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 .
Response to Arguments
This office action is in response to amendments filed 05/13/2026. Claims 1-2, 6, 12-13, 15-16, 20, 25-26 and 30-39 are pending.
Applicant’s arguments and amendments to the claims with respect to prior art rejections of Claims 1-2, 6, 12-13, 15-16, 20, 25-26 and 30-39 under 35 USC 102/103 have been fully considered but are not persuasive.
With respect to applicant’s arguments that Hu fails to teach “determine which nodes in said neural network are to be connected responsive to said compression factor, [and] configure said neural network responsive to said determining” in Claim 1 and “wherein said one or more processors are configured to determine which nodes in said network are to be connected when said compression factor is adjusted” in Claim 6, examiner respectfully disagrees. Examiner agrees with applicant’s characterization of the Hu reference, but applicant appears to be comparing the Hu reference to a narrower scope than is defined by the claims. Hu teaches selecting between different predetermined compressors at runtime that are inserted into the network to compress data for transmission between compute devices (see page 160). The different compressor layers will have different node connections in order to achieve different compression ratios (see Table 1), and the neural network is configured by inserting the compression layers by loading them to each corresponding device. The way Hu determines node connections and configures the neural network may be different from what applicant has disclosed but still falls within the scope of the claim language.
With respect to applicant’s arguments that Hu fails to teach “wherein only one DNN model is loaded to said device for different compression factors” in Claim 25, examiner respectfully disagrees. Hu discloses selecting a single autoencoder that corresponds to each compression factor. This falls within the scope of the claim because for each different compression factor, only one DNN model (the selected autoencoder) is loaded (see page 160-0161). If applicant would like this limitation to require being able to change the compression factor without loading a new DNN model, the claim language needs to be amended.
With respect to applicant’s arguments that the features relied upon in Shi and the invention of Hu cannot be combined, examiner respectfully disagrees. The way Hu and Shi communicate compression factors between devices is not mutually exclusive. Examiner agrees that the Hu teaches the runtime scheduler communicating to the sending device and receiving device which models to load. However, this does not mean that communicating the compression factor directly between devices would always be redundant. One of ordinary skill in the art would understand there could be situations where signal bandwidth or connection may be stronger between two devices directly than with the runtime scheduler taught by Hu or the base station taught by Shi. Shi teaches that providing flexibility in the communication in transmission schemes enables the system to adapt the schemes to make the system more efficient in more situations (see par. 0076)
Claim Rejections - 35 USC § 102
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claim(s) 1, 6, 12-13, 15, 20, 25-26, 30, 32-35, 37-39 is/are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Hu et al (D. Hu and B. Krishnamachari, "Fast and Accurate Streaming CNN Inference via Communication Compression on the Edge," 2020 IEEE/ACM Fifth International Conference on Internet-of-Things Design and Implementation (IoTDI), Sydney, NSW, Australia, 2020, pp. 157-163, doi: 10.1109/IoTDI49375.2020.00023., hereinafter Hu)
Regarding Claim 1, Hu teaches:
1. A Wireless Transmit/Receive Unit (WTRU) (see at least wireless communication between devices and device i starting on page 158 ) , comprising:
a receiver configured to receive a part of a Deep Neural Network (DNN) model, wherein said part is before a split point of said DNN model (see at least determining optimal DNN split points for load balancing between devices in a Wi-Fi environment on page 159 ) , and
wherein said part of said DNN model includes a neural network to compress feature at said split point of said DNN model; (see at least adaptively inserting AutoEncoder encoder portion to compress data on device i before wireless communication on page 159-161)
one or more processors (see at least Rasberry Pi and NVIDIA Jetson on page 159) configured to:
obtain a compression factor for said neural network (see compression ratio on page 159),
determine which nodes in said neural network are to be connected responsive to said compression factor, configure said neural network responsive to said determining, and (see adaptive communication compression with different autoencoder architectures to achieve a desired compression ratio on page 160)
perform inference with said part of said DNN model to generate compressed feature; and (see generation of compressed embedding on page 160)
a transmitter configured to transmit said compressed feature to another WTRU. (see transmitting of compressed intermediate activations between edge devices on page 158)
Regarding Claim 6, Hu teaches:
6. The device of claim 1, wherein said one or more processors are configured to determine which nodes in said network are to be connected when said compression factor is adjusted. (see at least "Lastly, by setting the layer parameters such as number of channels and stride, we can easily achieve arbitrary compression ratio." on page 160 )
Regarding Claim 12, Hu teaches:
12. The device of claim 1, wherein at least one of said split point and said compression factor is adapted based on one or more of (1) physical layer operations, (2) Media Access Control layer operations, (3) Radio Resource Control layer operations, (4) available processing resources and (5) control signaling. (see load balancing with optimal split points based on edge device capability on page 159 and adapting compression factor based on real-time bandwidth availability in Algorithm 2 on page 161)
Regarding Claim 13, Hu teaches:
13. The device of claim 1,
wherein at least one of said split point and said compression factor is adapted based on a transmission data rate. (see load balancing with optimal split points based on edge device capability on page 159 and adapting compression factor based on real-time bandwidth availability in Algorithm 2 on page 161)
Regarding Claim 15, Hu also teaches:
15. A method performed by a Wireless Transmit/Receive Unit (WTRU), the method comprising:
Performing, step by step, each function of the device of Claim 1 (see Claim 1 analysis for rejection of the device.)
Regarding Claim 20, Hu also teaches:
A method for performing, step by step, each function of the device of Claim 6 (see Claim 6 analysis for rejection of the device.)
Regarding Claim 25, Hu also teaches:
25. The method of claim 15, wherein only one DNN model is loaded to said device for different compression factors. (see inserting one suitable compressor neural network in the pipeline before the communication stage on page 159-160)
Regarding Claim 26, Hu also teaches:
A method for performing, step by step, each function of the device of Claim 12 (see Claim 12 analysis for rejection of the device.)
Regarding Claim 30, Hu also teaches:
30. A Wireless Transmit/Receive Unit (WTRU) (see at least wireless communication between devices and device i+1 starting on page 158 ), comprising:
a receiver configured to receive a part of a Deep Neural Network (DNN) model, wherein said part is after a split point of said DNN model (see at least determining optimal DNN split points for load balancing between devices in a Wi-Fi environment on page 159 ), and
wherein said part of said DNN model includes a neural network to expand feature at said split point of said DNN model, wherein said receiver is also configured to receive one or more features output from another WTRU (see at least adaptively inserting AutoEncoder decoder portion to decompress data after wireless communication on device i+1 on page 160-161 ); and
one or more processors (see at least Rasberry Pi and NVIDIA Jetson on page 159) configured to:
obtain a compression factor for said neural network (see compression ratio on page 159),
determine which nodes in said neural network are to be connected responsive to said compression factor, configure said neural network responsive to said determining (see adaptive communication decompression with different autoencoder architectures to achieve a desired expansion ratio on page 160-161), and
perform inference with said part of said DNN model, using said one or more features output from another WTRU as input to said neural network. (see real-time inference on page 161)
Regarding Claim 32, Hu also teaches:
32. The device of claim 30, wherein said one or more processors are configured to determine which nodes in said network are to be connected when said compression factor is adjusted. (see at least "Lastly, by setting the layer parameters such as number of channels and stride, we can easily achieve arbitrary compression ratio." on page 160)
Regarding Claim 33, Hu teaches:
33. The device of claim 30,
wherein only one DNN model is loaded to said device for different compression factors. (see inserting one suitable compressor (including decoder portion on device i+1) neural network in the pipeline before the communication stage on page 159-160)
Regarding Claim 34, Hu teaches:
34. The device of claim 30, wherein at least one of said split point and said compression factor is adapted based on one or more of (1) physical layer operations, (2) Media Access Control layer operations, (3) Radio Resource Control layer operations, (4) available processing resources and (5) control signaling. (see load balancing with optimal split points based on edge device capability on page 159 and adapting compression factor based on real-time bandwidth availability in Algorithm 2 on page 161)
Regarding Claim 35, Hu also teaches:
35. A method, comprising:
Performing, step by step, each function of the device of Claim 30 (see Claim 30 analysis for rejection of the device.)
Regarding Claim 37, Hu also teaches:
Performing, step by step, each function of the device of Claim 32 (see Claim 32 analysis for rejection of the device.)
Regarding Claim 38, Hu also teaches:
Performing, step by step, each function of the device of Claim 33 (see Claim 33 analysis for rejection of the device.)
Regarding Claim 39, Hu also teaches:
Performing, step by step, each function of the device of Claim 34 (see Claim 34 analysis for rejection of the device.)
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.
Claim(s) 2, 16, 31, 36 is/are rejected under 35 U.S.C. 103 as being unpatentable over Hu et al (D. Hu and B. Krishnamachari, "Fast and Accurate Streaming CNN Inference via Communication Compression on the Edge," 2020 IEEE/ACM Fifth International Conference on Internet-of-Things Design and Implementation (IoTDI), Sydney, NSW, Australia, 2020, pp. 157-163, doi: 10.1109/IoTDI49375.2020.00023., hereinafter Hu) in view of Shi et al US 20200382929, hereinafter Shi)
Regarding Claim 2, Hu teaches:
The device of claim 1,
Hu does not appear to explicitly teach all of the following, but Shi does teach:
wherein said transmitter is further configured to send an indication of said obtained compression factor to said another WTRU. (see at least "In any of the examples, the transmitter may be further configured to: generate a control message or header indicating the selected transmission scheme and assigned sub-channel for each transmission feature; and transmit the control message or header to the receiving ED. " in par. 0022 and “An information representation scheme defines the format (e.g., sampling rates, compression rates, quantization, source encoding) used for the information to be transmitted. A transmission scheme defines the characteristics of the transmission signal (e.g., segmentation, coding length, coding rate, channel coding, modulation, and waveform). Generally, the information representation scheme is implemented by a feature encoder (also referred to as a source encoder), and the transmission scheme is implemented by a channel encoder. There may be multiple schemes available for use by one transmitting ED 110a. For example, multiple schemes may be defined, and may be stored in a local memory (e.g., the memory 258) of the ED 110a. The ED 110a may use one scheme for a particular transmission, and use another scheme for another transmission. Similarly, the BS 120 may have multiple scheme for transmitting to the receiving ED 110b, and may select a scheme to use as appropriate.” In par. 0066)
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to have modified the device taught by Hu to incorporate the teachings of Shi wherein the transmitting device sends header information defining a transmission scheme including the compression ratio for decoding the compressed feature. The motivation to incorporate the teachings of Shi would be to adapt the transmission scheme to changes in wireless communication conditions and efficiently reduce wireless traffic when possible (see par. 0076)
Regarding Claim 16, Hu as modified by Shi also teaches:
A method comprising performing, step by step, each function of the device of Claim 2 (see Claim 2 analysis for rejection of the device).
Regarding Claim 31, Hu also teaches:
31. The device of claim 30,
Hu does not appear to explicitly teach all of the following, but Shi does teach:
wherein said receiver is further configured to receive a signal indicative of said compression factor. (see at least "In any of the examples, the transmitter may be further configured to: generate a control message or header indicating the selected transmission scheme and assigned sub-channel for each transmission feature; and transmit the control message or header to the receiving ED. " in par. 0022 and “An information representation scheme defines the format (e.g., sampling rates, compression rates, quantization, source encoding) used for the information to be transmitted. A transmission scheme defines the characteristics of the transmission signal (e.g., segmentation, coding length, coding rate, channel coding, modulation, and waveform). Generally, the information representation scheme is implemented by a feature encoder (also referred to as a source encoder), and the transmission scheme is implemented by a channel encoder. There may be multiple schemes available for use by one transmitting ED 110a. For example, multiple schemes may be defined, and may be stored in a local memory (e.g., the memory 258) of the ED 110a. The ED 110a may use one scheme for a particular transmission, and use another scheme for another transmission. Similarly, the BS 120 may have multiple scheme for transmitting to the receiving ED 110b, and may select a scheme to use as appropriate.” In par. 0066)
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to have modified the device taught by Hu to incorporate the teachings of Shi wherein the receiving device receives header information defining a transmission scheme including the compression ratio for decoding the compressed feature. The motivation to incorporate the teachings of Shi would be to adapt the transmission scheme to changes in wireless communication conditions and efficiently reduce wireless traffic when possible (see par. 0076)
Regarding Claim 36, Hu as modified by Shi also teaches:
A method comprising performing, step by step, each function of the device of Claim 31 (see Claim 31 analysis for rejection of the device).
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 DYLAN M KATZ whose telephone number is (571)272-2776. The examiner can normally be reached Mon-Thurs. 8:00-6:00.
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, Abby Lin can be reached on (571) 270-3976. 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.
/DYLAN M KATZ/Primary Examiner, Art Unit 3657
Prosecution Insights