DETAILED ACTION
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA . This action is responsive to claims filed 4/1/2026. Claims 1-18 are pending.
Claim Interpretation
The following is a quotation of 35 U.S.C. 112(f):
(f) Element in Claim for a Combination. – An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The claims in this application are given their broadest reasonable interpretation using the plain meaning of the claim language in light of the specification as it would be understood by one of ordinary skill in the art. The broadest reasonable interpretation of a claim element (also commonly referred to as a claim limitation) is limited by the description in the specification when 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is invoked.
As explained in MPEP § 2181, subsection I, claim limitations that meet the following three-prong test will be interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph:
(A) the claim limitation uses the term “means” or “step” or a term used as a substitute for “means” that is a generic placeholder (also called a nonce term or a non-structural term having no specific stru
ctural meaning) for performing the claimed function;
(B) the term “means” or “step” or the generic placeholder is modified by functional language, typically, but not always linked by the transition word “for” (e.g., “means for”) or another linking word or phrase, such as “configured to” or “so that”; and
(C) the term “means” or “step” or the generic placeholder is not modified by sufficient structure, material, or acts for performing the claimed function.
Use of the word “means” (or “step”) in a claim with functional language creates a rebuttable presumption that the claim limitation is to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites sufficient structure, material, or acts to entirely perform the recited function.
Absence of the word “means” (or “step”) in a claim creates a rebuttable presumption that the claim limitation is not to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is not interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites function without reciting sufficient structure, material or acts to entirely perform the recited function.
Claim limitations in this application that use the word “means” (or “step”) are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action. Conversely, claim limitations in this application that do not use the word “means” (or “step”) are not being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action.
This application includes one or more claim limitations that do not use the word “means,” but are nonetheless being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, because the claim limitation(s) uses a generic placeholder that is coupled with functional language without reciting sufficient structure to perform the recited function and the generic placeholder is not preceded by a structural modifier. Such claim limitation(s) is/are: data mapping unit (claim 1 and dependents); data encoding unit (claim 1 and dependents).
Because this/these claim limitation(s) is/are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, it/they is/are being interpreted to cover the corresponding structure described in the specification as performing the claimed function, and equivalents thereof.
If applicant does not intend to have this/these limitation(s) interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, applicant may: (1) amend the claim limitation(s) to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph (e.g., by reciting sufficient structure to perform the claimed function); or (2) present a sufficient showing that the claim limitation(s) recite(s) sufficient structure to perform the claimed function so as to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph.
Claim Rejections - 35 USC § 101
The 101 rejections have been withdrawn in view of the amendments.
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) 1-3, 10-12 are rejected under 35 U.S.C. 103 as being unpatentable over Pu (CN 107341113 A) in view of Lin (US 20160328646 A1).
For claim 1, Todorov discloses: a data compression device, comprising:
a data encoding unit used to encode plural data blocks of the plural items of data using at least two encoding modes to generate an encoding data, wherein the data are divided into the pleural data blocks by the data encoding unit (0037-65 gives a high level overview of the compression technique, with stage 1 steps 1-15 (0037-52) giving overview of data transformation; in particular, unused consecutive numbers in the number range are discovered in order to cheaply encode runs, with decision flags being set at step 8 (0045), step 10 (0047), and step 14 (0051) based on how many consecutive unused runs are detected, these decision flags being stored in the header (0053) for decoding; hence, these flags represent modes of encoding data; 0184-185: contemplate application to various data including pyramidally decomposed images, 2D images, video images, etc., hence, various files are received as a data stream for ingestion (fig.1:1), the data being divided into blocks in order to perform histogram calculation, etc. (fig.1:4)), and data size of each of the pleural data blocks encoded by the data encoding unit using each of the at least two encoding modes is calculated, and the pleural data blocks are encoded using an encoding mode producing smallest data size to generate the encoding data (ibid: comparisons at step 8, 10, 14 (0045, 47, 51) of the longest unused space represent a compression size heuristic for calculating how many zero runs are encoded, see 0055, hence, these calculations of unused runs of numbers in the histogram constitute a data size calculation for the encoded data, with the flags being set based on the smallest encoded data size, i.e., longest run of unused numbers);
wherein the encoding data includes an encoding mode column bit, the encoding mode column bit being used to record the encoding modes used in the data blocks, the encoding mode used in the data blocks (0053 flags in the header constitute encoding mode column bit), and the encoding modes used in the data block are not all identical (ibid: various decision flags would be set based on data histograms).
Todorov does not disclose: wherein the compression device is for a deep neural network;
a data mapping unit used to re-map pleural items of original data according to at least one offset value and a sign value to obtain pleural items of mapped data, wherein a distribution center of the pleural items of mapped data is aligned with 0 and all of the pleural items of mapped data are non-negative integers;
wherein the data is mapped data.
Lin discloses: wherein the compression device is for a deep neural network (fig.3b);
a data mapping unit used to re-map pleural items of original data according to at least one offset value and a sign value to obtain pleural items of mapped data, wherein a distribution center of the pleural items of mapped data is aligned with 0 (figs.6-7, 0066-69, fig.8, 0083: aligning moments via mapping for quantization, the quantization occurring based on the sign and magnitude values of a offset for the various distribution moments (e.g., mean)) and all of the pleural items of mapped data are non-negative integers (0059-61 contemplates fixed point quantization, however, greater than 1 resolution constitutes integer quantization; as the m.n format may be separate from a sign bit (“m does not include a sign bit”), the mapped data comprise non-negative integers);
wherein the data is mapped data (combination with Lin yielding application to quantized, aligned data).
It would have been obvious before the effective filing date to a person of ordinary skill in the art to modify the device of Todorov by incorporating the data mapping technique of Lin. Both concern the art of numerical storage formatting, and the incorporation would have, according to Lin, improve performance / complexity tradeoffs in neural network computation (0031-33).
For claim 2, Todorov modified by Lin discloses the device of claim 1, as described above. Todorov modified by Lin further discloses: wherein the pleural items of original data are pleural weights of the deep neural network (Lin 0058, 0064, fig.5).
For claim 3, Todorov modified by Lin discloses the device of claim 1, as described above. Todorov modified by Lin further discloses: wherein the pleural items of original data are pleural activation values of the deep neural network (Lin 0058, 0064, fig.5).
Claims 10-12 recite analogous methods and are hence rejected for the same reasons.
Claim(s) 4-6, 13-15 are rejected under 35 U.S.C. 103 as being unpatentable over Pu (CN 107341113 A) in view of Lin (US 20160328646 A1) in view of Elmer (US 20210157548 A1).
For claim 4, Todorov modified by Lin discloses the device of claim 1, as described above. Todorov modified by Lin does not disclose the limitations of claim 4.
Elmer discloses: wherein each of the pleural items of original data is in an integer format (0017).
It would have been obvious before the effective filing date to one of ordinary skill in the art to modify the device by incorporating the brain float technique of Lin. Both concern the art of neural networks, and the incorporation would have, according to Elmer, provided support for popular data types in the era of proliferation of neural network applications (0017).
For claim 5, Todorov modified by Lin discloses the device of claim 1, as described above. Todorov modified by Lin does not disclose the limitations of claim 5.
Elmer discloses: wherein each of the pleural items of original data is in a 16-bit brain floating-point (BF16) format or a 16-bit floating-point (FP16) format (0017).
It would have been obvious before the effective filing date to one of ordinary skill in the art to modify the device by incorporating the brain float technique of Lin. Both concern the art of neural networks, and the incorporation would have, according to Elmer, provided support for popular data types in the era of proliferation of neural network applications (0017).
For claim 6, Todorov modified by Lin modified by Elmer discloses the device of claim 5, as described above. Todorov modified by Lin modified by Elmer further discloses: wherein the data mapping unit is further used to re-map exponent parts of the pleural items of original data in the BF16 format according to the at least one offset value and the sign value (Lin fig.8, 0083: as the entire floating point number is remapped via quantization and the moment, the exponent parts of the original floating point data is mapped as well, the mapping occurring according the mean value and sign, see figs. 6-7, with Elmer 0017 disclosing the bf16 format).
Claims 13-15 recite analogous methods and are hence rejected for the same reasons.
Claim(s) 7, 16 are rejected under 35 U.S.C. 103 as being unpatentable over Pu (CN 107341113 A) in view of Lin (US 20160328646 A1) in view of Elmer (US 20210157548 A1) in view of Mellempudi (US 20180285733 A1).
For claim 7, Todorov modified by Lin modified by Elmer discloses the device of claim 6, as described above. Todorov modified by Lin modified by Elmer does not disclose the limitations of claim 7.
Mellempudi discloses: wherein when one of the exponent parts is 0, the data encoding unit does not encode corresponding sign bit and fraction (0025: removing zeros from message to further compression; as zeros would include exponential being zero, the zero is not encoded in the compressed message).
It would have been obvious before the effective filing date to one of ordinary skill in the art to modify the device by incorporating the quantization technique of Mellempudi. Both concern the art of neural network compression, and the incorporation would have, according to Mellempudi, compress size to improve efficiency in neural network operations.
Claims 16 recite analogous methods and are hence rejected for the same reasons.
Claim(s) 8, 17 are rejected under 35 U.S.C. 103 as being unpatentable over Pu (CN 107341113 A) in view of Lin (US 20160328646 A1) in view of Georgiadis (US 20200143249 A1).
For claim 8, Todorov modified by Lin discloses the device of claim 1, as described above. Todorov modified by Lin does not disclose the limitations of claim 8.
Georgiadis discloses: wherein the at least two encoding modes are at least two encoding modes of Golomb-Rice coding or n-bit fixed-length coding (0005, 0032-34 contemplates various forms of Golomb-Rice encoding, hence, combination with Todorov yielding selection of said multiple forms as potential compression algorithms).
It would have been obvious before the effective filing date to one of ordinary skill in the art to modify the device by incorporating the quantization technique of Georgiadis. Both concern the art of neural network compression, and the incorporation would have, according to Georgiadis, improve compression efficiency for neural networks, such as for low-power devices (0004).
Claims 17 recite analogous methods and are hence rejected for the same reasons.
Claim(s) 9, 18 are rejected under 35 U.S.C. 103 as being unpatentable over Pu (CN 107341113 A) in view of Lin (US 20160328646 A1) in view of Laude ("Neural network compression using transform coding and clustering", published 2018).
For claim 9, Pu modified by Lin discloses the device of claim 1, as described above. Pu modified by Lin does not disclose the limitations of claim 9.
Laude discloses: wherein the encoding data (p.2 fig.1 shows an output data with a compressed network alongside metadata specifying shape, offset, scale, etc., see p.3 col.1 for disclosure of the output file comprising metadata and unstructured weights) comprises a header column bit, an encoding mode column bit and the pleural data blocks which are encoded (fig.1: compressed network constitutes data blocks with the metadata constituting header and encoding metadata bits); the header column bit records the at least one offset value and the sign value (fig.1 contemplates offset value being encoded in a metadata or header in the output file, hence, the header column bit recording the offset value and sign), and the encoding mode column bit records one of the encoding mode used in each of the pleural data blocks (fig.1; combination with Todorov’s disclosure of using various encodings for each cluster yielding the metadata containing bits pertaining to the encoding mode being included in the metadata).
It would have been obvious before the effective filing date to one of ordinary skill in the art to modify the device by incorporating the encoding technique of Laude. Both concern the art of quantization based on data clustering (see fig.1), and the incorporation would have, according to Laude, address problems of memory footprint of neural networks by providing a complete code pipeline for neural network compression (p.1, “Introduction”).
Claims 18 recite analogous methods and are hence rejected for the same reasons.
Response to Arguments
Applicant’s arguments have been fully considered. In the remarks, Applicant argues:
1. Pu does not disclose the newly added limitations.
Applicant’s arguments are moot in view of newly applied art.
2. The amended claims do not evoke 101.
Examiner agrees.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Wang (US 20210160499 A1) discloses adaptive block compression for neural networks.
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 extension fee 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 date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to LIANG LI whose telephone number is (303)297-4263. The examiner can normally be reached Mon-Fri 9-12p, 3-11p MT (11-2p, 5-1a ET).
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor Jennifer Welch can be reached on (571)272-7212. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from Patent Center and the Private Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from Patent Center or Private PAIR. Status information for unpublished applications is available through Patent Center or Private PAIR to authorized users only. Should you have questions about access to Patent Center or the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free).
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. The examiner is available for interviews Mon-Fri 6-11a, 2-7p MT (8-1p, 4-9p ET).
/LIANG LI/
Primary examiner AU 2143
Prosecution Insights