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 .
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 6-7 and 14-15 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Regarding claims 6 and 14, it is unclear if the multiple adapter networks, multiple training datasets and tasks or domains include those already recited in claims 1 and 9 or if they are in addition to those already recited in claims 1 and 9.
Regarding claims 7 and 15, “the neural” lacks antecedent basis in the claims.
Claim Rejections - 35 USC § 102
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 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.
Claim(s) 1-4 and 6-8 is/are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Ye (arXiv:2312.02515v1 [cs.LG] 5 Dec 2023).
Regarding claim 1, Ye teaches a method of parallel training a neural network to perform multiple tasks, the method comprising:
training in parallel:
a first adapter neural network in conjunction with the neural network using a first training dataset comprising first training samples of performing a first task (see Figure 3(b), A1 and B1 [a first adapter neural network] and “Pretrained Weights” [in conjunction with the neural network] with x1 which corresponds to A1 and B1, see §4.1, “In Figure 3(b), consider multiple LoRA fine-tuning jobs, J1,…,Jn. Each job Ji involves fine-tuning input data denoted as xi” [a first training dataset comprising first training samples of performing a first task]), and
a second adapter neural network in conjunction with a copy of the neural network using a second training dataset comprising second training samples of performing a second task (see Figure 3(b), A2 and B2 [a second adapter neural network] and “Pretrained Weights” [in conjunction with the neural network] with x2 which corresponds to A2 and B2, see §4.1, “In Figure 3(b), consider multiple LoRA fine-tuning jobs, J1,…,Jn. Each job Ji involves fine-tuning input data denoted as xi” [a second training dataset comprising second training samples of performing a second task]);
merging the trained first adapter neural network, the trained second adapter neural network and the neural network to produce an adapted neural network (Figure 3, H = Concat(h1 [trained first adapter neural network],h2 [trained second adapter neural network],…,hn) and Equation (2), W0X [the neural network], where “Concat” is a merging operation which merges the results of each network and produces an adapted neural network); and
generating, by the adapted neural network, a first task output or a second task output in response to an input to perform the first task or the second task (Figure 3(b), “h1” and/or “h2” is a generated result from the adapted neural network and can be considered a first task output or a second task output in response to an input to perform the first task or the second task because data is input into the model (input to perform the first task or the second task) and the concatenation operation is the adapted neural network and ”H” is generated as a first task output or second task output because it results from the data corresponding to the first and/or second task).
PNG
media_image1.png
570
842
media_image1.png
Greyscale
Regarding claim 2, Ye teaches all of the limitations of claim 1, wherein training the first adapter neural network comprises:
jointly generating, by a combination of the first adapter neural network and the neural network, a training output based on a training input from the first training dataset (Figure 3(b), “H” is a training output generated based on a combination of the first adapter neural network and the neural network (see Equation (2)) and is generated based on x1, which is the first training dataset); and updating weights of the first adapter neural network based on a training loss computed from the training output while keeping weights of the neural network unchanged (§2, “..efficient fine-tuning by completely freezing the pre-trained model and introducing weight modifications through a trainable low-rank decomposition matrix…” [updating weights of the first adapter neural network based on a training loss computed from the training output while keeping weights of the neural network unchanged]).
Regarding claim 3, Ye teaches all of the limitations of claim 1, wherein
the first training dataset and the second training dataset are from different domains (see Figure 3, x1 can be from an “x1 domain” and x2 can be from an “x2 domain” i.e. from different domains).
Regarding claim 4, Ye teaches all of the limitations of claim 1, wherein
the first adapter neural network and the second adapter neural network are trained using different training methods (Figure 3(b) and Equation (2). The training is particular to each adapter neural network and therefore the specific manner in which each network is trained, i.e. the method, can be considered different because the parameters involved are not the same).
Regarding claim 6, Ye teaches all of the limitations of claim 1, further comprising:
training in parallel multiple adapter neural networks in conjunction with the neural network using multiple training datasets for multiple tasks or domains (see Figure 3(b), A1, B1…An,Bn [training in parallel multiple adapter neural networks with the neural network] see §4.1, “In Figure 3(b), consider multiple LoRA fine-tuning jobs, J1,…,Jn. Each job Ji involves fine-tuning input data denoted as xi” [using multiple training datasets for multiple tasks or domains]); and
selectively merging one or more of the trained multiple adapter neural networks with the neural network depending on an application request (see Equation 2 which merges one or more of the trained multiple adapter neural networks with the neural network. The merging is done “selectively” because each neural network is selected to be merged and it is done by the application running the training program, i.e. on an application request).
Regarding claim 7, Ye teaches all of the limitations of claim 1, wherein
the merging comprises merging a first set of layers of the trained first adapter neural network, a second set of layers of the trained second adapter neural network, and a third set of layers of the neural based on a per-layer basis (Equation (2), “Concat” is a merging operation and comprises merging ΔW1x1 which contains a first set of layers of the trained first adapter neural network, ΔW1x2 which contains a second set of layers of the trained second adapter neural network and W0X which contains a third set of layers of the neural [network] and their combination can be considered a “per-layer” basis because weights are added together based on their position in the corresponding matrices).
PNG
media_image2.png
112
692
media_image2.png
Greyscale
Regarding claim 8, Ye teaches all of the limitations of claim 1, wherein
the first adapter neural network and the second adapter neural network are trained in conjunction with a different neural network, wherein the different neural network is compatible with the neural network (see Equation (2), BnAnxn [a different neural network, wherein the different neural network is compatible with the neural network], which is trained in conjunction with the first adapter neural network and the second adapter neural network).
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.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claim(s) 9-12 and 14-16 is/are rejected under 35 U.S.C. 103 as being unpatentable over Ye (arXiv:2312.02515v1 [cs.LG] 5 Dec 2023) as applied to claims 1, 3, and 6-8 in view of Wang (US20210271956A1).
Regarding claim 9, Ye teaches a system performing the methodology of claim 9 (see rejection of claim 1), but does not teach the system comprising
a communication interface; a memory storing a plurality of processor-executable instructions; and one or more processors executing the plurality of processor-executable instructions to perform operations.
Wang teaches
a communication interface (Figure 5, 514 “I/O INTERFACES” [a communication interface]); a memory storing a plurality of processor-executable instructions (Figure 5, 504, “RAM”, “CACHE”, [a memory storing a plurality of processor-executable instructions]); and one or more processors executing the plurality of processor-executable instructions to perform operations (Figure 5, 502, [one or more processors executing the plurality of processor-executable instructions to perform operations]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Ye to include the components as taught by Wang in order to allow for efficient execution of the methodology of Ye on a computer.
Regarding claims 10-12 and 14-16, Ye as modified with Ye according to claims 2-4 and 6-8 teaches the limitations of claims 10-11 and 14-16.
Regarding claims 17-19, Ye as modified according to claims 9 and 11-12 teaches all of the limitations of claims 17-19.
Allowable Subject Matter
Claims 5, 13, and 20 are objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
Regarding claims 5, 13, and 20, Ye does not teach the particular training methods claimed. Furthermore, impermissible hindsight would be required to modify Ye to utilize the particular training methods for the first and second adapter neural networks because the prior art does not provide a prima facie case of obviousness for training the adapter networks using separate, particular training methodologies such as those claimed.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SCHYLER S SANKS whose telephone number is (571)272-6125. The examiner can normally be reached 06:30 - 15:30 Central Time, M-F.
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, Michael Huntley can be reached at (303) 297-4307. 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.
/SCHYLER S SANKS/Primary Examiner, Art Unit 2129
Prosecution Insights