DETAILED ACTION
This action is responsive to the Application filed on 04/24/2025. Claims 1-8 are pending in the case. Claim 1 is an independent claim. Claims 1 and 6 are amended.
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 .
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 04/24/2025 has been entered.
Response to Arguments
Applicant's arguments filed 04/24/2025 have been fully considered but they are not persuasive.
Applicant’s arguments:
Applicant appears to highlight that the cited are doesn’t teach “the inputting by user of the characters expressing the intention or emotion is to assist in training the AI model” as reflected in the claims. Applicant notes that Lee only relates to analysis by an algorithm.
Further, Applicant argues the references are silent concerning “generating sounds includes receiving a user selection from an intention/emotion list…”
Examiner response:
Examiner notes that Lee, does not appear to use the word “training” in the disclosure, however Lee is understood to describe a deep learning system which learns based on communication data between and object and a user. For example, Lee describes “The heterologous bidirectional communication service providing server 300 may be a server that receives a recognized situation between a communication object and a user, learns the received data, and transmits the received data to the user terminal 400…”. Nevertheless, the rejected has been updated in view of Molnar, which explicitly describes training a deep network by combining sound data with characters expressing intention or emotion input by a user.
Examiner notes Lee describes inputting user text, which is understood to be a selection from an intention list, by the user. Text in a human language is a description of intentions that is selected from the list of all possible language strings which as described by Lee is used to generate a sound “On the other hand, when the text from the user terminal 400 to the heterologous animal communication terminal 100 is collected, the heterologous animal bidirectional communication service providing server 300 transmits the collected text through the inverse translation process of the previously stored animal sound translation algorithm It can be created with animal sounds. Then, the heterologous animal bidirectional communication service providing server 300 can transmit the generated animal sounds”
See the updated rejection for details.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
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 factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-4, and 6-7 are rejected under 35 U.S.C. 103 as being unpatentable over Lee WO2018135819A1 as applied to claim 1 above, and further in view of Chang et al “learning representations of emotional speech with deep convolutional generative adversarial networks”, further in view of Molnar et al “Classification of dog barks: a machine learning approach”
Regarding Claim 1 Lee discloses A method for providing a companion animal sound service using an artificial intelligence model based on deep neural network machine learning, the method comprising the steps of (pg 7 para 003 “Accordingly, the user terminal 400 patterns the stored data into machine learning, deep learning and artificial intelligence neural network learning, and when the patterned output data is input from the heterogeneous animal communication terminal 100, The user terminal 400 can output the communication content matched with the output data. At this time, the patterning of the input data and the output data may be executed in the heterogeneous animal bidirectional communication service providing server 300”)
requesting, by a management server, recording and uploading of sounds for each intention or emotion of a user's companion animal to a companion animal application executed in a user terminal uploading, by the companion animal application, sound data of the requested and recorded sounds for each intention or emotion of the user's companion animal to the management server (pg 10 para 003 “In order to pattern the sounds of the dogs, five kinds of signals of different kinds of dogs were recorded. At this time, the situation can include anxiety, loneliness, pleasure, boundary and anger. Then, to analyze the recorded audio, the pitch was measured using a Logix Pro X program” pg 11 last paragraph “Referring to FIG. 17, when the sounds of dogs are input through the heterologous animal communication terminal 100 (S7100), the user terminal 400 or the heterologous animal bidirectional communication service providing server 300 transmits … the subject for processing the signal may be changed to the heterologous animal communication terminal 100, the user terminal 400, or the heterologous animal bidirectional communication service providing server 300 according to various embodiments”
training, by the management server, an artificial intelligence model based on deep neural network machine learning with the uploaded sound data of the sounds for each intention or emotion of the user’s companion animal( para 0030 Lee “The heterologous bidirectional communication service providing server 300 may be a server that receives a recognized situation between a communication object and a user, learns [trains] the received data, and transmits the received data to the user terminal 400”)
providing, by the management server, the trained artificial intelligence model to the companion animal application ( pg 7 para 003 “Accordingly, the user terminal 400 patterns the stored data into machine learning, deep learning and artificial intelligence neural network learning”
generating, by the companion animal application, sounds for the companion animal corresponding to a user input by using the artificial intelligence model, and outputting the sounds via a speaker (pg 4 para 002 “The heterogeneous animal communication terminal 100 may be a terminal that converts a message of the user terminal 400 into a sound of a communication object and outputs the message” pg 4 para 003 “In addition, the heterologous animal communication terminal 100 may be configured to generate a sound signal representing a cough sound expressing a disease, a righteous voice of an animal, a mutual communication sound between animals, and a sick feeling of an animal among cries of a plurality of animals” pg 6 para 009 “The output unit 340 outputs the collected human content to the user terminal 400 and the disparate animal communication terminal 100 when the human content including at least one of voice, image, and moving picture is collected from the user terminal 400 Output to the content input / output device… outputting a content through a beam project, a display, a speaker, a microphone, or the like)
The artificial intelligence model having been trained to generate sounds which are similar to the sounds for each intention or emotion of the user’s companion animal or to transform and generate sounds according to a criterion specified by a user ( pg 8 para 004 “Also, the pre-stored animal sound translation algorithm can be based on an animal translation database which interprets animal sounds including sound cycle, frequency, pitch, vibration, pattern, and interval by animal type and anatomical structure have. Here, the animal translation database may be a database that is updated based on the sound of the animal wearing the heterologous animal communication terminal 100 and the feedback data of the user terminal 400 and adapted to be customized using a gene learning algorithm, The algorithm can be an algorithm that bi-directionally translates by storing the mutual mapping of animal sounds and people, and is modified to be customized by a genetic learning algorithm” the algorithm is trained to transform and generate sounds based on feedback data specified by a user terminal, or customized using a specified gene learning algorithm, i.e criterion specified by a user.)
Preparing, by the management server, an inference system including the artificial intelligence learning based on deep neural network machine learning,… or generate new data having similar characteristics by training information about sounds for each intention or emotion of the user’s companion animal ( [Lee, 65] “Also, the pre-stored animal sound translation algorithm can be based on an animal translation database which interprets animal sounds including sound cycle,… the algorithm can be an algorithm that bi-directionally translates by storing the mutual mapping of animal sounds and people, and is modified to be customized by a genetic learning algorithm [training information]” the algorithm generates sounds based on information about sounds such as sound cycle,)
wherein the generating the sounds includes receiving a user selection from an intention/emotion list of intention/emotion items that can be expressed to the companion animal, presented by the companion animal application, and reproducing a sound corresponding to the user selection (pg 8 para 003 “On the other hand, when the text from the user terminal 400 to the heterologous animal communication terminal 100 is collected, the heterologous animal bidirectional communication service providing server 300 transmits the collected text through the inverse translation process of the previously stored animal sound translation algorithm It can be created with animal sounds. Then, the heterologous animal bidirectional communication service providing server 300 can transmit the generated animal sounds to the heterologous animal communication terminal 100. Accordingly, even if the user is a person and a communication object is personal, not only the voice of the user is analyzed and transmitted to the user, but also the user's text and the like are converted into four languages so that the intention and feelings of the opening user can be known”)
based on deep neural network machine learning (pg 4 para 005 “The heterogeneous animal bidirectional communication service providing server 300 analyzes and patterns the sound by type, object, etc., and performs machine learning and deep running using the big data.”)
Lee may not explicitly disclose combining and classifying, by the management server, the uploaded sound data of the sounds for each intention or emotion of the user's companion animal together with characters, input by the user, expressing the intention or emotion and sounds of the companion animal corresponding to the characters to refine and process the sound data as training data for training the artificial intelligence model …the artificial intelligence model including generative adversarial networks (GANs) which generate new data through a pair of generator and a classifier
However, Chang discloses the artificial intelligence model including generative adversarial networks (GANs) which generate new data through a pair of generator and a classifier (pg 2 para 002 “The unsupervised learning part of the investigated model builds upon an architecture known as the deep convolutional generative adversarial network, or DCGAN. DCGAN [generative adversarial network] consists of two components, known as the generator and discriminator,” Section 3 para 002-003 pg 2 “In particular, the model is trained by iteratively running the generator, discriminator, valence classifier, and activation classifier, and back-propagating the error for each component through the network… and discriminating between real and generated samples..” The GAN system includes a generator wo generate new data via a pair of a generator and classifier.)
It would be obvious to one with ordinary skill in the art to combine Lee and Chang address a system for generated new data with emotional expression , Chang notes “The subtle changes in the voice of the speaker that are indicative of positive or negative emotional states are often ”overshadowed” by voice characteristics relating to emotional intensity or emotional activation… to improve classifier performance: (1) utilization of unlabeled data using a deep convolutional generative adversarial network … experiments show that in particular the use of unlabeled data in our investigations improves performance of the classifiers and both fully supervised baseline approaches are outperformed considerably”, suggesting that a GAN is better able to capture positive and negative emotional states. (abstract Chang)
Lee/Chang does not explicitly teach, combining and classifying, by the management server, the uploaded sound data of the sounds for each intention or emotion of the user's companion animal together with characters, input by the user, expressing the intention or emotion and sounds of the companion animal corresponding to the characters to refine and process the sound data as training data for training the artificial intelligence model
Molnar however teaches, combining and classifying, by the management server, the uploaded sound data of the sounds for each intention or emotion of the user's companion animal together with characters, input by the user, expressing the intention or emotion and sounds of the companion animal corresponding to the characters to refine and process the sound data as training data for training the artificial intelligence model based on deep neural network machine learning (pg 2 “The machine is Wrst trained on a set of labeled examples and then tested on a second set for which it must predict the labels. During the training phase, parameters of the models are tuned automatically by the learning algorithm in order to obtain the best classiWcation performances on the training set …In this study, we analyzed more than 6,000 barks recorded in diVerent situations from several individuals using a computerized method… The most relevant descriptors are then fed into machine-learning algorithms to produce classiWers (or extractors)… We collected bark recordings in seven diVerent behavioral contexts, most of which could be arranged at the homes of the owners… “Stranger”… “Fight”...” pg 3 “As each recording possibly contained up to three or four barks, individual bark sounds were manually segmented and extracted.” the training data combines the animals sounds or barks with labeled input by users expressing the context or intent, which is refined and processed during training.)
It would be obvious to one with ordinary skill in the art to combine Lee/Change and Molnar address in order to perform steps related to training machine learning models on labeled sound data. Specifically, Molnar noted that “the efficiency of the algorithm was tested in a classification task in which unknown barks were analyzed. The recognition rates we found were highly above chance level” further noting that “A promising approach to handle the resulting information overload is to automate the process of knowledge extraction using data mining techniques, thereby extracting novel information and relationships between biological features” (Abstract and introduction Molnar)
Regarding Claim 2
Lee/Chang/Molnar teach claim 1
Lee teaches wherein the requesting of the recoding and uploading of the sound data of the sounds for each intention or emotion of the user’s companion animal is to request the recording and uploading of sounds expressing some intentions or emotions among all intentions or emotions which are expressed by the sounds for each intention or emotion of the user's companion animal ( pg 6 para 004 “The analysis unit 320 may analyze the sound data based on the pre-stored animal sound translation algorithm when the sound data is collected from the heterogeneous animal communication terminal 100. At this time, the pre-stored animal sound translation algorithm detects, analyzes, and transposes a note, a pitch, and a scale of sound data, and then compares it with reference data of a previously stored type of communication object And extracting the emotions. To this end, emotions can include feelings of anxiety, loneliness, friendliness, boundaries, and anger.
Regarding Claim 3
Lee/Chang/Molnar teach claim 2
Lee teaches,
checking, by the management server, a breed of the user's companion animal, wherein the requesting of the recoding and uploading of the sound data of the sounds for each intention or emotion of the user’s companion animal is to request the recording and uploading of the sounds expressing some intentions or emotions corresponding to the checked breed (pg 9 para 07 “In order to normalize the animal sound patterns and emotional expressions, each animal sound such as a number of breeds, sizes (weight), etc. are classified into frequency bands, sorted by scales, and analyzed using the same composition. Accordingly, in order to distinguish the sound patterns of the animals from the sounds of the emotional expressions and to perform a precise analysis, the quantized notes are firstly quantized, the increase and decrease of the notes, the lowest notes and the highest The notes and notes specific to each breed are detected and analyzed to analyze the feelings of the animals first” )
Regarding Claim 4
Lee/Chang/Molnar teach claim 3
Lee teaches,
Regarding Claim 4 Lee teaches wherein the checking of the breed comprises requesting images of the user's companion animal for the companion animal application and analyzing an image of the user's companion animal received from the companion animal application for checking the breed (pg 8 para 005 “The heterogeneous animal can be photographed and input, and the facial expression and behavior pattern of the divided subject can be identified from the photographed image and output to the user terminal 400 through the heterologous animal communication terminal 100. [ The subject wearing the heterologous animal communication terminal 100 can be photographed and inputted. The subject wearing the heterologous animal communication terminal 100 can be distinguished from the photographed image, and the facial expression and behavior pattern of the subject can be identified Can be identified” identification of the animal via photography and analysis amounts to requesting images for checking the breed.)
Regarding Claim 6
Lee/Chang/Molnar teach claim 1
Lee teaches,
wherein the refining and processing as the training data is to combine and classify the sound data of the sounds for each intention or emotion of the user’s companion animal together with pitches of the sounds for each intention or emotion of the user’s companion animal, the duration of the sounds for each intention or emotion of the user’s companion animal, the repetition number of the sounds for each intention of emotion of the user’s companion animal (pg 8 para 004 “the pre-stored animal sound translation algorithm can be based on an animal translation database which interprets animal sounds including sound cycle, frequency, pitch, vibration, pattern, and interval by animal type and anatomical structure have…Here, the animal translation database may be a database that is updated based on the sound of the animal wearing the heterologous animal communication terminal 100 and the feedback data of the user terminal [intention of the user’s companion animal] 400 and adapted to be customized using a gene learning algorithm” the gene learning algorithm adapts or refines the processing as the training data to combine and classify with pitch, duration, and frequency or repetition number)
Molnar however teaches, in addition to the characters expressing the intention or emotion, refine and process the combined and classified sound data as the training data, and train the artificial intelligence model (pg 2 “The machine is Wrst trained on a set of labeled examples and then tested on a second set for which it must predict the labels. During the training phase, parameters of the models are tuned automatically by the learning algorithm in order to obtain the best classiWcation performances on the training set …In this study, we analyzed more than 6,000 barks recorded in diVerent situations from several individuals using a computerized method… The most relevant descriptors are then fed into machine-learning algorithms to produce classiWers (or extractors)… We collected bark recordings in seven diVerent behavioral contexts, most of which could be arranged at the homes of the owners… “Stranger”… “Fight”...” pg 3 “As each recording possibly contained up to three or four barks, individual bark sounds were manually segmented and extracted.” the training data combines the animals sounds or barks with labeled input by users expressing the context or intent, which is refined and processed during training.)
Regarding Claim 7
Lee/Chang/Molnar teach claim 6
Lee teaches,
generating, by the management server, companion animal sounds corresponding characters expressing the intention or emotion by an inference system of the artificial intelligence model based on deep neural network machine learning together with pitches of the companion animal sounds, a duration of the companion animal sounds, and a repetition number of the companion animal sounds, when the user inputs intention or emotion characters to be expressed through the inference system including the trained artificial intelligence model ((pg 8 para 003 “On the other hand, when the text from the user terminal 400 to the heterologous animal communication terminal 100 is collected, the heterologous animal bidirectional communication service providing server 300 transmits the collected text through the inverse translation process of the previously stored animal sound translation algorithm It can be created with animal sounds. Then, the heterologous animal bidirectional communication service providing server 300 can transmit the generated animal sounds to the heterologous animal communication terminal 100. Accordingly, even if the user is a person and a communication object is personal, not only the voice of the user is analyzed and transmitted to the user, but also the user's text and the like are converted into four languages so that the intention and feelings of the opening user can be known” pg 4 para 0005 “The heterogeneous animal bidirectional communication service providing server 300 analyzes and patterns the sound by type, object, etc., and performs machine learning and deep running using the big data…In addition, the heterologous animal bidirectional communication service providing server 300 analyzes sounds to be expressed by animals using a note, an interval, and a scale, transposes the sounds…the animal sound translation algorithm can measure intensity and duration, and can classify emotional expression state according to intensity and duration.” The classification and generation of sounds is based on the patterns of the sounds which consist as described in the art at the pitch, duration in addition to user inputs.)
Claims 5 and 8 are rejected under 35 U.S.C. 103 as being unpatentable over Lee/Chang/Molnar further in view of Chen CN 104700829 A.
Regarding Claim 5
Lee/Chang/Molnar teaches claim 3
Lee discloses … an artificial intelligence model which has been pre-trained using pre-training data for the breed of the user's companion animal … pg 8 para 004 Also, the pre-stored animal sound translation algorithm can be based on an animal translation database which interprets animal sounds including sound cycle, frequency, pitch, vibration, pattern, and interval by animal type and anatomical structure [breed] have. Here, the animal translation database may be a database that is updated based on the sound of the animal wearing the heterologous animal communication terminal 100 and the feedback data of the user terminal 400 and adapted to be customized using a gene learning algorithm, The algorithm can be an algorithm that bi-directionally translates by storing the mutual mapping of animal sounds and people, and is modified to be customized by a genetic learning algorithm”)
wherein the training is to train the selected artificial intelligence model with the uploaded sound data of the sounds for each intention or emotion of the user’s companion animal (pg 8 para 002 “At this time, the heterogeneous animal bidirectional communication service providing server 300 performs data mining and learning using the collected data as big data, stores the categorized pattern for each type, estimates the disease of the communication object and transmits it to the user terminal 400 “ learning based on the collected data amounts to training with uploaded sound data. In the context of the art the data is sound data of the sounds of the animal.)
Lee may not explicitly disclose [selecting], by the management server, an artificial intelligence model which has been pre-trained using pre-training data for the breed of the user's companion animal [from a plurality of artificial intelligence models];
However, in analogous art, Chen discloses selecting, by the management server, an artificial intelligence model which has been pre-trained using pre-training data for the breed of the user's companion animal from a plurality of] artificial intelligence models ([0028, G; Chen] “comparing the posterior probability of each model (multiple models), selecting the maximal posterior probability corresponding to the emotion model”);
It would be obvious to one with ordinary skill in the art to combine Lee/Chang and Chen because Lee teaches a model for [lee, 8] “providing a two-way communication service for heterogeneous animals according to an embodiment of the present invention” using big data and AI while Chen also uses [0002 Chen] “machine learning and artificial intelligence development, people can through machine translation technology to communicate in different languages, therefore, this technology makes it possible to animal voice emotion recognition,” where both teaches AI but Chen has multiple models for [0028, Chen] “selecting the maximal posterior probability”
Regarding Claim 8
Lee/Chang/Molnar teaches claim 4
The limitations of claim 8 are rejected for the reasons set forth in the rejection of claim 5
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JOHNATHAN R GERMICK whose telephone number is (571)272-8363. The examiner can normally be reached M-F 7:30-4:30.
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, Kakali Chaki can be reached on 571-272-3719. 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.
/J.R.G./
Examiner, Art Unit 2122
/KAKALI CHAKI/Supervisory Patent Examiner, Art Unit 2122