DETAILED ACTION
This Office action is in response to Applicant's amendment and request for
reconsideration filed on January 27, 2026.
Claims 1, 5, 18, 20, and 27-42 are pending.
Response to Arguments
Applicant's arguments filed January 27, 2026 have been fully considered but they are not persuasive.
With respect to Applicant’s argument regarding the teachings of Lepeska US 9,037,638, see page 9, - “That is, the information related to the ‘download time’ mentioned in the cited portion of Lepeska is a statistical value, for example, an average or median value related to a server limited object download time. It is used to indicate the server performance. This is quite different from the expected download time recited in claim 1…”, the Examiner respectfully disagrees.
Regardless of whether download time in Lepeska is a statistical measure or indicative of server performance, similar to what is claimed, the download time in Lepeska still provides an expected time, i.e., “milliseconds from first to last byte in the response” (see for example, col. 43, line 16) to download an object related to a communication path/URL, and thus broadly reads on the “expected download time” as claimed.
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 of this title, 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.
Claims 1 and 27-39 are rejected under 35 U.S.C. 103 as being unpatentable over Han et al. (US 2021/0409789)(“Han”) in further in view of Crabtree et al. (US 2017/0118263)(“Crabtree”) and Lepeska et al. (US 9,037,638)(“Lepeska”).
As per claim 1, Han teaches a method performed by a wireless transmit/receive unit (WTRU) (i.e., client, see abstract), comprising:
receiving information (i.e., manifest) indicating a plurality of network communication paths (i.e., URLs) that are available for downloading an AI/ML model (i.e., trained content aware DNNs, see abstract, also see ¶0041-0042, i.e., “…divided scalable content-aware DNNs are stored as URLs in the video configuration or manifest file”), wherein said information further includes AI/ML information (i.e., information about trained content aware DNNs, see ¶0025, also see ¶0051, i.e., “index information of the content-aware DNNs”) …;
determining a plurality of AI/ML chunks (e.g., a requisite component 510 and an optional component 520, see ¶0041, and/or see ¶0056, i.e., DNN chunks to download) for said AI/ML model based on said received information (¶0041-0042, where the plurality of AI/ML chunks are determined from the manifest identifying the URLs of the divided DNN);
determining one network communication path (i.e., URL), from said plurality of communication network paths, to download a respective AI/ML chunk of said plurality of AI/ML chunks of said AI/ML model, based on said received information (see ¶0041-0042);
establishing communication with said one network communication path to download said respective model chunk of said AI/ML model (see ¶0043, i.e., “download the requisite component 510”, which implies establishing a network connection, based on a URL corresponding to the requisite component/DNN chunk);
building at least a subset of said AI/ML model based on said respective AI/ML model chunk of said AI/ML model (see ¶0043, i.e., “…execute the requisite component 510 of the scalable content-aware DNN”, where impliedly the in order to execute the requisite component 510, the requisite component must be first built/compiled); and
performing inference on said at least a subset of said AI/ML model (see ¶0043, i.e., “…execute the requisite component 510 of the scalable content-aware DNN”, also see ¶0032 and ¶0035, which provides further evidence of performing inference using the executed DNN).
As per claim 1, Han does not expressly teach wherein said information further … indicates a communication mode and expected download time for each one of said plurality of network communication paths.
Nevertheless, first, in the same art of downloading data in chunks according to a manifest, Crabtree teaches modifying a manifest file to indicate a communication mode (i.e., unicast or multicast) for delivery of data chunks (see ¶0036).
It would have been obvious to a person having ordinary skill in the art, prior to the earliest effective filing date of the claimed invention, to modify the manifest in Han, based on the teachings of Crabtree, to signal a communication mode (i.e., unicast or multicast). The obvious motivation for doing so would have been to optimize delivery of model chunks for the AI/ML model based on a number of requesting clients.
Secondly, in the same art of downloading data in chunks according to a manifest, Lepeska teaches a manifest file that includes supplemental information including a download time of an object indicated by the manifest (see col. 9, lines 38-50).
It would have been obvious to a person having ordinary skill in the art to modify the manifest file in Han, based on the teachings Lepeska, to further indicate an expected download time for each one of said chunks/objects indicating by the plurality of network communication paths. The obvious motivation for doing so would have been to enable improved prefetching and/or prioritization of chunk/object retrieval from the manifest (see for example, Lepeska, col. 5 lines 7-13).
As per claim 27, Han does not teach, however, Crabtree further teaches wherein said communication mode includes at least one of unicast, multicast, multicast carousel mode (i.e., unicast or multicast, see ¶0036).
The same motivation that was utilized for combining Han and Crabtree in claim 1 applies equally well to claim 27.
As per claim 28, Han further teaches wherein said AI/ML model information includes at least one of … a model size (i.e., a number of layers, see ¶0037).
As per claim 29, although Han teaches the AI model supporting an incremental model type (see ¶0041, “the content-aware DNN may be divided into a requisite component 510 and an optional component 520”, read as demonstrating an incremental execution), Han does not expressly teach the information about the DNN in the manifest (see ¶0037) including a “model usage type” indicating the capability (i.e., “…wherein said model usage type includes at least one of a full model type and an incremental model type”).
Nevertheless, prior to the earliest effective filing date of the claimed invention, making explicit in the manifest/configuration file a model usage type indicating an incremental model type for the AI model would have been obvious or a matter of common sense to a person having ordinary skill in the art (see MPEP §2141).
The obvious/common sense motivation for configuring the manifest to explicitly include a “model usage type” having at least an “incremental model type” designation - indicating the incremental execution capability of the AI model in Han – would have been to allow the client to easily determine whether or not all components in the manifest are necessary for executing the AI model.
As per claim 30, although Han teaches the AI model supporting adaptive bitrate or adaptive bitrate streaming and/or the ability to adapt to the client’s capabilities (see ¶0032 and ¶0036), read as an “adaptive type” capability, Han does not expressly teach the information about the DNN in the manifest (see ¶0037) including a “model usage type extension” indicating the capability (i.e., “…wherein said model usage type extension includes at least one of a regular type, a specialization type, and an adaptive type”).
Nevertheless, prior to the earliest effective filing date of the claimed invention, making explicit in the manifest/configuration file a model usage type extension - indicating the adaptive capability of the AI model - would have been would have been obvious or a matter of common sense to a person having ordinary skill in the art (see MPEP §2141).
The obvious/common sense motivation for configuring the manifest to explicitly include a “model usage type extension” having at least an “adaptive type” designation - indicating the capability of the AI model to support adaptive processing – would have to been to easily allow the client/user of the AI model to understand the capabilities of the AI model and thus determine whether the capabilities satisfy the client/user's needs or requirements.
As per claim 31, Han does not expressly teach, however in the same art as noted above, Crabtree further teaches wherein said AI/ML model information includes, for a model chunk, at least one of … a chunk number (i.e., “CHUNK_INDEX parameter”, see ¶0054).
It would have been obvious to a person having ordinary skill in the art, prior to the earliest effective filing date of the claimed invention, to modify the manifest in Han to indicate AI/ML model chunks using index numbers as taught by Crabtree. The obvious motivation for doing so would have been for easier tracking or identification purposes.
As per claim 32, Han further teaches wherein said information for a network communication path further indicates a network address (i.e., URLs, see ¶0042).
As per claim 32, Han does not expressly teach the information further indicates at least a network link type.
Nevertheless, in the same art as noted above, Lepeska further teaches the manifest indicates a network link type (i.e., a primary vs. alternate type of URL/network link, see col. 9, lines 50 – 65).
It would have been obvious to a person having ordinary skill in the art, prior to the earliest effective filing date of the claimed invention, to modify the manifest in Han with the ability to indicate different network links for downloading AI/ML model chunks. The obvious motivation for providing different sources/network links would have been to allow for fallback in case of poor latency or failure with respect to a primary URL source.
Claims 33-39 are rejected under the same rationale as claims 1 and 27-32 since they recite substantially identical subject matter. Any differences between the claims do not result in patentably distinct claims and all of the limitations are taught by the above cited art.
Claims 5, 18, 20, and 40-42 are rejected under 35 U.S.C. 103 as being unpatentable over Han, Crabtree and Lepeska, in further view of Kumar Addepalli et al. (US 2020/0265493)(“Kumar Addepalli”).
As per claim 5, Han teaches a method performed by a server, comprising:
selecting an AI/ML model for an event (i.e., selecting a DNN for rendering video based on a client request, see abstract and Fig. 1)…, said AI/ML model including a plurality of model chunks (e.g., a requisite component 510 and an optional component 520, also see ¶0041 and/or see ¶0056, i.e., “DNN chunks”);
generating information (i.e., a manifest) indicating a plurality of network communication paths (i.e., URLs) that are available for downloading respective model chunks of said plurality of model chunks of said AI/ML model (see ¶0025, ¶0041-0042 and ¶0049) and wherein said information further includes AI/ML model information (i.e., information about trained content aware DNNs, see ¶0025, also see ¶0051, i.e., “index information of the content-aware DNNs”); and
transmitting said generated information to a WTRU (i.e., client, see ¶0051).
Han does not expressly teach receiving model subscription information from a wireless transmit/receive unit (WTRU); and wherein the selecting and generating of the AI/ML model is based on said model subscription information.
Nevertheless, in the same art of AI/ML modeling, Kumar Addepalli teaches a system requiring a subscription (i.e., valid member credentials) by a User Device for accessing trained AI/ML models (see Fig. 3, abstract, ¶0031, and ¶0119-0120).
It would have been obvious to a person having ordinary skill in the art, prior to the earliest effective filing date of the claimed invention, to modify the teachings of Han with the teachings of Kumar Addepalli for requiring subscription information for accessing provided AI/ML model information in Han. The obvious motivaton for doing so would have been for revenue generating purposes.
Furthermore, Han does not expressly teach wherein said information further … indicates at least one of a communication mode and expected download time for each one of said plurality of network communication paths
Nevertheless, first, in the same art of as noted above, Crabtree teaches modifying a manifest file to indicate a communication mode (i.e., unicast or multicast) for delivery of data chunks (see ¶0036).
The same motivation that was utilized for combining Han and Crabtree in claim 1 applies equally well to claim 5.
Secondly, in the same art as noted above, Lepeska teaches a manifest file that includes supplemental information including a download time of an object indicated by the manifest (see col. 9, lines 38-50).
The same motivation that was utilized for combining Han and Lepeska in claim 1 applies equally well to claim 5.
As per claim 18, Han does not expressly teach, however in the same art as noted above, Crabtree further teaches wherein said AI/ML model information includes, for a model chunk, at least one of … a chunk number (i.e., “CHUNK_INDEX parameter”, see ¶0054).
It would have been obvious to a person having ordinary skill in the art, prior to the earliest effective filing date of the claimed invention, to modify the manifest in Han to indicate AI/ML model chunks using index numbers as taught by Crabtree. The obvious motivation for doing so would have been for easier tracking or identification purposes.
As per claim 20, Han further teaches wherein said information for a network communication path further indicates a network address (i.e., URLs, see ¶0042).
As per claim 20, Han does not expressly teach the information further indicates at least a network link type.
Nevertheless, in the same art as noted above, Lepeska further teaches the manifest indicates a network link type (i.e., a primary vs. alternate type of URL/network link, see col. 9, lines 50 – 65).
It would have been obvious to a person having ordinary skill in the art, prior to the earliest effective filing date of the claimed invention, to modify the manifest in Han with the ability to indicate different network links for downloading AI/ML model chunks. The obvious motivation for providing different sources/network links would have been to allow for fallback in case of poor latency or failure with respect to a primary URL source.
Claims 40-42 are rejected under the same rationale as claims 5, 18, and 20 since they recite substantially identical subject matter. Any differences between the claims do not result in patentably distinct claims and all of the limitations are taught by the above cited art.
Conclusion
THIS ACTION IS MADE FINAL. 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 BRENDAN HIGA whose telephone number is (571)272-5823. The examiner can normally be reached Monday - Friday 8:30 AM - 5:00 PM.
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, Wing Chan can be reached on (571) 272-7493. 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.
/BRENDAN Y HIGA/Primary Examiner, Art Unit 2441