determining second data, the second data being used to represent the relationship between the dispersion value of each observed signal and the time when each signal is observed;

[0011] Inputting the second data into a pre-trained first identification model to obtain an identification result output by the first identification model;

[0012] According to the identification result, identifying the signal of the fast radio burst from the observation data.

[0013] Optionally, the second data is input into a pre-trained first identification model to obtain an identification result output by the first identification model, specifically including:

[0014] The second data is input into the first identification model to determine, by the first identification model, the arrival time of each candidate signal in the observation data as the identification result output by the first identification model.

[0015] According to the identification result, the signal of the fast radio burst is identified from the observation data, specifically including:

[0016] For each candidate signal, according to the arrival time of the candidate signal and the observation data, third data corresponding to the candidate signal is determined, the third data being used to represent the relationship between the frequency of the candidate signal and the time when the candidate signal is observed.

[0017] The third data corresponding to the candidate signal is input into a pre-trained second identification model to obtain an identification result for the candidate signal.

[0018] According to the identification result of each candidate signal, the signal of the fast radio burst is identified from the observation data.

[0019] Optionally, before the third data corresponding to the candidate signal is input into the pre-trained second identification model to obtain the identification result for the candidate signal, the method further includes:

[0020] The third data corresponding to the candidate signal is subjected to a delay time correction process to obtain processed data corresponding to the candidate signal.

[0021] The third data corresponding to the candidate signal is input into a pre-trained second identification model to obtain an identification result for the candidate signal, specifically including:

[0022] The processed data corresponding to the candidate signal is input into the second identification model to obtain the identification result for the candidate signal.

[0023] Optionally, according to the first data, second data is determined, specifically including:

[0024] For each preset dispersion range, the first data is subjected to a delay time correction according to the dispersion range to obtain corrected data corresponding to the dispersion range.

[0025] The corrected data corresponding to each dispersion range is superimposed to obtain the second data.

[0026] Optionally, before inputting the second data into the pre-trained first identification model to obtain the identification result output by the first identification model, the method further comprises:

[0027] performing data segmentation on the second data in the time dimension and the dispersion dimension to obtain each data block;

[0028] inputting the second data into the pre-trained first identification model to obtain the identification result output by the first identification model, specifically comprising:

[0029] inputting each data block into the first identification model to obtain the identification result output by the first identification model.

[0030] Optionally, pre-training the first identification model specifically comprises:

[0031] obtaining first sample data according to the arrival time of the historical fast radio burst;

[0032] determining historical first data according to the first sample data;

[0033] determining historical second data according to the historical first data;

[0034] inputting the historical second data into the first identification model to be trained to obtain the identification result corresponding to the first sample data;

[0035] training the first identification model with the optimization target of minimizing the deviation between the identification result corresponding to the first sample data and the actual arrival time of the historical fast radio burst.

[0036] Optionally, obtaining first sample data according to the arrival time of the historical fast radio burst specifically comprises:

[0037] determining each data segment from historical observation data according to the arrival time of the historical fast radio burst;

[0038] performing data sampling on each data segment according to a preset data sampling mode to obtain sampling data;

[0039] performing delay time correction on the sampling data according to a preset dispersion range to obtain corrected sampling data;

[0040] superimposing the corrected data on the frequency of the signal to obtain first sample data in each frequency range.

[0041] Optionally, determining each data segment from historical observation data according to the arrival time of the historical fast radio burst specifically comprises:

[0042] According to the arrival time of the historical fast radio burst, the historical observation data is fragmented according to a preset step, and each data segment is obtained.

[0043] Optionally, the second identification model is pre-trained, and specifically includes:

[0044] Obtaining second sample data;

[0045] According to the dispersion value and the arrival time of the historical candidate signal in the second sample data, third data of each historical candidate signal is determined from the historical observation data.

[0046] The third data corresponding to each historical candidate signal is input into the second identification model to obtain an identification result corresponding to each historical candidate signal.

[0047] The second identification model is trained with the optimization target of minimizing the deviation between the identification result corresponding to each historical candidate signal and the label corresponding to each historical candidate signal, wherein for each historical candidate signal, the label corresponding to the historical candidate signal is used to indicate whether the historical candidate signal is actually a historical fast radio burst.

[0048] The present specification provides a device for identifying a fast radio burst, comprising:

[0049] An acquisition module is configured to acquire observation data.

[0050] A first determination module is configured to determine first data according to the observation data, wherein the first data is used to represent the relationship between the frequency of each observed signal and the time when each signal is observed.

[0051] A second determination module is configured to determine second data according to the first data, wherein the second data is used to represent the relationship between the dispersion value of each observed signal and the time when each signal is observed.

[0052] An input module is configured to input the second data into a pre-trained first identification model to obtain an identification result output by the first identification model.

[0053] An identification module is configured to identify the signal of a fast radio burst from the observation data according to the identification result.

[0054] Optionally, the input module is specifically configured to input the second data into the first identification model to determine the arrival time of each candidate signal in the observation data by the first identification model as the identification result output by the first identification model.

[0055] The identification module is specifically configured to: for each candidate signal, determine third data corresponding to the candidate signal according to the arrival time of the candidate signal and the observation data, the third data being used to represent the relationship between the frequency of the candidate signal and each time instant at which the candidate signal is observed; input the third data corresponding to the candidate signal into a second identification model that is pre-trained, to obtain an identification result for the candidate signal; and identify the signal of the fast radio burst from the observation data according to the identification result for each candidate signal.

[0056] Optionally, the identification module is further configured to, before inputting the third data corresponding to the candidate signal into the second identification model that is pre-trained to obtain the identification result for the candidate signal, perform delay time correction processing on the third data corresponding to the candidate signal, to obtain processed data corresponding to the candidate signal; and input the processed data corresponding to the candidate signal into the second identification model, to obtain the identification result for the candidate signal.

[0057] Optionally, the second determination module is specifically configured to: for each preset dispersion range, perform delay time correction on the first data according to the dispersion range, to obtain modified data corresponding to the dispersion range; and superimpose the modified data corresponding to each dispersion range, to obtain the second data.

[0058] Optionally, before inputting the second data into the first identification model that is pre-trained to obtain the identification result output by the first identification model, the input module is further configured to perform data segmentation on the second data in the time dimension and the dispersion dimension, to obtain data blocks.

[0059] The input module is specifically configured to input the data blocks into the first identification model, to obtain the identification result output by the first identification model.

[0060] Optionally, the apparatus further includes:

[0061] The first training module is configured to: obtain first sample data according to the arrival time of a historical fast radio burst; determine historical first data according to the first sample data; determine historical second data according to the historical first data; input the historical second data into a first identification model to be trained, to obtain an identification result corresponding to the first sample data; and train the first identification model, with the optimization objective being to minimize the deviation between the identification result corresponding to the first sample data and the actual arrival time of the historical fast radio burst.

[0062] Optionally, the first training module is specifically configured to: determine, according to the arrival time of the historical fast radio burst, each data segment from the historical observation data; perform data sampling on each data segment according to a preset data sampling mode to obtain sampled data; perform delay time correction on the sampled data according to a preset dispersion range to obtain corrected sampled data; and superimpose the corrected data on the frequency of the signal to obtain first sample data of each frequency range.

[0063] Optionally, the first training module is specifically configured to: according to the arrival time of the historical fast radio burst, perform slicing on the historical observation data according to a preset step size to obtain each data segment.

[0064] Optionally, the apparatus further comprises:

[0065] The second training module is configured to: obtain second sample data; determine, according to the dispersion value and the arrival time of the historical candidate signal in the second sample data, third data of each historical candidate signal from the historical observation data; input the third data corresponding to each historical candidate signal into the second identification model to obtain an identification result corresponding to each historical candidate signal; and train the second identification model with the optimization target of minimizing the deviation between the identification result corresponding to each historical candidate signal and a label corresponding to the historical candidate signal, wherein the label corresponding to each historical candidate signal is used to represent whether the historical candidate signal is actually a historical fast radio burst.

[0066] The present specification provides a computer-readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the method for identifying a fast radio burst.

[0067] The present specification provides an electronic device, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the method for identifying a fast radio burst when executing the program.

[0068] The above at least one technical solution adopted by the present specification can achieve the following beneficial effects:

[0069] In the method for identifying a fast radio burst provided by the present specification, after the observation data is obtained, the first data used to represent the relationship between the frequency of each observed signal and the time at which each signal is observed can be determined according to the observation data, and the second data used to represent the relationship between the dispersion value of each observed signal and the time at which each signal is observed can be determined according to the first data. Then, the second data can be input into the pre-trained first identification model to obtain an output identification result, and the signal of the fast radio burst can be identified from the observation data according to the identification result.

[0070] As can be seen from the above method, after the observation data is obtained, the second data used for input into the first identification model can be obtained based on the observation data, so that the signal of the fast radio burst is quickly and accurately identified from the observation data through the first identification model, and the search efficiency of the fast radio burst is improved. BRIEF DESCRIPTION OF DRAWINGS

[0071] The accompanying drawings, which are included to provide a further understanding of the present specification, constitute a part of the present specification, and the illustrative embodiments of the present specification and their description serve to explain the present specification, and do not constitute an improper limitation on the present specification. In the drawings:

[0072] Figure 1 A flowchart of a method for identifying a fast radio burst in the present specification;

[0073] Figure 2 A graph of the dispersion value of a radio signal versus time provided in the present specification;

[0074] Figures 3A-3C A schematic diagram of a graph of each frequency versus time provided in the present specification;

[0075] Figure 4A And Figure 4B A schematic diagram of data before and after delay time processing provided in the present specification;

[0076] Figure 5 A schematic diagram of an apparatus for identifying a fast radio burst provided in the present specification;

[0077] Figure 6 A schematic diagram of an electronic device corresponding to Figure 1 provided in the present specification. DETAILED DESCRIPTION

[0078] In order to make the purpose, technical solutions and advantages of the present specification clearer, the technical solutions of the present specification will be described clearly and completely in combination with the specific embodiments of the present specification and the corresponding drawings. Obviously, the described embodiments are only some of the embodiments of the present specification, not all the embodiments. Based on the embodiments in the present specification, all other embodiments obtained by those skilled in the art without making creative efforts fall within the scope of the present application.

[0079] The technical solutions provided by the embodiments of the present specification will be described in detail below in combination with the drawings.

[0080] Figure 1 A flowchart of a method for identifying a fast radio burst in the present specification, specifically comprising the following steps:

[0081] S101: Obtain observation data.

[0082] In the present specification, the observation data for identifying a fast radio burst can be obtained, wherein the observation data can be obtained by observing radio signals in the universe through an astronomical radio telescope. It should be noted that, when the astronomical radio telescope observes radio signals in the universe, it can also collect radio signals generated by various human activities such as television, mobile phone, satellite, etc., which are called radio frequency interference (RFI). Since it is necessary to identify the signal of the fast radio burst from the observation data, the radio signals of the radio frequency interference can be regarded as noise in the observation data.

[0083] Therefore, the method for identifying a fast radio burst provided in the present specification actually needs to quickly and accurately identify the radio signal of the fast radio burst from the observation data containing a large amount of noise.

[0084] Further, in the present specification, the data format of the observation data can be a standard data format FITS used in the field of space astronomy, that is, a data format for transmitting, analyzing and archiving scientific data files recognized by astronomical and related organizations and organizations such as the International Astronomical Union (IAU), the National Aeronautics and Space Administration (NASA).

[0085] On this basis, the execution subject for executing the method for identifying a fast radio burst provided in the present specification can be an electronic terminal device such as a desktop computer, a notebook computer, etc., or a server. In order to facilitate description, the method for identifying a fast radio burst provided in the present specification will be described below by taking the terminal device as the execution subject. Correspondingly, the terminal device can obtain the observation data in the above data format to execute the subsequent steps.

[0086] S102: According to the observation data, determine first data, wherein the first data is used to represent the relationship between the frequency of each observed signal and the time when each signal is observed.

[0087] After the terminal device obtains the observation data, it can determine the first data used to represent the relationship between the frequency of each observed signal and the time when each signal is observed. That is, the first data is used to represent the relationship between the frequency and time of each observed signal.

[0088] For a radio signal, the first data can represent each time when the radio signal is observed, and the frequency of the radio signal corresponding to each time when the radio signal is observed.

[0089] S103: determining second data according to the first data, the second data being used to represent the relationship between the dispersion value of each observed signal and the time at which each observed signal is observed.

[0090] After the first data is determined, the terminal device can perform data conversion to obtain the second data used to represent the relationship between the dispersion value of each observed signal and the time at which each observed signal is observed.

[0091] The reason why the first data needs to be converted into the second data is that, by studying the relationship graph between the dispersion value of each signal and time, it is found that the signal of a fast radio burst will present a special image feature in the relationship graph between the dispersion value and time, such as Figure 2 .

[0092] Figure 2 The relationship graph between the dispersion value of a radio signal and time provided in the specification.

[0093] Figure 2 The relationship graph between the dispersion value and time shown, Figure 2 is actually obtained by superimposing a plurality of frequency-time relationship graphs according to different dispersion value steps. As can be seen from Figure 2 , the arrival time of a fast radio burst is actually the time corresponding to the intersection of a plurality of lines in Figure 2 . That is, the arrival time of a fast radio burst can be determined by finding the intersection of a plurality of lines in Figure 2 to form an "X" shape.

[0094] It should be noted that, as mentioned above, Figure 2 is actually obtained by superimposing a plurality of frequency-time relationship graphs, so in the specification, the terminal device can perform delay time correction on the first data according to each preset dispersion range to obtain the corrected data corresponding to the dispersion range, and then the terminal device can superimpose the corrected data corresponding to each dispersion range to obtain the second data, as shown in Figures 3A-3C .

[0095] Figures 3A-3C The schematic diagram of the relationship graph between each frequency and time provided in the specification.

[0096] Figure 3A The relationship graph obtained by performing delay time correction on the relationship between the frequency of a signal and the time at which the signal is observed according to the dispersion value of 100, Figure 3B The relationship graph obtained by performing delay time correction on the relationship between the frequency of a signal and the time at which the signal is observed according to the dispersion value of 200, Figure 3CA graph showing the relationship between the frequency of the signal and the time of the observed signal after the delay time correction according to a dispersion value of 300.

[0097] The reason for determining the delay time correction of the first data according to multiple dispersion ranges and superimposing the corrected data is that in actual application, it is unknown in which dispersion range the fast radio burst can be observed clearly, so the delay time correction needs to be performed according to different dispersion ranges. Regardless of which dispersion range is used for the delay time correction, the data of the fast radio burst are contained in the corrected data. Since the signal strength of the fast radio burst is the highest at the optimal dispersion value, and gradually weakens along both sides of the optimal dispersion value, superimposing the data after the delay time correction will continuously enhance the data of the fast radio burst at the time of arrival of the fast radio burst, so that the "X" shape shown in Figure 2 Figure 2 The intersection of the multiple lines in the figure is the result of the continuous superposition of the data of the fast radio burst at the time of arrival of the fast radio burst.

[0098] S104: inputting the second data into the first recognition model trained in advance to obtain a recognition result output by the first recognition model.

[0099] After obtaining the second data, the terminal device can input the second data into the first recognition model trained in advance. The first recognition model can recognize the image features generated when the fast radio burst bursts in the second data, and further determine the arrival time of the fast radio burst.

[0100] It should be noted that before inputting the second data into the first recognition model, the terminal device can first perform data division on the second data in the time dimension and the dispersion to obtain data blocks. In this way, inputting the second data into the first recognition model is actually inputting the data blocks into the first recognition model to obtain the recognition result.

[0101] Specifically, the data division manner can be data division according to a specified time length and dispersion value size. For example, assuming that the data in the time period of 0.15s to 0.2s is divided into a data block, the data in the time period of 0.16s to 0.21s is divided into another data block, and the data in the time period of 0.17s to 0.22s is divided into a data block, and so on, to finally obtain multiple data blocks.

[0102] ​Of course, due to the existence of various noise interference in actual application, if only the first identification model is used to identify the fast radio burst, misidentification may occur. Therefore, in the present specification, another identification model, i.e., the second identification model, can be introduced, and the first identification model and the second identification model are used to accurately identify the signal of the fast radio burst from the obtained observation data.

[0103] Specifically, after the second data is input into the first identification model, the arrival time of each candidate signal in the observation data can be determined by the first identification model as the identification result output by the first identification model. The candidate signal mentioned here can be understood as a signal that may be a fast radio burst identified by the first identification model. These candidate signals include actual fast radio burst signals and signals that are similar to fast radio bursts in signal characteristics but are not actually fast radio bursts.

[0104] Therefore, the subsequent use of the second identification model is actually to identify the actual fast radio burst from these candidate signals.

[0105] Specifically, for each candidate signal, the terminal device can determine the third data corresponding to the candidate signal according to the arrival time of the candidate signal and the above observation data, and the third data is used to represent the relationship between the frequency of the candidate signal and the time when the candidate signal is observed. That is, the corresponding relationship between the frequency of the candidate signal and the observation time is determined.

[0106] Then, the terminal device can input the third data corresponding to the candidate signal into the pre-trained second identification model to obtain the identification result for the candidate signal.

[0107] After the terminal device determines the identification result of each candidate signal, the terminal device can identify the signal of the fast radio burst from the observation data according to the identification results.

[0108] It should be noted that if only the first identification model is used, the arrival time of the identified fast radio burst can be output by the first identification model, and the terminal device can search for the signal arriving at the arrival time in the observation data, i.e., the signal of the fast radio burst.

[0109] If the first identification model and the second identification model are used simultaneously, the first identification model can output the arrival time of each candidate signal that can be a fast radio burst, and the second identification model can output which of the candidate signals is the signal of the fast radio burst and which is not. Accordingly, based on the result output by the second identification model, the terminal device can determine the arrival time of the fast radio burst, and then search for the signal arriving at the arrival time from the observation data, that is, the signal of the fast radio burst.

[0110] It should be noted that before the third data corresponding to each candidate signal is input into the second identification model, the third data needs to be processed. Specifically, for each candidate signal, the terminal device can perform delay time correction processing on the third data corresponding to the candidate signal to obtain processed data corresponding to the candidate signal.

[0111] The purpose of this is that a fast radio burst often contains signals of multiple frequencies, some of which are high and some of which are low. Because the signal will disperse when it propagates in the universe, that is, the propagation speed of radio signals of different frequencies is different, so in a fast radio burst, some signals of high frequency will arrive on Earth first, and some will arrive after a period of time.

[0112] Because of the dispersion, in order to accurately determine the arrival time of the signal of the fast radio burst, the third data corresponding to the candidate signal needs to be processed for delay time, so that in the processed data, the arrival time of the fast radio burst can be concentrated as much as possible at one time or a small time range, as shown in Figure 4A and Figure 4B .

[0113] Figure 4A and Figure 4B Data before and after delay time processing provided for this specification.

[0114] As can be seen from Figure 4A , the time of the signal observed is earlier than that of the signal of low frequency, but these actually belong to one signal of a fast radio burst, so delay time correction is needed to obtain the data shown in Figure 4B . As can be seen from Figure 4B , after delay time correction processing, the signals of the fast radio burst whose arrival time is lengthened are concentrated in a small time range.

[0115] After the above delay time correction is completed, the terminal device can input the processed data corresponding to each candidate signal into the second identification model to obtain the identification result for the candidate signal.

[0116] In addition, before identifying the fast radio burst using the first and second identification models, the two models need to be trained. The training processes of the two identification models will be introduced respectively.

[0117] In training the first identification model, the first sample data can be obtained according to the arrival time of the historical fast radio burst, that is, the first sample data can be obtained according to the data of the fast radio burst observed in history. In addition to the data of the historical fast radio burst, the first sample data can also include other noise data.

[0118] After obtaining the first sample data, the historical first data can be determined, and the way of determining the historical first data is basically the same as the way of determining the first data, which will not be described in detail here. The corresponding historical first data can be understood as representing the relationship between the frequency of each signal observed in history and the time of each signal observed in history.

[0119] After determining the historical first data, the historical second data can be further determined according to the historical first data, and the way of determining the historical second data is basically the same as the way of determining the second data, which will not be described in detail here. The second historical data can be used to represent the relationship between the dispersion value of each signal observed and the time of each signal observed in history.

[0120] The historical second data can be input into the first identification model to be trained to obtain the identification result corresponding to the first sample data, which can be understood as the first identification model identifying the arrival time of the historical fast radio burst from the historical second data.

[0121] Finally, the deviation between the identification result corresponding to the first sample data and the actual arrival time of the historical fast radio burst can be minimized as the optimization target to train the first identification model, that is, the loss value can be determined according to the deviation between the arrival time of the historical fast radio burst identified by the first identification model and the actual arrival time of the historical fast radio burst, and then the first identification model is trained by minimizing the loss value as the optimization target.

[0122] As for the training method of the second identification model, the second sample data can be obtained first, wherein the second sample data includes actual historical fast radio bursts and noise data that is suspected to be a fast radio burst but is not actually a fast radio burst.

[0123] Then, the third data of each historical candidate signal can be determined from the historical observation data according to the dispersion value and the arrival time of the historical candidate signal in the second sample data. The third data of the historical candidate signal is determined in a manner substantially the same as the third data of the candidate signal, and thus will not be described in detail here.

[0124] The third data corresponding to each historical candidate signal can be input into the second identification model to obtain an identification result corresponding to each historical candidate signal. Finally, the second identification model is trained with the optimization objective of minimizing the deviation between the identification result corresponding to each historical candidate signal and the label corresponding to each historical candidate signal. The label corresponding to each historical candidate signal is used to indicate whether the historical candidate signal is actually a historical fast radio burst.

[0125] That is, after determining the identification result corresponding to each historical candidate signal, the loss value corresponding to each historical candidate signal can be determined according to the deviation between the identification result corresponding to the historical candidate signal and the label corresponding to the historical candidate signal, and then the second identification model is trained with the optimization objective of minimizing the sum of the loss values corresponding to each historical candidate signal.

[0126] It should be noted that in the present specification, the execution subject performing the training process of the first identification model and the second identification model can be a terminal device such as a desktop computer, a notebook computer, or a server, and the present specification does not limit the specific execution subject.

[0127] In addition, in order to further improve the identification ability of the first identification model, in the present specification, each data segment can be determined from the historical observation data according to the arrival time of the historical fast radio burst. The determination of each data segment can be performed in the following manner: according to the arrival time, the historical observation data is divided into data segments with a predetermined step size. The data segments obtained here can be understood as some containing fast radio bursts and some not containing fast radio bursts. The data segments containing fast radio bursts can be used to construct positive samples, and the data segments not containing fast radio bursts can be used to construct negative samples.

[0128] That is, according to the arrival time of the fast radio burst, the time-frequency data of the fast radio burst is cut out from the historical observation data, and since this data is cut out from the historical observation data, it can be called a data segment, and the front end and the rear end of the data segment can each reserve a segment of noise data. Then, the cut-out data segment can be used to construct positive samples in subsequent processes, and the end of the cut-out data segment can be used as the starting point to cut out a data segment of the same length as the fast radio burst data segment for constructing negative samples.

[0129] After determining each data segment, the terminal device can perform data sampling on each data segment according to a preset data sampling mode to obtain sampled data, and then the terminal device can perform delay time correction on the obtained sampled data according to a preset dispersion range to obtain corrected sampled data. The terminal device can superimpose the corrected data on the frequency of the signal to obtain first sample data in each frequency range.

[0130] For example, the terminal device can perform 4-fold, 8-fold and 16-fold downsampling on the obtained data segments respectively to obtain sampled data. Then, the sampled data is corrected by time delay in a manner that the dispersion value is from 1 to 1024 with a step of 1. Finally, the corrected sampled data is superimposed on the frequency of the signal, and the data set A superimposed on 2048 frequency channels is saved, and the superposition is continued to 4096 frequency channels, and the time-dispersion data set B is saved. The difference set of the data set B and the data set A is obtained to obtain the time-dispersion data superimposed on only 2048 frequency channels. Ultimately, the sample data can be multiplied by the augmentation in this way, and the sample data obtained after the augmentation actually contains low-frequency, high-frequency and full-frequency data.

[0131] If the data segment used to construct the positive sample is frequency superimposed in the above example, a plurality of positive samples containing the fast radio burst will also be obtained, and correspondingly, if the data segment used to construct the negative sample is frequency superimposed in the above example, a plurality of negative samples will also be obtained.

[0132] Through the above-mentioned manner, samples of different frequencies and different sampling rates can be obtained, so that the generalization ability of the first recognition model can be effectively improved by training the first recognition model through these samples, thereby further improving the recognition accuracy of the first recognition model.

[0133] It should be noted that in the process of augmenting the sample data, a multi-thread parallel mode can be used for execution, that is, when sampling data according to each data sampling mode, at least one thread can be used for each sampling mode to achieve parallel data sampling. The delay time correction according to the preset dispersion range can also be processed in parallel through multi-thread mode, that is, when the delay time correction is performed according to 5 dispersion ranges, 5 threads can be used for parallel processing, and one thread is responsible for processing the delay time correction under one dispersion range. Among them, these steps can be implemented by GPU for multi-thread parallel processing.

[0134] In this specification, the specific algorithm used by the first recognition model and the second recognition model can be a conventional neural network algorithm, for example, the first recognition model can use the YOLO v5 algorithm, and the second recognition model can use ResNet50. Of course, the first recognition model and the second recognition model can also use other algorithms that can be applied to identify fast radio bursts, so the specific algorithm used by the first recognition model and the second recognition model is not limited in this specification.

[0135] S105: According to the recognition result, the signal of the fast radio burst is recognized from the observation data.

[0136] After obtaining the above recognition result, the terminal device can actually determine the arrival time of the fast radio burst, and then the signal arriving at the arrival time can be further searched from the obtained observation data, and the signal of the fast radio burst is identified.

[0137] In the prior art, using the traditional fast radio burst search method (such as Presto, Heimdall), first, according to the threshold of signal intensity, the position of noise signal (i.e. radio frequency interference) in the two-dimensional data of time-frequency of each signal is blocked (although such a way cannot completely remove the radio frequency interference), and then a certain dispersion value is selected for delay time correction. After that, the frequencies are superimposed to only keep one-dimensional time sequence, and a complex signal-to-noise ratio calculation method is applied in the one-dimensional time sequence to search for candidate signals with signal intensity exceeding the threshold. Then, the dispersion value is modified, the delay time is corrected, and the subsequent process is continued. After multiple iterations of the search step, the search for the fast radio burst is finally completed.

[0138] Although this method can search for the signal of the fast radio burst from the observation data, it needs to calculate the signal-to-noise ratio and select the signal for different dispersion values, so the calculation amount is large and the repeated calculation process is more, and the selection of different dispersion value ranges also has a great impact on the calculation amount.

[0139] But in the method for identifying a fast radio burst provided in the specification, after the observation data is obtained, the second data used for input into the first identification model can be obtained based on the observation data, so that the signal of the fast radio burst is quickly and accurately identified from the observation data through the first identification model, the whole process does not need a complex calculation process, thereby effectively improving the search efficiency of the fast radio burst.

[0140] And, after combining the above-mentioned second identification model, the signal of the fast radio burst can be accurately searched from the observation data containing noise signals, thereby improving the search accuracy of the fast radio burst.

[0141] The above is the method for identifying a fast radio burst provided by one or more embodiments of the specification, based on the same idea, the specification also provides a corresponding device for identifying a fast radio burst, as shown in Figure 5 .

[0142] Figure 5 A device for identifying a fast radio burst provided in the specification includes:

[0143] The acquisition module 501 is configured to acquire observation data.

[0144] The first determination module 502 is configured to determine first data according to the observation data, and the first data is used to represent the relationship between the frequency of each observed signal and the time when each signal is observed.

[0145] The second determination module 503 is configured to determine second data according to the first data, and the second data is used to represent the relationship between the dispersion value of each observed signal and the time when each signal is observed.

[0146] The input module 504 is configured to input the second data into a pre-trained first identification model to obtain an identification result output by the first identification model.

[0147] The identification module 505 is configured to identify the signal of the fast radio burst from the observation data according to the identification result.

[0148] Optionally, the input module 504 is specifically configured to input the second data into the first identification model to determine the arrival time of each candidate signal in the observation data as the identification result output by the first identification model through the first identification model.

[0149] The identification module 505 is specifically configured to: for each candidate signal, determine third data corresponding to the candidate signal according to an arrival time of the candidate signal and the observation data, the third data being used to represent a relationship between a frequency of the candidate signal and each time instant at which the candidate signal is observed; input the third data corresponding to the candidate signal into a second identification model that is pre-trained, to obtain an identification result for the candidate signal; and identify a signal of a fast radio burst from the observation data according to the identification result for each candidate signal.

[0150] Optionally, the identification module 505 is further configured to: before inputting the third data corresponding to the candidate signal into the second identification model that is pre-trained, to obtain the identification result for the candidate signal, perform delay time correction processing on the third data corresponding to the candidate signal, to obtain processed data corresponding to the candidate signal; and input the processed data corresponding to the candidate signal into the second identification model, to obtain the identification result for the candidate signal.

[0151] Optionally, the second determination module 503 is specifically configured to: for each preset dispersion range, perform delay time correction on the first data according to the dispersion range, to obtain modified data corresponding to the dispersion range; and superimpose the modified data corresponding to each dispersion range, to obtain the second data.

[0152] Optionally, before inputting the second data into the first identification model that is pre-trained, to obtain the identification result output by the first identification model, the input module 504 is further configured to: perform data segmentation on the second data in a time dimension and a dispersion dimension, to obtain data blocks.

[0153] The input module 504 is specifically configured to: input the data blocks into the first identification model, to obtain the identification result output by the first identification model.

[0154] Optionally, the apparatus further includes:

[0155] The first training module 506 is configured to: obtain first sample data according to arrival times of historical fast radio bursts; determine historical first data according to the first sample data; determine historical second data according to the historical first data; input the historical second data into a first identification model to be trained, to obtain an identification result corresponding to the first sample data; and train the first identification model, with the optimization objective being to minimize a deviation between the identification result corresponding to the first sample data and actual arrival times of the historical fast radio bursts.

[0156] Optionally, the first training module 506 is specifically configured to determine each data segment from the historical observation data according to the arrival time of the historical fast radio burst; sample the data in each data segment according to a preset data sampling mode to obtain sampled data; correct the delay time of the sampled data according to a preset dispersion range to obtain corrected sampled data; and superimpose the corrected data on the frequency of the signal to obtain first sample data of each frequency range.

[0157] Optionally, the first training module 506 is specifically configured to slice the historical observation data according to a preset step size according to the arrival time of the historical fast radio burst to obtain each data segment.

[0158] Optionally, the apparatus further comprises:

[0159] The second training module 507 is configured to obtain second sample data; determine third data of each historical candidate signal from the historical observation data according to the dispersion value and the arrival time of the historical candidate signal in the second sample data; input the third data corresponding to each historical candidate signal into the second identification model to obtain an identification result corresponding to each historical candidate signal; and train the second identification model with the optimization target of minimizing the deviation between the identification result corresponding to each historical candidate signal and a label corresponding to the historical candidate signal, wherein the label corresponding to each historical candidate signal is used to indicate whether the historical candidate signal is actually a historical fast radio burst.

[0160] The present specification also provides a computer-readable storage medium storing a computer program, and the computer program can be used to execute the above Figure 1 The present specification provides a method for identifying a fast radio burst.

[0161] The present specification also provides a computer program product Figure 6 The schematic structural diagram of the electronic device is shown in FIG. 6. As shown in FIG. 6, at the hardware level, the electronic device comprises a processor, an internal bus, a network interface, a memory and a non-volatile memory, and of course, other hardware required by the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs to implement the above Figure 1 The method for identifying a fast radio burst. Of course, in addition to the software implementation, the present specification does not exclude other implementation manners, such as a logic device or a combination of software and hardware, and the like, that is, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or a logic device.

[0162] In the 1990s, it was quite obvious to distinguish whether an improvement in a technology was in hardware (e.g., improvement in circuit structures of diodes, transistors, switches, etc.) or in software (improvement in method flow). However, as technology has evolved, many improvements in method flow today can be considered as direct improvements in hardware circuit structures. Designers almost always obtain the corresponding hardware circuit structures by programming the improved method flow into hardware circuits. Therefore, it cannot be said that an improvement in a method flow cannot be implemented by hardware entity modules. For example, a programmable logic device (PLD) such as a field programmable gate array (FPGA) is an integrated circuit whose logic function is determined by user programming of the device. A digital system is "integrated" on a PLD by the designer programming it, rather than by ordering a chip manufacturer to design and fabricate a custom integrated circuit chip. Moreover, instead of manually fabricating integrated circuit chips, this programming is now mostly implemented by "logic compiler" software, which is similar to software compilers used in program development, and the original code before compilation must also be written in a specific programming language, which is called a hardware description language (HDL), and there are many such languages, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc., and the most commonly used are VHDL (Very-High-Speed Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should be aware that, as long as the method flow is logically programmed in the above-mentioned hardware description languages and programmed into an integrated circuit, a hardware circuit that implements the logical method flow can be easily obtained.

[0163] The controller can be implemented in any suitable way, for example, the controller can take the form of a microprocessor or processor and a computer readable medium storing computer readable program code, such as software or firmware, executable by the (micro)processor, logic gates, switches, an application specific integrated circuit (ASIC), a programmable logic controller and an embedded microcontroller, examples of which include but are not limited to the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20 and Silicone Labs C8051F320, the memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art will also know that, in addition to being implemented in pure computer readable program code, the controller can equally well be implemented to perform the same functions using logic gates, switches, an application specific integrated circuit, a programmable logic controller and an embedded microcontroller, etc. by means of a logical programming of the method steps. The controller can thus be considered as a hardware component, and the means comprised therein for performing the various functions can be considered as structures within the hardware component. Alternatively, the means for performing the various functions can even be considered as both a software module implementing the method and a structure within the hardware component.

[0164] The systems, apparatuses, modules or units illustrated by the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer can be a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0165] For the sake of description, the above apparatuses are described in various units by functions respectively. Of course, the functions of each unit can be implemented in one or more software and / or hardware in implementing the present specification.

[0166] Those skilled in the art will understand that the embodiments of the present application can be provided as a method, a system or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program code.

[0167] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks in the flowcharts and / or combination thereof. Figure 1 one or more flowcharts and / or blocks in the flowcharts and / or combination thereof.

[0168] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks in the flowcharts and / or combination thereof. Figure 1 one or more flowcharts and / or blocks in the flowcharts and / or combination thereof.

[0169] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks in the flowcharts and / or combination thereof. ​ one or more flowcharts and / or blocks in the flowcharts and / or combination thereof.

[0170] In one typical configuration, the computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0171] The memory can include non-persistent memory and / or volatile memory, such as random access memory (RAM) and / or cache memory, non-volatile memory, such as read-only memory (ROM), EPROM, and / or flash memory. The memory is an example of computer-readable media.

[0172] Computer-readable media includes permanent and non-permanent, movable and non-movable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to a computing device. According to the definition herein, computer-readable media does not include transitory media such as modulated data signals and carriers.

[0173] It should also be noted that the terms "comprising", "containing", or any other variant thereof are intended to cover non-exclusive inclusion, such that a process, method, article or apparatus that comprises a list of elements does not only include those elements, but also other elements not explicitly listed or inherent to such process, method, article or apparatus. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article or apparatus that includes the element.

[0174] Those skilled in the art will appreciate that embodiments of the present specification can be provided as methods, systems or computer program products. Therefore, the present specification can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Moreover, the present specification can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) containing computer-usable program code.

[0175] The present specification can be described in the general context of computer-executable instructions, such as program modules, executed by computers. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The present specification can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in both local and remote computer storage media including storage devices.

[0176] Each of the embodiments in the specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. Each of the embodiments focuses on the difference from other embodiments. In particular, for the system embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the part of the method embodiments.

[0177] The above only describes the embodiments of the specification and is not used to limit the specification. The specification can have various changes and variations for those skilled in the art. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the specification shall be included in the scope of the claims of the application.

Claims

1. A method for identifying fast radio bursts, characterized in that, include: Acquire observation data; Based on the observation data, first data is determined, which is used to represent the relationship between the frequency of each observed signal and the time when each signal was observed. Based on the first data, second data is determined, which is used to represent the relationship between the dispersion value of each observed signal and the time when each signal was observed. The process of inputting the second data into a pre-trained first recognition model to obtain the recognition result output by the first recognition model includes: inputting the second data into the first recognition model to determine the arrival time of each candidate signal in the observation data through the first recognition model, and using it as the recognition result output by the first recognition model; Based on the identification results, the signal of a fast radio burst is identified from the observation data; this includes: for each candidate signal, determining third data corresponding to the candidate signal based on the arrival time of the candidate signal and the observation data, wherein the third data is used to represent the relationship between the frequency of the candidate signal and the times at which the candidate signal is observed; inputting the third data corresponding to the candidate signal into a pre-trained second identification model to obtain the identification result for the candidate signal; and identifying the signal of a fast radio burst from the observation data based on the identification result of each candidate signal; wherein the identification result of the second identification model is the classification result of the fast radio burst.

2. The method as described in claim 1, characterized in that, Before inputting the third data corresponding to the candidate signal into a pre-trained second recognition model to obtain the recognition result for the candidate signal, the method further includes: The third data corresponding to the candidate signal is subjected to a time delay correction process to obtain the processed data corresponding to the candidate signal. The third data corresponding to the candidate signal is input into a pre-trained second recognition model to obtain the recognition result for the candidate signal, specifically including: The processed data corresponding to the candidate signal is input into the second recognition model to obtain the recognition result for the candidate signal.

3. The method as described in claim 1, characterized in that, Based on the first data, the second data is determined, specifically including: For each preset dispersion range, the first data is delayed according to the dispersion range to obtain the corrected data corresponding to the dispersion range; The corrected data corresponding to each dispersion range are superimposed to obtain the second data.

4. The method as described in claim 1, characterized in that, Before inputting the second data into a pre-trained first recognition model to obtain the recognition result output by the first recognition model, the method further includes: The second data is segmented along the time and dispersion dimensions to obtain data blocks. The second data is input into a pre-trained first recognition model to obtain the recognition result output by the first recognition model, specifically including: The data blocks are input into the first recognition model to obtain the recognition results output by the first recognition model.

5. The method as described in claim 1, characterized in that, Pre-training the first recognition model specifically includes: The first sample data was obtained based on the arrival times of historical fast radio bursts; Based on the first sample data, determine the first historical data; Based on the first historical data, determine the second historical data; The historical second data is input into the first recognition model to be trained to obtain the recognition result corresponding to the first sample data; The first identification model is trained with the optimization objective of minimizing the deviation between the identification result corresponding to the first sample data and the actual arrival time of the historical fast radio bursts.

6. The method as described in claim 5, characterized in that, Based on the arrival times of historical fast radio bursts, the first sample of data was obtained, specifically including: Based on the arrival times of the historical fast radio bursts, each data segment is determined from the historical observation data; According to the preset data sampling methods, data segments are sampled to obtain sampled data; According to the preset dispersion range, the sampling data is delayed and corrected to obtain the corrected sampling data; The corrected data is superimposed on the frequency of the signal to obtain the first sample data for each frequency range.

7. The method as described in claim 6, characterized in that, Based on the arrival times of the historical fast radio bursts, data segments are determined from historical observation data, specifically including: Based on the arrival time of the historical fast radio bursts, the historical observation data is segmented according to a preset step size to obtain each data segment.

8. The method as described in claim 1, characterized in that, Pre-training the second recognition model specifically includes: Obtain the second sample data; Based on the dispersion value and arrival time of the historical candidate signals in the second sample data, the third data of each historical candidate signal is determined from the historical observation data. The third data corresponding to each historical candidate signal is input into the second recognition model to obtain the recognition result for each historical candidate signal; The second identification model is trained with the optimization objective of minimizing the deviation between the identification result corresponding to each historical candidate signal and the label corresponding to each historical candidate signal. For each historical candidate signal, the label corresponding to the historical candidate signal is used to indicate whether the historical candidate signal is actually a historical fast radio burst.

9. A device for identifying fast radio bursts, characterized in that, include: The acquisition module is used to acquire observation data; The first determining module is used to determine first data based on the observation data, wherein the first data is used to represent the relationship between the frequency of each observed signal and the time when each signal is observed; The second determining module is used to determine second data based on the first data, wherein the second data is used to represent the relationship between the dispersion value of each observed signal and the time when each signal was observed; An input module is used to input the second data into a pre-trained first recognition model to obtain the recognition result output by the first recognition model; including: inputting the second data into the first recognition model to determine the arrival time of each candidate signal in the observation data through the first recognition model, and using it as the recognition result output by the first recognition model; An identification module is used to identify fast radio burst signals from the observation data based on the identification results; including: for each candidate signal, determining third data corresponding to the candidate signal based on the arrival time of the candidate signal and the observation data, wherein the third data is used to represent the relationship between the frequency of the candidate signal and the times when the candidate signal is observed; inputting the third data corresponding to the candidate signal into a pre-trained second identification model to obtain an identification result for the candidate signal; and identifying fast radio burst signals from the observation data based on the identification results of each candidate signal; wherein the identification result of the second identification model is the classification result of fast radio bursts.

10. The apparatus as claimed in claim 9, characterized in that, The recognition module is further configured to perform a time delay correction process on the third data corresponding to the candidate signal before inputting the third data corresponding to the candidate signal into the pre-trained second recognition model to obtain the recognition result for the candidate signal, so as to obtain the processed data corresponding to the candidate signal. The processed data corresponding to the candidate signal is input into the second recognition model to obtain the recognition result for the candidate signal.

11. The apparatus as claimed in claim 9, characterized in that, The second determining module is specifically used to perform a delay time correction on the first data according to each preset dispersion range to obtain the corrected data corresponding to the dispersion range. The corrected data corresponding to each dispersion range are superimposed to obtain the second data.

12. The apparatus as claimed in claim 9, characterized in that, Before inputting the second data into the pre-trained first recognition model and obtaining the recognition result output by the first recognition model, the input module is further used to divide the second data into data blocks in the time dimension and dispersion dimension. The input module is specifically used to input the data into the first recognition model in blocks to obtain the recognition result output by the first recognition model.

13. The apparatus as claimed in claim 9, characterized in that, The device further includes: The first training module is used to acquire first sample data based on the arrival time of historical fast radio bursts; determine historical first data based on the first sample data; determine historical second data based on the historical first data; input the historical second data into the first recognition model to be trained to obtain the recognition result corresponding to the first sample data; and train the first recognition model with the optimization objective of minimizing the deviation between the recognition result corresponding to the first sample data and the actual arrival time of the historical fast radio bursts.

14. The apparatus as claimed in claim 13, characterized in that, The first training module is specifically used to determine each data segment from historical observation data based on the arrival time of the historical fast radio bursts; and to sample each data segment according to preset data sampling methods to obtain sampled data. According to the preset dispersion range, the sampling data is delayed and corrected to obtain the corrected sampling data; The corrected data is superimposed on the frequency of the signal to obtain the first sample data for each frequency range.

15. The apparatus as claimed in claim 14, characterized in that, The first training module is specifically used to divide the historical observation data into segments according to the arrival time of the historical fast radio bursts and a preset step size to obtain each data segment.

16. The apparatus as claimed in claim 9, characterized in that, The device further includes: The second training module is used to acquire second sample data; determine the third data of each historical candidate signal from the historical observation data based on the dispersion value and arrival time of the historical candidate signals in the second sample data; input the third data corresponding to each historical candidate signal into the second recognition model to obtain the recognition result corresponding to each historical candidate signal; train the second recognition model with the optimization objective of minimizing the deviation between the recognition result corresponding to each historical candidate signal and the label corresponding to each historical candidate signal, wherein, for each historical candidate signal, the label corresponding to the historical candidate signal is used to indicate whether the historical candidate signal is actually a historical fast radio burst.

17. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the method described in any one of claims 1 to 8.

18. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method described in any one of claims 1 to 8.

Images