DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
The following is a Final Office action in response to communications received on 06/22/2026.
Response to Amendment
Claims 2, 3, 9, and 18 have been cancelled.
Claims 21-24 have been newly added.
Claims 1, 8, 12, 13, and 17 have been amended.
Claims 1, 4-8, 10-17, and 19-24 have been examined.
The rejections of claims 2, 9, and 18 are withdrawn in light of the cancellation of the claims.
The objections to claim 1 are withdrawn in light of the amendments made to the claim.
Applicant’s arguments with respect to claims 1, 8, and 17 regarding the new limitations: “wherein the obfuscation key comprises a plurality of key values, and wherein the derivative of the machine learning model was generated by a process comprising: obtaining a plurality of obfuscation values respectively from the plurality of key values, and respectively modifying at least some weights of the machine learning model based on corresponding instances of the obfuscation values” have been considered but are moot in view of the new ground of rejection presented in the current office action. As per the applicant’s argument that prior arts of record fail to teach the new limitations, the examiner respectfully disagrees. Prior art of record Zhang teaches: Zhang: [0051]: S101. The user equipment establishes a connection with the server through the interface, so as to complete the registration of the user equipment on the server. [0054]: S102. The server generates key information according to the device information of the user equipment and the set key generation method. [0057]: S103. Select one or more data from the key information (obfuscation values selected from the plurality of key values), use an encryption algorithm to encrypt the DNN model of the deep neural network to be encrypted, and generate an encrypted DNN model and a key. [0058]: Specifically, the parameters and hyperparameters of the Faster RCNN model are encrypted using the AES algorithm; [0059]-[0060]: Wherein, the parameters and hyperparameters include at least one of the following: Connection weights, bias values, control signals or instructions corresponding to the neural network structure, the number of layers or the number of neurons in each layer, and any static parameters of the neural network model. [0062]: S104. Send the encrypted DNN model and the key to the user equipment, i.e., the weights of the neural network model is encrypted using one or more data (obfuscation values) selected from the key.
Double Patenting
The nonstatutory double patenting rejection is based on a judicially created doctrine grounded in public policy (a policy reflected in the statute) so as to prevent the unjustified or improper timewise extension of the “right to exclude” granted by a patent and to prevent possible harassment by multiple assignees. A nonstatutory double patenting rejection is appropriate where the conflicting claims are not identical, but at least one examined application claim is not patentably distinct from the reference claim(s) because the examined application claim is either anticipated by, or would have been obvious over, the reference claim(s). See, e.g., In re Berg, 140 F.3d 1428, 46 USPQ2d 1226 (Fed. Cir. 1998); In re Goodman, 11 F.3d 1046, 29 USPQ2d 2010 (Fed. Cir. 1993); In re Longi, 759 F.2d 887, 225 USPQ 645 (Fed. Cir. 1985); In re Van Ornum, 686 F.2d 937, 214 USPQ 761 (CCPA 1982); In re Vogel, 422 F.2d 438, 164 USPQ 619 (CCPA 1970); In re Thorington, 418 F.2d 528, 163 USPQ 644 (CCPA 1969).
A timely filed terminal disclaimer in compliance with 37 CFR 1.321(c) or 1.321(d) may be used to overcome an actual or provisional rejection based on nonstatutory double patenting provided the reference application or patent either is shown to be commonly owned with the examined application, or claims an invention made as a result of activities undertaken within the scope of a joint research agreement. See MPEP § 717.02 for applications subject to examination under the first inventor to file provisions of the AIA as explained in MPEP § 2159. See MPEP § 2146 et seq. for applications not subject to examination under the first inventor to file provisions of the AIA . A terminal disclaimer must be signed in compliance with 37 CFR 1.321(b).
The filing of a terminal disclaimer by itself is not a complete reply to a nonstatutory double patenting (NSDP) rejection. A complete reply requires that the terminal disclaimer be accompanied by a reply requesting reconsideration of the prior Office action. Even where the NSDP rejection is provisional the reply must be complete. See MPEP § 804, subsection I.B.1. For a reply to a non-final Office action, see 37 CFR 1.111(a). For a reply to final Office action, see 37 CFR 1.113(c). A request for reconsideration while not provided for in 37 CFR 1.113(c) may be filed after final for consideration. See MPEP §§ 706.07(e) and 714.13.
The USPTO Internet website contains terminal disclaimer forms which may be used. Please visit www.uspto.gov/patent/patents-forms. The actual filing date of the application in which the form is filed determines what form (e.g., PTO/SB/25, PTO/SB/26, PTO/AIA /25, or PTO/AIA /26) should be used. A web-based eTerminal Disclaimer may be filled out completely online using web-screens. An eTerminal Disclaimer that meets all requirements is auto-processed and approved immediately upon submission. For more information about eTerminal Disclaimers, refer to www.uspto.gov/patents/apply/applying-online/eterminal-disclaimer.
Claims 1, 4-8, 10-17, and 19-24 are rejected on the ground of nonstatutory double patenting as being unpatentable over claims 1-20 of U.S. Patent No. 12204663. Although the claims at issue are not identical, they are not patentably distinct from each other because:
Instant application
U.S. Patent No. 12204663
1. A non-transitory computer-readable medium, storing program instructions that, when executed by one or more processors of a computing system, cause the computing system to perform operations comprising: receiving, from a user device, a request related to a machine learning model; determining whether that a derivative of the machine learning model was not provided to the user device; and in response to determining that the derivative of the machine learning model was not provided to the user device, generating the derivative of the machine learning model with one or more weights of the machine learning model obfuscated by way of an obfuscation key, and providing, to the user device, the derivative of the machine learning model and the obfuscation key, wherein the obfuscation key comprises a plurality of key values, and wherein generating the derivative of the machine learning model comprises: obtaining a plurality of obfuscation values respectively from the plurality of key values, and respectively modifying at least some of the weights based on corresponding instances of the obfuscation values.
8. (Currently amended) A method comprising:
requesting, from a computing system, a machine learning model;
receiving, from the computing system, a derivative of the machine learning model and an obfuscation key,
wherein the obfuscation key comprises a plurality of key values, and wherein the derivative of the machine learning model was generated by a process comprising: obtaining a plurality of obfuscation values respectively from the plurality of key values, and respectively modifying at least some weights of the machine learning model based on corresponding instances of the obfuscation values;
storing, in memory, the derivative of the machine learning model and the obfuscation key; applying the obfuscation key to the derivative of the machine learning model during runtime of an application; in response to the application closing, erasing, from the memory, the obfuscation key; and in response to the application reopening, requesting and receiving, from the computing system, a new copy of the obfuscation key.
17. (Currently amended) A non-transitory computer-readable medium, storing program instructions that, when executed by one or more processors of a user device, cause the user device to perform operations comprising: requesting, from a computing system, a machine learning model; receiving, from the computing system, a derivative of the machine learning model and an obfuscation key,
wherein the obfuscation key comprises a plurality of key values, and wherein the derivative of the machine learning model was generated by a process comprising: obtaining a plurality of obfuscation values respectively from the plurality of key values, and respectively modifying at least some weights of the machine learning model based on corresponding instances of the obfuscation values;
storing, in memory, the derivative of the machine learning model and the obfuscation key; applying the obfuscation key to the derivative of the machine learning model during runtime of an application; in response to the application closing, erasing, from the memory, the obfuscation key; and in response to the application reopening, requesting and receiving, from the computing system, a new copy of the obfuscation key.
18. Non-transitory computer readable storage media storing executable instructions which, when executed by at least one processing device, cause the at least one processing device to:
receive, from an application executing on a device of a user, a request to download a machine learning model to the device for enabling a feature of the application, the request including information associated with the user and the device; create an obfuscation key based on the information included in the request; generate a derivative model using a reference copy of the machine learning model and the obfuscation key by: including one or more obfuscation parameters as inputs to the reference copy of the machine learning model in addition to a dataset input to the reference copy of the machine learning model, each of the one or more obfuscation parameters having an assigned value derived from the obfuscation key; modifying one or more weights of the reference copy of the machine learning model with the assigned value of a respective obfuscation parameter;…; and send the derivative model and the obfuscation key to the application; …
1. A method for controlling access to an on-device machine learning model, the method comprising: receiving, from an application executing on a device of a user, a request to download a machine learning model to the device for enabling a feature of the application, the request including information associated with the user and the device; creating an obfuscation key based on the information included in the request; generating a derivative model using a reference copy of the machine learning model and the obfuscation key by: including one or more obfuscation parameters as inputs to the reference copy of the machine learning model in addition to a dataset input to the reference copy of the machine learning model, each of the one or more obfuscation parameters having an assigned value derived from the obfuscation key; modifying one or more weights of the reference copy of the machine learning model with the assigned value of a respective obfuscation parameter; …;
persistently storing the derivative model in a storage of the device; accessing the derivative model by the application using the obfuscation key; once the application is closed on the device of the user: removing the obfuscation key from the storage of the device by the application; once the application is re-opened on the device of the user; receiving a re-created obfuscation key by the application; and accessing the derivative model persistently stored in the storage of the device by the application using the re-created obfuscation key.
18. Non-transitory computer readable storage media storing executable instructions which, when executed by at least one processing device, cause the at least one processing device to: receive, from an application executing on a device of a user, a request to download a machine learning model to the device for enabling a feature of the application, the request including information associated with the user and the device; create an obfuscation key based on the information included in the request; generate a derivative model using a reference copy of the machine learning model and the obfuscation key by: including one or more obfuscation parameters as inputs to the reference copy of the machine learning model in addition to a dataset input to the reference copy of the machine learning model, each of the one or more obfuscation parameters having an assigned value derived from the obfuscation key; modifying one or more weights of the reference copy of the machine learning model with the assigned value of a respective obfuscation parameter; and for each of the one or more modified weights, …; persistently store the derivative model in a storage of the device; access the derivative model by the application using the obfuscation key; once the application is closed on the device of the user: remove the obfuscation key from the storage of the device by the application; once the application is re-opened on the device of the user: receive a re-created obfuscation key by the application; and access the derivative model persistently stored in the storage of the device by the application using the re-created obfuscation key.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action.
Claims 1, 4-6, and 21 are rejected under 35 U.S.C. 103 as being unpatentable over US 20210058653 to Jang (hereinafter Jang) and prior art of record CN109040091A to Zhang et al (hereinafter Zhang).
Examiner’s Note: The examiner used an English translation of Zhang which was provided in the previous office action.
As per claim 1, Jang teaches:
A non-transitory computer-readable medium, storing program instructions that, when executed by one or more processors of a computing system, cause the computing system to perform operations comprising:
receiving, from a user device, a request related to a machine learning model; determining whether that not provided to the user device (Jang: [0114] Thereafter, the server 100 may perform an operation 720 of confirming that a streaming request (or a download request) for the video data is received from the user device 200. In response to a user request, the server 100 may perform an operation 725 of transmitting a low-quality version of the requested video data to the user device 200 together with the generated neural network file for improving the image quality, i.e., the server determines, based on the user request, that the neural network file has not been provided to the user); and
in response to determining that learning model (Jang: [0114]: In response to a user request, the server 100 may perform an operation 725 of transmitting a low-quality version of the requested video data to the user device 200 together with the generated neural network file for improving the image quality).
Jang does not teach a derivative of the machine learning model and generating the derivative of the machine learning model with one or more weights of the machine learning model obfuscated by way of an obfuscation key, and providing, to the user device, the derivative of the machine learning model and the obfuscation key, wherein the obfuscation key comprises a plurality of key values, and wherein generating the derivative of the machine learning model comprises: obtaining a plurality of obfuscation values respectively from the plurality of key values, and respectively modifying at least some of the weights based on corresponding instances of the obfuscation values. However, Zhang teaches:
a derivative of the machine learning model; generating the derivative of the machine learning model with one or more weights of the machine learning model obfuscated by way of an obfuscation key, and providing, to the user device, the derivative of the machine learning model and the obfuscation key, wherein the obfuscation key comprises a plurality of key values, and wherein generating the derivative of the machine learning model comprises: obtaining a plurality of obfuscation values respectively from the plurality of key values, and respectively modifying at least some of the weights based on corresponding instances of the obfuscation values (Zhang: [0051]: S101. The user equipment establishes a connection with the server through the interface, so as to complete the registration of the user equipment on the server. [0054]: S102. The server generates key information according to the device information of the user equipment and the set key generation method. [0057]: S103. Select one or more data from the key information (obfuscation values selected from the plurality of key values), use an encryption algorithm to encrypt the DNN model of the deep neural network to be encrypted, and generate an encrypted DNN model and a key. [0058]: Specifically, the parameters and hyperparameters of the Faster RCNN model are encrypted using the AES algorithm; [0059]-[0060]: Wherein, the parameters and hyperparameters include at least one of the following: Connection weights, bias values, control signals or instructions corresponding to the neural network structure, the number of layers or the number of neurons in each layer, and any static parameters of the neural network model. [0062]: S104. Send the encrypted DNN model and the key to the user equipment).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to employ the teachings of Zhang in the invention of Jang to include the above limitations. The motivation to do so would be to protect the DNN model (Zhang: [0005]).
As per claim 4, Jang in view of Zhang teaches:
The non-transitory computer-readable medium of claim 1, wherein the request is from an application having a feature executable on the user device, wherein the feature uses the machine learning model (Jang: [0114] Thereafter, the server 100 may perform an operation 720 of confirming that a streaming request (or a download request) for the video data is received from the user device 200. Accordingly, since the user device 200 receives a low-quality version of a video, contents may be received in a relatively easy way without being restricted by a network environment, and a high-quality video at a level desired by the user may be played back through applying the universal neural network file. It was well known to one of ordinary skill in the art before the effective filing date of the claimed invention that a streaming application is used to request video data).
As per claim 5, Jang in view of Zhang teaches:
The non-transitory computer-readable medium of claim 4, wherein the request is associated with a user account of a media service that uses the application (Zhang: [0051]: S101. The user equipment establishes a connection with the server through the interface, so as to complete the registration of the user equipment on the server. It was well known to one of ordinary skill in the art before the effective filing date of the claimed invention that a registration process is used to create a user account. [0052]: the decrypted deep neural network model can be used to detect pedestrians and vehicles. It was well known to one of ordinary skill in the art before the effective filing date of the claimed invention that the pedestrians and vehicles are detected based on the video (media service) taken of the surroundings).
The examiner provides the same rationale to combine prior arts Jang and Zhang as in claim 1 above.
As per claim 6, Jang in view of Zhang teaches:
The non-transitory computer-readable medium of claim 5, wherein obtaining the obfuscation key comprises generating the obfuscation key based on: a user identifier of the user account, or a device identifier of the user device (Zhang: [0054]: S102. The server generates key information according to the device information (device identifier) of the user equipment and the set key generation method).
The examiner provides the same rationale to combine prior arts Jang and Zhang as in claim 1 above.
As per claim 21, Jang in view of Zhang teaches:
The non-transitory computer-readable medium of claim 1, wherein obtaining the plurality of obfuscation values respectively from the plurality of key values comprises applying respective first operations to the key values (Zhang: [0057]: S103. Select one or more data from the key information, use an encryption algorithm to encrypt the DNN model of the deep neural network to be encrypted, and generate an encrypted DNN model and a key), the operations further comprising: storing, associated with the derivative of the machine learning model, respective second operations that are inverses of the respective first operations, the respective second operations configured to be performed during execution of the derivative of the machine learning model (Zhang: [0065]: When the matching is successful, the user equipment passes the verification, and the user equipment decrypts the encrypted DNN model. [0066]: If it is a legitimate device, the encrypted Faster RCNN model can continue to be decrypted and used; [0067]: provide the model to users in an encrypted form and restrict the DNN model to be decrypted and run only on authorized devices).
The examiner provides the same rationale to combine prior arts Jang and Zhang as in claim 1 above.
Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over Jang in view of Zhang as applied to claim 1 above, and further in view of prior art of record US 20200302313 to Jeong (hereinafter Jeong).
As per claim 7, Jang in view of Zhang does not teach the limitations of claim 7. However, Jeong teaches:
wherein applying the obfuscation key to the derivative of the machine learning model facilitates playout of media on the user device, wherein the media is from the media service (Jeong: [0037]-[0039]. [0243]. [0298] For example, when the command word is “Please play the story of Snow White”, the artificial intelligence unit 130 may determine that the function corresponding to the command word is output of the story of Snow White. [0300] For example, when the function corresponding to the command word is the first type of function, the artificial intelligence unit 130 may determine the tablet PC for performing the first type of function among the plurality of devices as the device for performing the function corresponding to the command word. [0301] In this case, the controller 180 may transmit the command for performing the function corresponding to the command word to the determined device, under control of the artificial intelligence unit 130. [0302] For example, the controller 180 may transmit the command for outputting the story of Snow White to the tablet PC).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to employ the teachings of Jeong in the invention of Jang in view of Zhang to include the above limitations. The motivation to do so would be to provide direct feedback to a user who wants to receive a function, by distinguishing a speaker who has uttered a command word and outputting a function corresponding to a command word to a personal device of the speaker who has uttered the command word (Jeong: [0011]).
Claims 8, 16, 17, 19, 22, and 23 are rejected under 35 U.S.C. 103 as being unpatentable over Zhang and prior art of record US 2014033129 to Ramallo et al (hereinafter Ramallo).
As per claims 8 and 17, Zhang teaches:
A method comprising:
requesting, from a computing system, a machine learning model; receiving, from the computing system, a derivative of the machine learning model and an obfuscation key (Zhang: [0051]: S101. The user equipment establishes a connection with the server through an interface, so as to complete the registration of the user equipment on the server. [0054]: S102. The server generates key information according to the device information of the user equipment and the set key generation method. [0062]: S104. Send the encrypted DNN model and the key to the user equipment),
wherein the obfuscation key comprises a plurality of key values, and wherein the derivative of the machine learning model was generated by a process comprising: obtaining a plurality of obfuscation values respectively from the plurality of key values, and respectively modifying at least some weights of the machine learning model based on corresponding instances of the obfuscation values (Zhang: [0054]: S102. The server generates key information according to the device information of the user equipment and the set key generation method. [0057]: S103. Select one or more data from the key information (obfuscation values selected from the plurality of key values), use an encryption algorithm to encrypt the DNN model of the deep neural network to be encrypted, and generate an encrypted DNN model and a key. [0058]: Specifically, the parameters and hyperparameters of the Faster RCNN model are encrypted using the AES algorithm; [0059]-[0060]: Wherein, the parameters and hyperparameters include at least one of the following: Connection weights, bias values, control signals or instructions corresponding to the neural network structure, the number of layers or the number of neurons in each layer, and any static parameters of the neural network model);
storing, in memory, the derivative of the machine learning model and the obfuscation key (Zhang: [0062]: S104. Send the encrypted DNN model and the key to the user equipment. Storing the received encrypted DNN model and key was well known to one of ordinary skill in the art);
applying the obfuscation key to the derivative of the machine learning model during runtime of an application (Zhang: [0073]: the user equipment decrypts the encrypted DNN model. [0052]: computing device, the decrypted deep neural network model can be used to detect pedestrians and vehicles. It was well known to one of ordinary skill in the art before the effective filing date of the claimed invention that detecting pedestrians and vehicles is performed during the runtime of an application);
Zhang does not teach: in response to the application closing, erasing, from the memory, the obfuscation key; and in response to the application reopening, requesting and receiving, from the computing system, a new copy of the obfuscation key. However, Ramallo teaches:
in response to the application closing, erasing, from the memory, the obfuscation key (Ramallo: [0006] Thus, in accordance with the invention, as soon as the computing device ceases to be in an operational mode, the access key is removed from the storage location. [0007] The operational mode may be normal operation of the computing device or may be operation of a particular part of the computing device, such as an application. [0008] The access key may be an encryption key. [0010] Removing the access key from the storage location may comprise deleting the access key from the storage location); and
in response to the application reopening, requesting and receiving, from the computing system, a new copy of the obfuscation key (Ramallo: [0037]: a new key will be requested from the server when the device is to be used. [0016]: The key is restored in a secure way by identifying the user and/or checking other required conditions and/or credentials (hardware or software) before initiating the restore procedure. [0025] The method may use the communication technique described in WO 2010/039041 to restore keys in a secure way. In this case, the client device first sends a request to the server using any method of communication identifying that the client requests a replacement key. This may also be done from an alternative device, such as a mobile phone, by calling, SMS, e-mail, or from any other device. This will in turn trigger an event in the server using the techniques of WO 2010/039041 to send a request to the device that is rejected, and the device will search for the identifier, and if found connect back using any mean of communications to the server).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to employ the teachings of Ramallo in the invention of Zhang to include the above limitations. The motivation to do so would be to secure a computing device (Ramallo: [0004]).
As per claim 16, Zhang in view of Ramallo teaches:
The method of claim 8, further comprising: applying the new copy of the obfuscation key to the derivative of the machine learning model during runtime of the application (Ramallo: [0037]: a new key will be requested from the server when the device is to be used. Zhang: [0073]: the user equipment decrypts the encrypted DNN model. [0052]: computing device, the decrypted deep neural network model can be used to detect pedestrians and vehicles. It was well known to one of ordinary skill in the art before the effective filing date of the claimed invention that detecting pedestrians and vehicles is performed during the runtime of an application).
The examiner provides the same rationale to combine prior arts Zhang and Ramallo as in claim 8 above.
As per claim 19, Zhang in view of Ramallo teaches:
The non-transitory computer-readable medium of claim 17, wherein requesting the machine learning model comprises transmitting, to the computing system, a request from the application, wherein the request is associated with a user account of a media service that uses the application (Zhang: [0051]: S101. The user equipment establishes a connection with the server through the interface, so as to complete the registration of the user equipment on the server. It was well known to one of ordinary skill in the art before the effective filing date of the claimed invention that a registration process is used to create a user account. [0052]: the decrypted deep neural network model can be used to detect pedestrians and vehicles. It was well known to one of ordinary skill in the art before the effective filing date of the claimed invention that the pedestrians and vehicles are detected based on the video (media service) taken of the surroundings), and wherein the obfuscation key was generated based on: a user identifier of the user account, or a device identifier of the user device (Zhang: [0054]: S102. The server generates key information according to the device information (device identifier) of the user equipment and the set key generation method).
As per claims 22 and 23, Zhang in view of Ramallo teaches:
The method of claim 8, wherein obtaining the plurality of obfuscation values respectively from the plurality of key values comprises applying respective first operations to the key values (Zhang: [0057]: S103. Select one or more data from the key information, use an encryption algorithm to encrypt the DNN model of the deep neural network to be encrypted, and generate an encrypted DNN model and a key), the method further comprising: storing, associated with the derivative of the machine learning model, respective second operations that are inverses of the respective first operations, the respective second operations configured to be performed during execution of the derivative of the machine learning model (Zhang: [0065]: When the matching is successful, the user equipment passes the verification, and the user equipment decrypts the encrypted DNN model. [0066]: If it is a legitimate device, the encrypted Faster RCNN model can continue to be decrypted and used; [0067]: provide the model to users in an encrypted form and restrict the DNN model to be decrypted and run only on authorized devices).
Claims 10 are rejected under 35 U.S.C. 103 as being unpatentable over Zhang in view of Ramallo as applied to claim 8 above, and further in view of prior art of record US 20210117805 to Kyakuno (hereinafter Kyakuno).
As per claim 10, Zhang in view of Ramallo does not explicitly teach the limitations of claim 10. However, Kyakuno teaches:
wherein requesting the machine learning model comprises transmitting, to the computing system, a request from the application having a feature that uses the machine learning model, and wherein the application is configured to execute on a user device that provided the request (Kyakuno: [0084] When an application is executed by a user, the determination unit 12 determines whether a learned model has been encrypted by referring to an encryption identifier attached to the learned model acquired by the acquisition unit 11. [0088] The inference unit 14 performs inference processing by using the decrypted learned model. The inference unit 14 outputs an inference result to the application).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to employ the teachings of Kyakuno in the invention of Zhang in view of Ramallo to include the above limitations. The motivation to do so would be to promote collaboration between the developer of the learned model and the application developer, while reducing the risk such that the learned model is misused without permission (Kyakuno: [0158]).
As per claim 11, Zhang in view of Ramallo and Kyakuno teaches:
The method of claim 10, wherein the request is associated with a user account of a media service that uses the application (Zhang: [0051]: S101. The user equipment establishes a connection with the server through the interface, so as to complete the registration of the user equipment on the server. It was well known to one of ordinary skill in the art before the effective filing date of the claimed invention that a registration process is used to create a user account. [0052]: the decrypted deep neural network model can be used to detect pedestrians and vehicles. It was well known to one of ordinary skill in the art before the effective filing date of the claimed invention that the pedestrians and vehicles are detected based on the video (media service) taken of the surroundings).
As per claim 12, Zhang in view of Ramallo and Kyakuno teaches:
The method of claim 11, wherein the obfuscation key was generated based on: a user identifier of the user account, or a device identifier of the user device (Zhang: [0054]: S102. The server generates key information according to the device information (device identifier) of the user equipment and the set key generation method).
Claim 13 is rejected under 35 U.S.C. 103 as being unpatentable over Zhang in view of Ramallo and Kyakuno as applied to claim 11 above, and further in view of Jeong.
As per claim 13, Zhang in view of Ramallo and Kyakuno does not teach the limitations of claim 13. However, Jeong teaches:
wherein applying the obfuscation key to the derivative of the machine learning model facilitates playout of media on the user device, wherein the media is from the media service (Jeong: [0037]-[0039]. [0243]. [0298] For example, when the command word is “Please play the story of Snow White”, the artificial intelligence unit 130 may determine that the function corresponding to the command word is output of the story of Snow White. [0300] For example, when the function corresponding to the command word is the first type of function, the artificial intelligence unit 130 may determine the tablet PC for performing the first type of function among the plurality of devices as the device for performing the function corresponding to the command word. [0301] In this case, the controller 180 may transmit the command for performing the function corresponding to the command word to the determined device, under control of the artificial intelligence unit 130. [0302] For example, the controller 180 may transmit the command for outputting the story of Snow White to the tablet PC).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to employ the teachings of Jeong in the invention of Zhang in view of Ramallo and Kyakuno to include the above limitations. The motivation to do so would be to provide direct feedback to a user who wants to receive a function, by distinguishing a speaker who has uttered a command word and outputting a function corresponding to a command word to a personal device of the speaker who has uttered the command word (Jeong: [0011]).
Claim 14 is rejected under 35 U.S.C. 103 as being unpatentable over Zhang in view of Ramallo as applied to claim 8 above, and further in view of prior art of record CN112698848A to Zhang et al (hereinafter Zhang’848).
Examiner’s Note: The examiner used an English translation of Zhang’848 which was provided in the previous office action.
As per claim 14, Zhang in view of Ramallo does not teach the limitations of claim 14. However, Zhang’848 teaches:
wherein requesting the machine learning model is in response to the application opening (Zhang’848: [n0068]: In a possible manner, the application programs or system services shown in this application continue to run when the terminal is turned on. In this scenario, when a continuously running application or system service detects that the application environment information has changed, the application or system service will execute the method for downloading the machine learning model shown in this application).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to employ the teachings of Zhang’848 in the invention of Zhang in view of Ramallo to include the above limitations. The claim would have been obvious because a particular known technique was recognized as part of the ordinary capabilities of one skilled in the art (see KSR Int’l Co. v. Teleflex Inc. 550 U.S. ___, 82 USPQ2d 1385 (Supreme Court 2007) (KSR)).
Claims 15, 20, and 24 are rejected under 35 U.S.C. 103 as being unpatentable over Zhang in view of Ramallo as applied to claims 8 and 17 above, and further in view of Jeong.
As per claims 15 and 24, Zhang in view of Ramallo teaches:
The method of claim 8, wherein the derivative of the machine learning model is configured such that it receives as output when the obfuscation key is not available in the memory (Zhang: [0066]: The user equipment is verified according to the incentive corresponding signal X sent by the server, and returns a corresponding response signal to the server. If the returned response signal is Y, it means that the user equipment is a device that has been registered on the server side. If it is a legitimate device, the encrypted Faster RCNN model can continue to be decrypted and used; if the returned signal is not Y, that is, the signal returned by the user equipment is different from the stimulus response pair stored in the server, it means that the user equipment is A device that has not been registered on the server side, the user device is an illegal device, and the user device cannot decrypt and use the encrypted Faster RCNN model, i.e., an illegal device cannot decrypt and use the machine learning model so that there is no output from the machine learning model).
Zhang in view of Ramallo does not teach: such that it receives audio as input and: produces the audio as output. However, Jeong teaches:
such that it receives audio as input and: produces the audio as output (Jeong: [0037]-[0039]. [0247] For example, when the acquired command word is “Please play music” (audio input), the artificial intelligence unit 130 of the artificial intelligence device 600 may transmit a control command for outputting music to the second personal device 800. The controller of the second personal device 800, which has received the command, may output music using the content (music) stored in the storage of the second personal device 800).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to employ the teachings of Jeong in the invention of Zhang in view of Ramallo to include the above limitations. The motivation to do so would be to provide direct feedback to a user who wants to receive a function, by distinguishing a speaker who has uttered a command word and outputting a function corresponding to a command word to a personal device of the speaker who has uttered the command word (Jeong: [0011]).
As per claim 20, Zhang in view of Ramallo does not teach the limitations of claim 20. However, Jeong teaches:
wherein applying the obfuscation key to the derivative of the machine learning model facilitates playout of media on the user device, wherein the media is from a media service that uses the application (Jeong: [0037]-[0039]. [0243]. [0298] For example, when the command word is “Please play the story of Snow White”, the artificial intelligence unit 130 may determine that the function corresponding to the command word is output of the story of Snow White. [0300] For example, when the function corresponding to the command word is the first type of function, the artificial intelligence unit 130 may determine the tablet PC for performing the first type of function among the plurality of devices as the device for performing the function corresponding to the command word. [0301] In this case, the controller 180 may transmit the command for performing the function corresponding to the command word to the determined device, under control of the artificial intelligence unit 130. [0302] For example, the controller 180 may transmit the command for outputting the story of Snow White to the tablet PC).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to employ the teachings of Jeong in the invention of Zhang in view of Ramallo to include the above limitations. The motivation to do so would be to provide direct feedback to a user who wants to receive a function, by distinguishing a speaker who has uttered a command word and outputting a function corresponding to a command word to a personal device of the speaker who has uttered the command word (Jeong: [0011]).
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MADHURI R HERZOG whose telephone number is (571)270-3359. The examiner can normally be reached 8:30AM-4:30PM.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Taghi Arani can be reached at (571)272-3787. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
MADHURI R. HERZOG
Primary Examiner
Art Unit 2438
/MADHURI R HERZOG/Primary Examiner, Art Unit 2438
Prosecution Insights