Notice of Pre-AIA or AIA Status
1. The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
DETAILED ACTION
2. This office action is in response to communication filed on 12/17/2024. Claims1-20 are pending on this application.
Claim Rejections - 35 USC § 112
3. 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:
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.
4. Claims 2, 7, 12, 14-17 and 18-20 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.
Claim 2 recites the limitation "the value" in claim 1. There is insufficient antecedent basis for this limitation in the claim.
Claim 7 recites the limitation " the decoded weight” in claim 1. There is insufficient antecedent basis for this limitation in the claim.
Claim 12 recites the limitation " the preceding 0s" in claim 8. There is insufficient antecedent basis for this limitation in the claim.
Claim 14 recites the limitation “the preceding code” in the claim. There is insufficient antecedent basis for this limitation in the claim.
Claims 15-17 are depended on claim 14; therefore, claims 15-17 included insufficient antecedent basis of claim 14.
Claim 18 recites the limitation “the first bit” in the claim. There is insufficient antecedent basis for this limitation in the claim.
Claims 19-20 are depended on claim 18, therefore claims 19-20 included is insufficient antecedent basis of claim 18.
Claim Rejections - 35 USC § 102
5. 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.
6. Claims 1- 4, 6 and 14-17 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Satpathy et al. Pub. No. 2018/0341722.
Regarding claim 1. Fig. 2 of Satpathy et al. discloses an operating method of a decoding device (Fig. 2a) comprising processing hardware (202, 204, 206) and storage hardware (210, 212) the operating method (Fig. 2) comprising: receiving a compressed weight (paragraph 0031) comprising a preceding code and subsequent bits (XXX…XXX) following the preceding code (preceding numbers 0s or 1s of Huffman code of 212); and decoding (Fig. 2) the compressed weight (paragraph 0031) by applying the preceding code (preceding numbers 0s or 1s of Huffman code of 212) of the compressed weight (paragraph 0031) to a Huffman tree (paragraph 0034 discloses “Huffman tree 202”), wherein the Huffman tree (202) decodes the preceding code (preceding numbers 0s or 1s of Huffman code of 212), and wherein the decoded preceding code (preceding numbers 0s or 1s of Huffman code of 212) is joined with the subsequent bits (XXX…XXX) to form a decompressed version (206) of the compressed weight (paragraph 0031).
Regarding claim 2, the operating method of claim 1, Fig. 2 Satpathy et al. further discloses wherein the decoding the compressed weight (paragraph 0031) comprises: not decoding the subsequent bits (XXX…XXXX) that are subsequent to the preceding code (preceding numbers 0s or 1s of Huffman code of 212) while decompressing (206) the compressed weight (paragraph 0031) without converting the value (value of XXX….XXX).
Regarding claim 3. The operating method of claim 1, Fig. 2 further discloses wherein the Huffman tree (202) is configured to record a decoding value (value of Huffman code) corresponding to cases of the preceding code (preceding numbers 0s or 1s of Huffman code of 210, 212).
Regarding claim 4. The operation of claim 1, Fig. 2 further discloses wherein the preceding code (preceding numbers 0s or 1s of Huffman code of 212) is determined according to, in the weight before being compressed (paragraph 0031), the number of consecutive 0s (number of 0s in 212) preceding before a first 1 (first 1 of in 212) first in the weight before compression (paragraph 0031).
Regarding claim 6. The operating method of claim 1, Fig. 2a further discloses wherein the receiving the compressed weight (plain pay load 1015) comprises receiving the compressed weight (plain pay load 1015) stored in static random-access memory (paragraph 0077).
Regarding claim 14. Fig.1 and Fig. 2 of Satpathy et al. discloses decoding device comprising: one or more processors; a memory; and one or more programs stored in the memory that when executed by the one or more processors (paragraph 0117): receiving a compressed weight (paragraph 0031) comprising: a compression code (compression code 0s or 1s of Huffman code of 212) followed by subsequent uncompressed bits (XXX….XXX) adjoining the preceding code (preceding numbers 0s or 1s of Huffman code of 212) in the compressed weight (paragraph 0031); and decoding (202) the compressed weight (paragraph 0031) , to a decoded bit string (output of 202), by hashing the compression code (paragraph 0031) the compressed weight (weight of Huffman Code) with reference to a Huffman tree ((paragraph 0034 discloses “Huffman tree 202”), wherein the Huffman tree (202) comprises decoding information (Lenth of Symbol) of the compression code (Huffman Code) .
Regarding claim 15. The decoding device of claim 14, Fig. 2 further discloses wherein the decoding (202) the compressed weight (paragraph 0031) comprises: forming, in a decompressed weight (206) corresponding to the compressed weight (paragraph 0031), the decompressed bits (Symbol) joined (see Fig. 4 to the uncompressed bits (XXX….XXXX).
Regarding claim 16. The decoding device of claim 14, Fig. 2 further discloses wherein the compression code (paragraph 0031) is determined according to the number of 0s preceding before 1 (number of 0s preceding 1 in 212) first appears (first appear of 1 in 212) in a weight (symbol value) corresponding to the compressed weight (paragraph 0031).
Regarding claim 17. The decoding device of claim 14, Fig. 2 further discloses wherein the decoding the compressed weight (paragraph 0031) comprises: filling an insufficient bit (XXXX….XXX) of the compressed weight with 1 (1s in Huffman Coding) such that the length of the compressed weight (length of each Huffman Code) corresponds to a fixed length bit (each Huffman have same number of bits).
7. Claims 8 -11, 13, and 18-20 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Kim et al. Pub. No. 2022/0164143.
Regarding claim 8. Fig. 1, Fig. 5 and Fig. 6 of Kim et al. discloses an operating method of an encoding device (150; paragraph 122), the operating method (Fig. 5, Fig. 6) comprising: receiving a weight (weight of Soff Data in Fig. 6) , the weight (Soft Data) having a prefix of bits of all 0s (0’s prior to 1 and first 1 in soft data from left to right of Soft data) followed by a postfix (all bits after the first 1 of Soft Data) tarting with a bit of a 1 (first 1 of Soft data); determining the number of bits of 0s (number of all 0’s prior to the firs 1 of Soft Data) in the prefix (0’s prior to first 1 in soft data ) of the weight (weight of Soff Data in Fig. 6); compressing (Fig. 6) the prefix (0’s prior to first 1 in soft data ) into a code (Code alphabet in Fig. 5) that is predetermined (Fig. 5) according to the number of bits of 0s (number of all 0’s prior to the firs 1 of Soft Data) in the prefix (0’s prior to 1 and first 1 in soft data ); and forming a compressed version (compressed data) of the weight (weight of Soff Data in Fig. 6) by joining the code (Code alphabet in Fig. 5) with the postfix (all bits after the first 1 of Soft Data).
Regarding claim 9. The operating method of claim 8, Fig. 5 and Fig. 6 further comprising generating a Huffman tree (Fig. 5{500)) configured to record a compressing method (Huffman compressing of Fig. 5) based on which zeros-prefix lengths (0’s prefix length of Source alphabet in Fig. 5) are most common among a set (01…..000000000000001 ) of weights including the weight (weight of Soft Data).
Regarding claim 10. The operating method of claim 8, Fig. 5 and Fig. 6 further discloses wherein the compressing (Fig. 6) of the prefix 0’s prior to 1 and first 1 in soft data ) into the code (Code alphabet) comprises: when a first bit is 1 (first bit of Source alphabet), compressing the first bit 1 (first bit of source alphabet) and connecting a bit value of a subsequent bit (value of subsequence bits after the first 1 of soft data ) to the first bit 1 (first bit 1 of Soft data) to the compressed code (compressed data) and displaying the bit value (displaying value of subsequence bits after the first 1 of soft data).
Regarding claim 11. The operating method of claim 8, Fig. 5 and Fig. 6 further disclose wherein the compressing (Fig. 6) of the prefix (0’s prior to 1 and first 1 in soft data from left to right of Soft data) into the code (Code alphabet) comprises: converting bits (compressed data), excluding a first bit (first bit 1 of source alphabet) from among the bits of 0s (Source alphabet), into the code (code alphabet) that is predetermined according to the number of bits of 0s (predetermined number 0 of source alphabet).
Regarding claim 13. Fig. 1 of Kim et al. further discloses non-transitory computer-readable storage medium storing instructions (paragraph 0243) that, when executed by a processor (134), cause the processor (134) to perform the method of claim 8 (Fig. 5 and Fig. 6).
Regarding claim 18. Fig. 1, Fig. 5 and Fig. 6 of Kim et al. disclose an encoding device (150; paragraph 122) comprising: one or more processors; a memory storing instructions configured to cause the one or more processors (paragraph 0243) to: receive a weight (weight of Soft data in Fig. 6); determine the number of bits of consecutive 0s (Number of 0’s of prior to first 1 in soft data from left to right) in a prefix (number of 0’s prior to 1 and first 1 in soft data from left to right of Soft data) of the weight (weight of Soft data) before a first 1 in the weight (first 1 in the weight in soft data), the first 1 (first 1 of soft data from left to right) comprising a first bit (bit after the first 1 in soft data) in a postfix of the weight (bits after the first 1 in Soft data) that follows prefix (number of 0’s prior to 1 and first 1 in soft data from left to right of Soft data); compress the prefix (number of 0’s prior to 1 and first 1 in soft data from left to right of Soft data) according to the number of bits of 0s (Source alphabet) ; and form a compressed weight (Compressed data) corresponding to the weight (weight of soft data) by joining the compressed prefix (Code alphabet) with the postfix (bits after the first 1 of soft data).
Regarding claim 19. The encoding device of claim 18, Fig. 5 and Fig. 6 further discloses wherein the compressing of the prefix (number of 0’s prior to 1 and first 1 in soft data from left to right of Soft data) comprises: when a first bit in the prefix is 1 (first 1 of soft data) , compressing the first bit 1 (first 1 of soft data) and connecting a bit value after the first bit 1 (values of bits after the first 1 in soft data) to the compressed code (Code alphabet) and displaying the bit value (display the bit value of bits after the first 1 in soft data).
Regarding claim 20. The encoding device of claim 18, Fig. 5 and Fig. 6 further discloses wherein the compressing of the prefix (number of 0’s prior to 1 and first 1 in soft data from left to right of Soft data) comprises: mapping (mapping of Huffman table) the prefix (number of 0’s prior to 1 and first 1 in soft data from left to right of Soft data) to a code (Code alphabet) that is predetermined according to the number of bits of 0s of the prefix (number of 0’s prior to 1 and first 1 in soft data from left to right of Soft data).
Claim Rejections - 35 USC § 103
8. 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.
9. Claim 5 is rejected under 35 U.S.C. 103 as being unpatentable over Satpathy et al. applied to claim 1 above, in view of Wu et al. Pub. No. 2023/0037575.
Regarding claim 5. Fig. 2 of Satpathy et al. applied to claim 1 above further discloses do not disclose wherein the decoding the compressed weight comprises: decoding the compressed weight (Huffman code) comprises: filling an insufficient bit (XX…XX) of the compressed weight with (Huffman code) with 1’s such that a length of the compressed weight (number of bits of Huffman code) corresponds to a fixed length bit (fixed bit length of Huffman code).
However, the length of the compressed weight corresponds to a fixed length bit prior to compression.
Fig. 2 of Wu et al. discloses a compressed data (412) comprising: length of the compressed weight (412) corresponds to a fixed length bit (fixed length of uncompressed data 418) prior to compression (412).
Satpathy et al. and “Wu et al. are common subject matter of decompress based of Huffman code; therefore, it would have been obvious before the effective filing date of claimed invention to one ordinary skill in the art to which the claimed invention pertains to incorporate Wu et al. into Satpathy et al. for the purpose of providing lossless sparse encoding may reduce complexity of some data compression techniques by leveraging the fact that, for many machine learning (ML) use cases, weight and activation values can be sparse (e.g., contain many zero values).; the weights can be retrained to be sparse, for example, by setting near zero weights to zero and retraining the ML model to maintain classification accuracy (paragraph 0068 of Wu et al.)
10. Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over Satpathy et al. applied to claim 1 above, in view of Kung et al. U.S. patent Pub. No. 2021/0209190.
Satpathy et al. applied to claim 1 above do not disclose inputting the decoded weight to an operator for a multiply-accumulate (MAC) operation.
Fig. 9 of Kung et al. discloses a Huffman decoder (paragraph 0058) comprising inputting the decoded weight (outputs of Decoder 0 and Decoder 1) to an operator for a multiply-accumulate (MAC) operation (LUT; paragraph 0058).
Satpathy et al. and Kung et al. are common subject matter of compress and decompress; therefore, it would have been obvious before the effective filing date of claimed invention to one ordinary skill in the art to which the claimed invention pertains to incorporate Kung et al. into Satpathy et al. for the purpose of providing a matrix data processing apparatus performing a matrix multiplication operation of an artificial neural network may improve accuracy in data classification, categorization, or labelling by performing the pruning operation (paragraph 0043 of Kung et al.).
11. Claim 12 is rejected under 35 U.S.C. 103 as being unpatentable over Satpathy et al. applied to claim 1 above, in view of Joung et al. U.S. patent No. 8,842,7724.
Fig. 5 and Fig. 6 of Spatpathy et al. further discloses wherein the determining the number of bits of the preceding 0s (number of 0’ prior to the first 1 of Soft data).
However, Satpathy et al. do not discloses a counter corresponding to the number of bits of the weight; and determining a bit value of a digit corresponding to the counter while increasing the counter.
Fig. 2 and Fig. 3 of Jount et al. discloses a Huffman encoding (Col. 2 lines 62-63) comprising: a counter (S10) corresponding to the number of bits of a weight (Fig. 3 discloses a weight of data); and determining a bit value (S30, S40, S60) of a digit D’MSB, Q’MSB, P’LSB) corresponding to the counter (S10) while increasing the counter (S100).
Satpathy et al. and Joung et al are common subject matter of Huffman compression; therefore, it would have been obvious before the effective filing date of claimed invention to one ordinary skill in the art to which the claimed invention pertains to incorporate Kung et al. into Satpathy et al. for the purpose of providing an significantly reduce an amount of data transmitted and received between the DU and the RU in a way such that basic units of compression are defined as bundles of basic frames defined in a Common Public Radio Interface (CPRI) standard, and a header having information about an amount of data remaining after compression for each of the basic units is defined so as to be transmitted and received (Col. 1 lines 22-29 of Joung et al.).
Contact Information
12. Any inquiry concerning this communication or earlier communications from the examiner should be directed to Linh Van Nguyen whose telephone number is (571) 272-1810. The examiner can normally be reached from 8:30 – 5:00 Monday-Friday.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Mr. Dameon E. Levi can be reached at (571) 272-2105. The fax phone numbers for the organization where this application or proceeding is assigned are (571-273-8300) for regular communications and (571-273-8300) for After Final communications.
06/29/2026
/LINH V NGUYEN/Primary Examiner, Art Unit 2845
Prosecution Insights