Information processor, magnetic recording and playback device, and magnetic recording and playback system

Abstract

To provide an information processor, a magnetic recording and playback device, and a magnetic recording and playback system capable of improving processing accuracy.SOLUTION: According to an embodiment, an information processor includes an acquisition part and a processing part. The acquisition part is capable of acquiring a playback signal based on a signal obtained from a magnetic recording medium. The playback signal includes first input data, second input data, and third input data. The processing part is capable of carrying out first processing and second processing. In the first processing, the processing part derives a first derived characteristic amount from a first inner product of a first characteristic amount based on the first input data and a second characteristic amount based on the second input data. In the second processing, the processing part derives a second derived characteristic amount from a second inner product of a third characteristic amount based on the third input data and the first derived characteristic amount.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] An embodiment of the present invention relates to an information processing device, a magnetic recording and reproducing device, and a magnetic recording and reproducing system. [Background technology]

[0002] For example, information recorded on a magnetic recording medium is reproduced by an information processing device. For example, by improving the processing accuracy of the information processing device, the magnetic recording density can be improved. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] US Patent Application Publication No. 2020 / 0389188 Summary of the Invention [Problem to be solved by the invention]

[0004] The embodiments of the present invention provide an information processing device, a magnetic recording and reproducing device, and a magnetic recording and reproducing system that are capable of improving processing accuracy. [Means for solving the problem]

[0005] According to an embodiment of the present invention, an information processing device includes an acquisition unit and a processing unit. The acquisition unit is capable of acquiring a reproduction signal based on a signal obtained from a magnetic recording medium. The reproduction signal includes first input data, second input data, and third input data. The processing unit is capable of performing a first process and a second process. In the first process, the processing unit derives a first derived feature from a first inner product of a first feature based on the first input data and a second feature based on the second input data. In the second process, the processing unit derives a second derived feature from a second inner product of a third feature based on the third input data and the first derived feature. [Brief description of the drawings]

[0006] [Figure 1] FIG. 1 is a schematic diagram illustrating an information processing device and an information recording / reproducing device according to the first embodiment. [Diagram 2] FIG. 2 is a schematic diagram illustrating a part of the information processing device according to the first embodiment. [Diagram 3] FIG. 3 is a schematic diagram illustrating a part of the information processing device according to the first embodiment. [Figure 4] FIG. 4 is a schematic diagram illustrating an information processing device and an information recording / reproducing device according to the second embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0007] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In this specification and each drawing, elements similar to those described above with reference to the previous drawings are given the same reference numerals and detailed descriptions thereof will be omitted as appropriate.

[0008] (First embodiment) FIG. 1 is a schematic diagram illustrating an information processing device and an information recording / reproducing device according to the first embodiment. 1, the magnetic recording and reproducing device 210 according to the embodiment includes an information processing device 70, a magnetic recording medium 80, and a reproducing unit 80R. The magnetic recording medium 80 and the reproducing unit 80R may be included in a recording unit 80D, which may be included in, for example, a magnetic disk (HDD: Hard Disk Drive). The magnetic recording and reproducing device 210 may further include, for example, an SSD (Solid State Drive).

[0009] The reproducing unit 80R can reproduce a signal 80S corresponding to information recorded on the magnetic recording medium 80. The reproducing unit 80R includes, for example, a magnetic head 80H. A signal 80S corresponding to the reproduced information is output from the reproducing unit 80R. The signal 80S is an electrical signal. Information may be recorded on the magnetic recording medium 80 by the magnetic head 80H.

[0010] The information processing device 70 includes an acquisition unit 70i and a processing unit 70p. The acquisition unit 70i is capable of acquiring a reproduction signal RS based on a signal 80S obtained from a magnetic recording medium 80. The reproduction signal RS may be based on a signal obtained by amplifying the signal 80S. The reproduction signal RS may be a digital signal obtained by AD (Analog to Digital) converting the signal 80S. The acquisition unit 70i may be, for example, an interface.

[0011] The reproduction signal RS acquired by the acquisition unit 70i is processed by the processing unit 70p. The processing unit 70p includes, for example, an input unit 70pi and an output unit 70po. The reproduction signal RS is supplied to the input unit 70pi. The processing result of the processing unit 70p is output from the output unit 70po.

[0012] 1, the reproduction signal RS includes first input data 11, second input data 12, and third input data 13. The processing unit 70p is capable of performing a first process PR1 and a second process PR2. In one example, the first input data 11 includes data related to bits to be decoded. For example, the second input data 12 and the third input data 13 include data related to other bits.

[0013] In the first process PR1, the processing unit 70p derives a first derived feature FD1 from a first inner product 31 of a first feature F11 based on the first input data 11 and a second feature F12 based on the second input data 12.

[0014] In the second process PR2, the processing unit 70p derives a second derived feature FD2 from a second inner product 32 of a third feature F13 based on the third input data 13 and the first derived feature FD1.

[0015] By performing the first process PR1 and the second process PR2, the reproduction signal RS corresponding to the information recorded on the magnetic recording medium 80 can be decoded with high accuracy.

[0016] As shown in FIG. 1, the processing unit 70p can further perform a third process PR3. In the third process PR3, the processing unit 70p outputs a bit decoded result BD1 for the decoding target bit included in the first input data 11 based on at least a part of the second derived feature FD2. The obtained bit decoded result BD1 is output from the output unit 70po. In one example, the bit decoded result BD1 corresponds to information related to 0 / 1 judgment. The information related to 0 / 1 judgment may be, for example, 0 or 1. In one example, the bit decoded result BD1 may be information usable for 0 / 1 judgment. The information usable for 0 / 1 judgment may include, for example, the probability of becoming 0 or 1. The information usable for 0 / 1 judgment may be, for example, the likelihood for 0 or 1.

[0017] The processing unit 70p is capable of performing neural network processing (NN processing). As shown in Fig. 1, the processing unit 70p may include an attention layer 71 and a classification layer 72. The attention layer 71 is capable of performing a first process PR1 and a second process PR2. The classification layer 72 is capable of performing a third process PR3. The classification layer 72 includes a fully connected neural network (FC-NN).

[0018] In embodiments, a neural network may be a component that includes parameters whose values ​​are adjustable (eg, learnable) based on information, such as data, etc. In embodiments, a neural network may include a perceptron.

[0019] In the embodiment, the reproduction signal RS obtained from the magnetic recording medium 80 is processed by attention processing including two inner product operations. By such processing, feature extraction of the waveform data of the reproduction signal RS is performed (first processing PR1 and second processing PR2). Then, after the feature extraction, classification processing is performed by a fully connected NN. The classification processing corresponds to, for example, a process of using information extracted by the feature extraction and converting it into information related to 0 / 1 determination. This allows, for example, decoding with high accuracy. For example, decoding can be performed quickly and with high accuracy. It is possible to provide an information processing device capable of improving processing accuracy.

[0020] In the embodiment, a feature amount that takes into account the relationship between input data can be appropriately extracted by calculation based on the inner product. In processing based on the inner product, parallel processing is easy. In the embodiment, high speed parallel calculation and appropriate feature amount allow high-precision decoding to be performed quickly. For example, in the NN model of the reference example that takes into account the relationship between input data, a Recurrent Neural Network (RNN) (e.g., LSTM: Long Short Term Memory) is applied. In such a reference example, parallel calculation is difficult, and therefore the calculation time is long. In contrast, in the embodiment, high-speed processing is possible.

[0021] In the embodiment, the first input data 11, the second input data 12, the third input data 13, etc. correspond to information recorded on the magnetic recording medium 80. For example, the first input data 11 includes data related to the bit to be decoded. The second input data 12 includes data related to the first other bit. The third input data 13 includes data related to the second other bit. In one example, the bit to be decoded is between the first other bit and the second other bit. In one example, the bit to be decoded may be next to the first other bit. The bit to be decoded may be next to the second other bit. Another bit may be provided between the bit to be decoded and the first other bit. Another bit may be provided between the bit to be decoded and the second other bit. The first other bit may include one or more bits. The second other bit may include one or more bits. In another example, the first other bit may be provided between the bit to be decoded and the second other bit.

[0022] At least one of the first input data 11, the second input data 12, and the third input data 13 may be determined, for example, by a method of attention processing. As will be described later, for example, self-attention may be applied.

[0023] 1, for example, the first feature F11 includes a "query" related to the first input data 11. For example, the second feature F12 includes a "key" related to the second input data 12.

[0024] 1, the attention layer 71 may include a first neural network 11N and a second neural network 12N. The attention layer 71 can derive the above-mentioned "query" (first feature F11) by processing the first input data 11 with the first neural network 11N. The attention layer 71 can derive the above-mentioned "key" (second feature F12) by processing the second input data 12 with the second neural network 12N.

[0025] The first derived feature FD1 includes, for example, a first weight indicating the relationship between the first input data 11 and the second input data 12.

[0026] For example, the third feature F13 includes a "value" related to the third input data 13. For example, the attention layer 71 further includes a third neural network 13N. The attention layer 71 can derive the above-mentioned "value" (the third feature F13) by processing the third input data 13 with the third neural network 13N.

[0027] 1, the attention layer 71 may further include a fourth neural network 14N. The attention layer 71 can derive a second derived feature FD2 by processing the second inner product 32 with the fourth neural network 14N.

[0028] In the embodiment, for example, in signal reproduction of a HDD, interference between bits is taken into consideration. In the attention process in the embodiment, the relationship of input data is taken into consideration and feature amounts can be extracted. By using the attention process in signal reproduction of a HDD, signal reproduction can be performed taking into consideration the interference (relationship) between bits.

[0029] In this manner, in the embodiment, the reproduction signal RS obtained from the magnetic recording medium 80 is processed by the attention process including two inner product calculations. This allows decoding with high accuracy, and provides an information processing device capable of improving processing accuracy.

[0030] The number of attention layers 71 may be one, or may be two or more.

[0031] As shown in Fig. 1, the processing unit 70p may be capable of performing an embedding process 71e. For example, the processing unit 70p performs an embedding process 71e of the reproduction signal RS (the first input data 11, the second input data 12, and the third input data 13). The embedding process may include, for example, embedding of position information. The embedding process may include, for example, a class token used when classifying input data for performing attention processing.

[0032] FIG. 2 is a schematic diagram illustrating a part of the information processing device according to the first embodiment. 2 shows an example of the configuration of the attention layer 71. The processing unit 70p performs the above-mentioned first process PR1 after the embedding process 71e.

[0033] As shown in Fig. 2, the processing unit 70p may be capable of performing a multi-head attention process 71m. For example, the attention layer 71 may divide at least one of the first input data 11, the second input data 12, and the third input data 13 into a plurality of vectors. The attention layer 71 may be capable of performing processes on at least two of the plurality of vectors in parallel. For example, the processing result can be obtained at high speed.

[0034] 2, the processing unit 70p may include normalization layers (such as a first normalization layer 71a and a second normalization layer 71b). The normalization layers may be capable of performing a normalization process in at least a portion of at least one of the first process PR1 and the second process PR2.

[0035] 2, after processing by the first normalization layer 71a, processing on at least two of the plurality of vectors (multi-head attention processing 71m) can be performed, and after processing on at least two of the plurality of vectors, processing by the second normalization layer 71b can be performed.

[0036] 2, the processing unit 70p may include a first skip connection unit 71p and a second skip connection unit 71q. The first skip connection unit 71p, for example, connects the result of a process (multi-head attention process 71m) for at least two of the multiple vectors to a first value set 71u input to the first normalization layer 71a. The connection may include, for example, a residual connection. In the case of the residual connection, for example, a result of adding the two is supplied to the second normalization layer 71b.

[0037] The second skip connection unit 71q, for example, adds together the result of processing by the second normalization layer 71b through the neural network 71n and the second value group 71v input to the second normalization layer 71b. Through this addition, a second derived feature FD2 (see FIG. 1) is derived.

[0038] FIG. 3 is a schematic diagram illustrating a part of the information processing device according to the first embodiment. Figure 3 shows one example of the configuration of the attention layer 71. As shown in Figure 4, a processing unit 70p may be capable of implementing the embedding process 71e described with respect to Figure 2.

[0039] For example, at least one of the first input data 11, the second input data 12, and the third input data 13 is divided into a plurality of vectors in the attention layer 71. As shown in Fig. 4, for example, after the embedding process 71e, a process (multi-head attention process 71m) for at least two of the plurality of vectors is performed.

[0040] The process for at least two of the multiple vectors (multi-head attention process 71m) is normalized, for example, in the first normalization layer 71a. In this example, the process by the first normalization layer 71a is performed at least after the first process PR1.

[0041] For example, the first skip connection unit 71p connects the result of the processing in the first normalization layer 71a to a first value group 71u input to the multihead attention processing unit 71m.

[0042] 3, the result of the processing by the first skip connection unit 71p is processed by the neural network 71n. The result of the processing by the neural network 71n is normalized by the second normalization layer 71b.

[0043] The second skip connection unit 71q, for example, adds the result of the processing in the second normalization layer 71b and the second value group 71v input to the neural network 71n. This addition process derives a second derived feature FD2 (see FIG. 1).

[0044] Second embodiment FIG. 4 is a schematic diagram illustrating an information processing device and an information recording / reproducing device according to the second embodiment. 4, the information processing device 70A according to the embodiment also includes an acquisition unit 70i and a processing unit 70p. In the operation of the information processing device 70A, self-attention is performed. Except for this, the configuration and operation of the information processing device 70A may be similar to the configuration and operation of the information processing device 70.

[0045] 4, the reproduction signal RS includes first input data 11. The processing unit 70p is capable of carrying out a first process PR1 and a second process PR2.

[0046] In the first process PR1, the processing unit 70p derives a first derived feature FD1 from a first inner product 31 of a first feature F11 based on the first input data 11 and a second feature F12 based on the first input data 11.

[0047] In the second process PR2, the processing unit 70p derives a second derived feature FD2 from a second inner product 32 of the third feature F13 based on the first input data 11 and the first derived feature FD1.

[0048] The processing unit 70p may be further capable of performing a third process PR3. In the third process PR3, the processing unit 70p outputs a bit-decoded result for the first input data 11 based on at least a part of the second derived feature FD2.

[0049] In the information processing device 70A, feature extraction of the waveform data of the reproduction signal RS is performed based on the calculation of the inner product (first process PR1 and second process PR2). Then, after the feature extraction, classification processing is performed by a fully connected NN. This allows, for example, decoding with high accuracy. For example, decoding with high accuracy at high speed can be performed. It is possible to provide an information processing device capable of improving processing accuracy.

[0050] For example, the first input data 11 includes data on the bit to be decoded and the other bit. In one example, the other bit includes a first other bit and a second other bit. In one example, the bit to be decoded is between the first other bit and the second other bit. In one example, the bit to be decoded may be adjacent to the first other bit. The bit to be decoded may be adjacent to the second other bit. Another bit may be provided between the bit to be decoded and the first other bit. Another bit may be provided between the bit to be decoded and the second other bit. The first other bit may include one or more bits. The second other bit may include one or more bits. In another example, the first other bit may be provided between the bit to be decoded and the second other bit.

[0051] 4, in the information processing device 70A, the processing unit 70p can also perform neural network processing. The processing unit 70p may include an attention layer 71 and a classification layer 72. The attention layer 71 can perform a first process PR1 and a second process PR2. The classification layer 72 can perform a third process PR3.

[0052] For example, the first feature F11 includes a "query" related to the first input data 11. For example, the second feature F12 includes a "key" related to the first input data 11.

[0053] For example, the attention layer 71 includes a first neural network 11N and a second neural network 12N. The attention layer 71 can derive a "query" by processing the first input data 11 with the first neural network 11N. The attention layer 71 can derive a "key" by processing the first input data 11 with the second neural network 12N.

[0054] For example, the third feature F13 includes a "value" related to the first input data 11. For example, the attention layer 71 may further include a third neural network 13N. The attention layer 71 may be capable of deriving the "value" by processing the first input data 11 with the third neural network 13N.

[0055] The attention layer 71 may further include a fourth neural network 14N. The attention layer 71 can derive a second derived feature FD2 by processing the second inner product 32 with the fourth neural network 14N.

[0056] The information processing device 70A may perform the process described with reference to Fig. 3 or 4. In this case, the information processing device 70A uses one type of input data (first input data 11) as input for the process.

[0057] In the information processing device 70A, the processing unit 70p may perform an embedding process 71e of the reproduction signal RS (first input data 11). The embedding process may include, for example, embedding of position information. The embedding process may include, for example, a class token used when classifying input data for which attention processing is performed. For example, the processing unit 70p may be capable of performing a multi-head attention process 71m. For example, the attention layer 71 divides the first input data 11 into multiple vectors.

[0058] Third embodiment The third embodiment relates to a magnetic recording and reproducing device 210 (see FIG. 1). The magnetic recording and reproducing device 210 includes an information processing device according to the embodiment (information processing device 70 or information processing device 70A), a magnetic recording medium 80, and a reproducing unit 80R. The reproducing unit 80R is capable of reproducing a signal 80S corresponding to information recorded on the magnetic recording medium 80. In the magnetic recording and reproducing device 210, it is possible to improve processing accuracy.

[0059] (Fourth embodiment) The fourth embodiment relates to a magnetic recording and reproducing system 310 (see FIG. 1). The magnetic recording and reproducing system 310 includes an information processing device (information processing device 70 or information processing device 70A) according to the embodiment. In the magnetic recording and reproducing system 310, it is possible to improve processing accuracy. In the magnetic recording and reproducing system 310, the information processing device 70 may be provided at a location away from the magnetic recording medium 80.

[0060] In an embodiment, a signal 80S obtained from the magnetic recording medium 80 is AD converted and supplied to the processing unit 70p. The processing unit 70p may include a "Channel Estimation", a waveform equalizer, and a maximum likelihood decoder. The result obtained by the processing unit 70p may be supplied to an error correction decoder (e.g., LDPC: low-density parity-check code). The processing result of the error correction decoder may be supplied to the processing unit 70p. For example, an iterative process may be performed.

[0061] The embodiments may include the following technical solutions. (Technical proposal 1) An acquisition unit; A processing section; Equipped with the acquiring unit is capable of acquiring a reproduction signal based on a signal obtained from a magnetic recording medium, the reproduction signal includes first input data, second input data, and third input data; The processing unit is capable of performing a first process and a second process, In the first processing, the processing unit derives a first derived feature from a first inner product of a first feature based on the first input data and a second feature based on the second input data; In the second process, the processing unit derives a second derived feature from a second inner product of a third feature based on the third input data and the first derived feature.

[0062] (Technical proposal 2) The processing unit is further capable of performing a third process, the first input data includes data on bits to be decoded; The information processing device according to Technical Solution 1, wherein in the third process, the processing unit outputs a bit decoding result for the bits to be decoded based on at least a part of the second derived features.

[0063] (Technical proposal 3) The processing unit is capable of performing neural network processing; The processing unit includes an attention layer and a classification layer; The attention layer is capable of performing the first processing and the second processing, The information processing device described in Technical Proposal 2, wherein the classification layer is capable of performing the third processing.

[0064] (Technical proposal 4) the first feature includes a “query” related to the first input data; The information processing device described in Technical Proposal 3, wherein the second feature includes a "key" related to the second input data.

[0065] (Technical proposal 5) the attention layer includes a first neural network and a second neural network; The attention layer is capable of deriving the "query" by processing the first input data with the first neural network; The information processing device described in Technical Proposal 4, wherein the attention layer is capable of deriving the "key" by processing the second input data with the second neural network.

[0066] (Technical proposal 6) The information processing device described in Technical Proposal 4 or 5, wherein the third feature includes a "value" related to the third input data.

[0067] (Technical proposal 7) the attention layer further includes a third neural network; The information processing device described in Technical Proposal 6, wherein the attention layer is capable of deriving the "value" by processing the third input data with the third neural network.

[0068] (Technical proposal 8) The attention layer further includes a fourth neural network; The information processing device described in Technical Proposal 7, wherein the attention layer is capable of deriving the second derived feature by processing the second inner product with the fourth neural network.

[0069] (Technical proposal 9) The processing unit is capable of embedding processing of the first input data, the second input data, and the third input data, The information processing device according to any one of Technical Solutions 3 to 8, wherein the processing unit performs the first processing after the embedding processing.

[0070] (Technical proposal 10) The attention layer divides at least one of the first input data, the second input data, and the third input data into a plurality of vectors; The information processing device described in any one of Technical Proposals 3 to 9, wherein the attention layer is capable of performing processing on at least two of the multiple vectors in parallel.

[0071] (Technical proposal 11) An acquisition unit; A processing section; Equipped with the acquiring unit is capable of acquiring a reproduction signal based on a signal obtained from a magnetic recording medium, the reproduction signal includes first input data; The processing unit is capable of performing a first process and a second process, In the first processing, the processing unit derives a first derived feature from a first inner product of a first feature based on the first input data and a second feature based on the first input data; In the second processing, the processing unit derives a second derived feature from a second inner product of a third feature based on the first input data and the first derived feature.

[0072] (Technical proposal 12) The processing unit is further capable of performing a third process, the first input data includes data on a bit to be decoded and other bits; The information processing device according to Technical Proposal 11, wherein in the third process, the processing unit outputs a bit decoding result for the bits to be decoded based on at least a portion of the second derived features.

[0073] (Technical proposal 13) The processing unit is capable of performing neural network processing; The processing unit includes an attention layer and a classification layer; The attention layer is capable of performing the first processing and the second processing, The information processing device described in Technical Proposal 12, wherein the classification layer is capable of performing the third processing.

[0074] (Technical proposal 14) the first feature includes a “query” related to the first input data; The information processing device described in Technical Proposal 13, wherein the second feature includes a "key" related to the first input data.

[0075] (Technical proposal 15) the attention layer includes a first neural network and a second neural network; The attention layer is capable of deriving the "query" by processing the first input data with the first neural network; The information processing device described in Technical Proposal 14, wherein the attention layer is capable of deriving the "key" by processing the first input data with the second neural network.

[0076] (Technical proposal 16) The information processing device described in Technical Proposal 14 or 15, wherein the third feature includes a "value" related to the first input data.

[0077] (Technical proposal 17) the attention layer further includes a third neural network; The information processing device described in Technical Proposal 16, wherein the attention layer is capable of deriving the "value" by processing the first input data with the third neural network.

[0078] (Technical proposal 18) The attention layer further includes a fourth neural network; The information processing device described in Technical Proposal 17, wherein the attention layer is capable of deriving the second derived feature by processing the second inner product with the fourth neural network.

[0079] (Technical proposal 19) An information processing device according to any one of technical proposals 1 to 18; The magnetic recording medium; a reproducing unit capable of reproducing the signal corresponding to the information recorded on the magnetic recording medium; A magnetic recording and reproducing device comprising:

[0080] (Technical proposal 20) A magnetic recording and reproducing system including an information processing device according to any one of Technical Solutions 1 to 18.

[0081] According to the embodiments, it is possible to provide an information processing device, a magnetic recording and reproducing device, and a magnetic recording and reproducing system that are capable of improving processing accuracy.

[0082] The above describes the embodiment of the present invention with reference to examples. However, the present invention is not limited to these examples. For example, the specific configurations of each element, such as an acquisition unit, a processing unit, and a magnetic recording medium, included in an information processing device, a magnetic recording and reproducing device, and a magnetic recording and reproducing system, are included in the scope of the present invention as long as a person skilled in the art can implement the present invention in the same way and obtain the same effect by appropriately selecting from the known range.

[0083] Any combination of two or more elements of each example, within the scope of technical feasibility, is also included within the scope of the present invention as long as it includes the gist of the present invention.

[0084] All information processing devices, magnetic recording and reproducing devices, and magnetic recording and reproducing systems that can be implemented by a person skilled in the art by making appropriate design modifications based on the information processing devices, magnetic recording and reproducing devices, and magnetic recording and reproducing systems described above as embodiments of the present invention also fall within the scope of the present invention, so long as they include the gist of the present invention.

[0085] Within the scope of the concept of the present invention, a person skilled in the art may conceive of various modifications and alterations, and it is understood that these modifications and alterations also fall within the scope of the present invention.

[0086] Although some embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope and spirit of the invention, and are included in the scope of the invention and its equivalents described in the claims. [Explanation of symbols]

[0087] 11-13: first to third input data, 11N-14N: first to fourth neural networks, 31, 32: first and second inner products, 70, 70A: information processing device, 70i: acquisition unit, 70p: processing unit, 70pi: input unit, 70po: output unit, 71: attention layer, 71a, 71b: first and second normalization layers, 71e: embedding processing, 71m: multi-head attention processing, 71n: neural network, 71p, 71q: first and second skip connection units, 71u, 71v: first and second value groups, 72: classification layer, 80: magnetic recording medium, 80D: recording unit, 80H: magnetic head, 80R: reproduction unit, 80S: signal, 210: magnetic recording and reproduction device, 310: magnetic recording and reproduction system, BD1: bit decoded result; F11 to F13: first to third feature quantities; FD1, FD2: first and second derived feature quantities; PR1 to PR3: first to third processes; RS: playback signal

Claims

1. An acquisition unit; A processing section; Equipped with the acquiring unit is capable of acquiring a reproduction signal based on a signal obtained from a magnetic recording medium, the reproduction signal includes first input data, second input data, and third input data; The processing unit is capable of performing a first process and a second process, In the first process, the processing unit derives a first derived feature from a first inner product of a first feature based on the first input data and a second feature based on the second input data; In the second process, the processing unit derives a second derived feature from a second inner product of a third feature based on the third input data and the first derived feature.

2. The processing unit is further capable of performing a third process, the first input data includes data on bits to be decoded; The information processing device according to claim 1 , wherein in the third process, the processing unit outputs a bit decoding result for the bits to be decoded based on at least a part of the second derived feature.

3. The processing unit is capable of performing neural network processing; The processing unit includes an attention layer and a classification layer; The attention layer is capable of performing the first process and the second process, The information processing device according to claim 2 , wherein the classification layer is capable of performing the third process.

4. The processing unit is capable of embedding processing of the first input data, the second input data, and the third input data, The information processing apparatus according to claim 3 , wherein the processing unit performs the first processing after the embedding processing.

5. The attention layer divides at least one of the first input data, the second input data, and the third input data into a plurality of vectors; The information processing device according to claim 3 , wherein the attention layer is capable of performing processing on at least two of the plurality of vectors in parallel.

6. An acquisition unit; A processing section; Equipped with the acquiring unit is capable of acquiring a reproduction signal based on a signal obtained from a magnetic recording medium, the reproduction signal includes first input data; The processing unit is capable of performing a first process and a second process, In the first process, the processing unit derives a first derived feature amount from a first inner product of a first feature amount based on the first input data and a second feature amount based on the first input data; In the second process, the processing unit derives a second derived feature from a second inner product of a third feature based on the first input data and the first derived feature.

7. The processing unit is further capable of performing a third process, the first input data includes data on a bit to be decoded and other bits; The information processing device according to claim 6 , wherein in the third process, the processing unit outputs a bit decoding result for the bits to be decoded based on at least a part of the second derived feature.

8. The processing unit is capable of performing neural network processing; The processing unit includes an attention layer and a classification layer; The attention layer is capable of performing the first process and the second process, The information processing device according to claim 7 , wherein the classification layer is capable of performing the third process.

9. An information processing device according to any one of claims 1 to 8; The magnetic recording medium; a reproducing unit capable of reproducing the signal corresponding to the information recorded on the magnetic recording medium; A magnetic recording and reproducing device comprising:

10. 9. A magnetic recording and reproducing system comprising the information processing device according to claim 1.

Images