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 .
Claims 1-61 are presented for examination.
Claim Objections
Claim 61 is objected to under 37 CFR 1.75(c) as being in improper form because a multiple dependent claim should refer to other claims in the alternative only, and cannot depend from any other multiple dependent claims. See MPEP § 608.01(n). Accordingly, the claim has not been further treated on the merits.
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 1-60 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 pre-AIA the applicant regards as the invention.
Claim(s) 1, 11, 21 and 41, each recite(s) “each of the N blocks comprising a sequence length that is shorter than ... input sequence length”. It is unclear whether all the N blocks have the exact same length or each of the N blocks has a respective sequence length, rendering the claim(s) indefinite. For examination purposes the examiner has interpreted “each of the N blocks comprising a sequence length that is shorter than ... input sequence length” to be “each of the N blocks comprising a respective sequence length that is shorter than ... input sequence length”.
Claim(s) 1 and 11, recite(s) “any mode of data” without any examples of modes to convey what constitutes a mode, and without limitations on modes, and therefore is an open-ended limitation, rendering the claim(s) indefinite.
Claim(s) 21 recite(s) 2 operations labelled (e), two operations labelled (f) and “operations (a) - (f)”, although the claim does not recite operations a-d, rendering the claim(s) indefinite.
Claim(s) 41 recite(s) 2 operations labelled (i), two operations labelled (j) and “operations (a) - (f)”, although the claim does not recite operations a-f, rendering the claim(s) indefinite.
Claim(s) 2 and 12, depend on claims 1 and 11 respectively and recite(s) “wherein operations (a) - (e)”. However claims 1 and 11 do not recite operation (e), rendering the claim(s) indefinite. For examination purposes the examiner has interpreted “wherein operations (a) - (e)” to be “wherein operations (a) - (d)”.
Claim(s) 22, depends on claim 21 and recite(s) “repeating operations (d) - (f)”. However claim 21 does not recite operation (d), rendering the claim(s) indefinite.
Claim(s) 42, depends on claim 41 and recite(s) “repeating operations (d) - (f)”. However claim 41 does not recite operations (d-f), rendering the claim(s) indefinite.
Claim(s) 6 and 16, recite(s) “each unit”. There is lack of antecedent basis for this limitation in these claim(s), rendering the claim(s) indefinite.
Claim(s) 2-10, 12-20, 22-40 and 42-60 do not contain claim limitations that cure the indefiniteness of claim(s) 1, 11, 21 and 41 respectively, and therefore are also indefinite under 35 U.S.C. 112(b).
Claim Rejections - 35 USC § 103
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 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.
Claims 1-31, 33-38, 40-51, 53-58, 60, are rejected under 35 U.S.C. 103 as being unpatentable over Szegedy (US 20240412054 A1), in view of Li et al “Sequence Parallelism: Long Sequence Training from System Perspective” published in Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics Volume 1: Long Papers, pages 2391–2404, July 9-14, 2023.
Regarding claim 1, Szegedy teaches a non-transitory computer readable medium having instructions thereon, the instructions when executed by a computer causing the computer to perform operations that provide long range attention in a machine learning model for an input comprising any mode of data, the operations comprising (Szegedy [10, 121] invention may be instructions in medium, and provides long range model attention for input sequence);
(a) dividing the input into N blocks, each of the N blocks ...shorter than an input sequence ..., the input received with a layer of a neural network of the machine learning model, the input comprising a sequence having the input sequence length (Szegedy [9, 11, 37, 38, 40] input sequence has sequence length and is split into sub-sequences (blocks that are shorter than sequence as they are portions of sequence), input may be received by neural network (model) block, model blocks can be layers);
(b) determining a first block output for a first of the N blocks (Szegedy [37, 38, 89, 90] output determined for a first sub-sequence);
(c) providing the first block output as input for determining a next block output for a next of the N blocks (Szegedy [37, 38 ] sub-sequence output is used to determine next block output); and
(d) repeating operations (b) - (c) for each remaining block, such that each next block output is determined based on output from a previous block, thus merging smaller blocks to retain context associated with the input and provide the long range attention in the machine learning model (Szegedy [11, 38, 41, 42, 46] each sub-sequence output is determined using previous sub-sequence output, and repeated for all sub-sequences, final output merges outputs for all sub-sequences, retaining long-range attentio).
Szegedy does not specifically teach each of the N blocks comprising a sequence length that is shorter than an input sequence length.
However Li teaches each of the N blocks comprising a sequence length that is shorter than an input sequence length, determine a next block output based on previous block output (Li Intro-1stBulletPoint, Secs 3 and 3.1, sub-sequences have length (L/N) less than input sequence length (L), each block output can be based on a previous block output, making sub-sequences length shorter than input and based on length limitations, is memory-efficient).
It would have been obvious to one of an ordinary skill in the art before the effective filing date of the claimed invention, to have incorporated the concept taught by Li of each of the N blocks comprising a sequence length that is shorter than an input sequence length, determine a next block output based on previous block output, into the invention suggested by Szegedy; since both inventions are directed towards determining model output for an input sequence by dividing the sequence into shorter sub-sequences, and incorporating the teaching of Li into the invention suggested by Szegedy would provide the added advantage of being memory efficient,
and the combination would perform with a reasonable expectation of success (Li Intro-1stBulletPoint, Secs 3 and 3.1).
Regarding claim 2, Szegedy and Li teach the invention as claimed in claim 1 above.
Szegedy further teaches wherein operations (a) - (e) facilitate providing the neural network with longer sequence lengths compared to what was possible with prior neural networks, by dividing the sequence length into smaller blocks to mitigate computational complexity, and subsequently merging the smaller blocks to retain the context associated with the input and provide the long range attention (Szegedy [11, 38, 41, 42, 46] each sub-sequence output is determined using previous sub-sequence output, and repeated for all sub-sequences, computational complexity is reduced, final output merges outputs for all sub-sequences, retaining long-range attention)
Regarding claim 3, Szegedy and Li teach the invention as claimed in claims 1 above.
Szegedy further teaches wherein a block output coming from previous sequence blocks comprises collective information from all previous sequence blocks (Szegedy [10] each current sub-sequence output is based on previous sub-sequence output, therefore current block uses information based on all previous sub-sequences).
Regarding claim 4, Szegedy and Li teach the invention as claimed in claim 1 above.
Szegedy further teaches wherein the input comprises text (Szegedy [24] inout can be text).
Regarding claim 5, Szegedy and Li teach the invention as claimed in claim 1 above.
Szegedy further teaches converting the input into the sequence, the sequence having units associated with the input, a quantity of the units comprising the input sequence length (Szegedy [41, 112, 114] input may be converted to embedding sequence).
Regarding claim 6, Szegedy and Li teach the invention as claimed in claim 1 above.
Szegedy further teaches converting each unit of the sequence into a vector and providing each vector to the layer (Szegedy [41, 112, 114] input units may be converted to vectors which as used as input to layer(s) of model)
Regarding claim 7, Szegedy and Li teach the invention as claimed in claim 1 above.
Szegedy further teaches wherein the neural network comprises a transformer module (Szegedy [120] neural network may have transformer module).
Szegedy does not specifically teach of a large language model
However Li teaches a transformer module of a large language model (Li Intro transformer module may be for large model which can be a language model).
Regarding claim 8, Szegedy and Li teach the invention as claimed in claim 7 above.
Szegedy further teaches wherein the layer of the neural network comprises a multi-head attention layer of the transformer module (Szegedy [40, 119, 120] layer can be multi-head attention layer of transformer).
Regarding claim 9, Szegedy and Li teach the invention as claimed in claim 8 above.
Szegedy further teaches wherein the multi-head attention layer comprises multiple self- attention modules, each self-attention module associated with one of the N blocks (Szegedy 105, 106, 119] multiple self-head modules may each process one of the sub-sequences).
Regarding claim 10, Szegedy and Li teach the invention as claimed in claim 9 above.
Szegedy further teaches wherein the multiple self-attention modules are configured to operate in parallel (Szegedy [119] multiple self-attention modules can work in parallel).
Claim 11 is directed towards a method performing instructions similar in scope to the instructions stored by the medium of claim 1, and is rejected under the same rationale.
Claim(s) 12-20 is/are dependent on claim 11 above, is/are directed towards a method performing instructions similar in scope to the instructions stored by the medium of claim(s) 2-10 respectively, and is/are rejected under the same rationale.
Regarding claim 21, Szegedy teaches a non-transitory computer readable medium having instructions thereon, the instructions when executed by a computer causing the computer to perform operations comprising (Szegedy [10, 121] invention may be instructions in medium, and provides long range model attention for input sequence):
(e) receiving an input (X) with a layer of a neural network, the input comprising a sequence having an input sequence length (ξ); (f) dividing the input into N blocks (Xo, X1, ..., Xi), each of the N blocks ... that is shorter than the input sequence ... (Szegedy [3, 9, 11, 37, 38, 40] received input sequence has sequence length and is split into sub-sequences (blocks that are shorter than sequence as they are portions of sequence), input may be received by neural network (model) block, model blocks can be layers);
(g) determining an initial key (Ko), query (Qo), and value (Vo) of a first (Xo) of the N blocks; (h) determining an initial key (Ki), query (Qi) and value (V1) of a next (X1) of the N blocks (Szegedy [48-50, 58-62, 101, 37, 38, 89, 90] (K,Q,V) of each sub-sequence is determined),
(e) determining a first block output (Xo ) for the first (Xo) of the N blocks based on the initial key (Ko), query (Qo), and value (Vo) of the first (Xo) of the N blocks (Szegedy [58-62, 101, 37, 38, 89, 90] output determined for a first sub-sequence based on (K,Q,V) of first sub-sequence); and
(f) providing the first block output (Xo) as input for determining a next block output (X1) for the next (Xi) of the N blocks, the next block output (Xi ) determined based on the first block output (Xo ), and the initial key (Ki), query (Qi), and value (Vi) of the next (Xi) of the N blocks (Szegedy [37, 38, 58-62, 101] for each current subsequent sub-sequence, based on (K,Q,V) of current sub-sequence and output of previous sub-sequence, determine current block output);
wherein operations (a) - (f) facilitate training the neural network with a longer sequence length compared to training of prior neural networks, by dividing the sequence length into smaller blocks to mitigate computational complexity, and subsequently merging the smaller blocks to retain long-range context (Szegedy [11, 38, 41, 42, 46] each sub-sequence output is determined using previous sub-sequence output, and repeated for all sub-sequences, computational complexity is reduced, final output merges outputs for all sub-sequences, retaining long-range attention).
Szegedy does not specifically teach each of the N blocks comprising a sequence length that is shorter than the input sequence length.
However Li teaches each of the N blocks comprising a sequence length that is shorter than the input sequence length; providing the first block output (Xo) as input for determining a next block output (X1) for the next (Xi) of the N blocks, the next block output (Xi ) determined based on the first block output (Xo) (Li Intro-1stBulletPoint, Secs 3 and 3.1, sub-sequences have length (L/N) less than input sequence length (L), each block output can be based on a previous block output, making sub-sequences length shorter than input and based on length limitations, is memory-efficient).
It would have been obvious to one of an ordinary skill in the art before the effective filing date of the claimed invention, to have incorporated the concept taught by Li of each of the N blocks comprising a sequence length that is shorter than the input sequence length; providing the first block output (Xo) as input for determining a next block output (X1) for the next (Xi) of the N blocks, the next block output (Xi ) determined based on the first block output (Xo); into the invention suggested by Szegedy; since both inventions are directed towards determining model output for an input sequence by dividing the sequence into shorter sub-sequences, and incorporating the teaching of Li into the invention suggested by Szegedy would provide the added advantage of being memory efficient, and the combination would perform with a reasonable expectation of success (Li Intro-1stBulletPoint, Secs 3 and 3.1).
Regarding claim 22, Szegedy and Li teach the invention as claimed in claim 21 above.
Szegedy further teaches (g) repeating operations (d) - (f) for each remaining block of the N blocks, such that, for each block (Xi), each next block output (Xi) is output from a previous block (Xi thus merging smaller blocks to retain long-range context associated with the input (Szegedy [11, 33, 38, 41, 42, 46] each sub-sequence output is determined using previous sub-sequence output, and repeated for all sub-sequences, computational complexity is reduced, final output merges outputs for all sub-sequences, retaining long-range attention and context).
Regarding claim 23, Szegedy and Li teach the invention as claimed in claim 21 above.
Szegedy further teaches wherein the model is a neural network (Szegedy [11]).
Szegedy does not specifically teach wherein the neural network comprises at least a portion of a large language model
However Li teaches wherein the ...model... comprises at least a portion of a large language model (Li Intro transformer module may be for large model which can be a language model).
Regarding claim 24, Szegedy and Li teach the invention as claimed in claim 21 above.
Szegedy further teaches wherein the neural network comprises a transformer module (Szegedy [120] neural network may have transformer module).
Regarding claim 25, Szegedy and Li teach the invention as claimed in claim 24 above.
Szegedy further teaches wherein the layer of the neural network comprises a multi-head attention layer of the transformer module (Szegedy [40, 119, 120] layer can be multi-head attention layer of transformer).
Regarding claim 26, Szegedy and Li teach the invention as claimed in claim 25 above.
Szegedy further teaches wherein the multi-head attention layer comprises multiple self- attention modules, each self-attention module associated with one of the N blocks (Szegedy 105, 106, 119] multiple self-head modules may each process one of the sub-sequences).
Regarding claim 27, Szegedy and Li teach the invention as claimed in claim 26 above.
Szegedy further teaches wherein the multiple self-attention modules are configured to operate in parallel (Szegedy [119] multiple self-attention modules can work in parallel).
Regarding claim 28, Szegedy and Li teach the invention as claimed in claim 21 above.
Szegedy further teaches wherein determining the first block output (Xo) for the first (Xo) of the N blocks comprises performing a first Norm + Softmax operation using the initial key (Ko) and query (Qo), and then performing a second Norm + Softmax operation using output from the first Norm + Softmax operation and the initial value (Vo) of the first (Xo) of the N blocks (Szegedy [58-62, 101, 37, 38, 89, 90] output determined for a first sub-sequence based on (K,Q,V) of first sub-sequence, Szegedy [37, 38, 58-62, 101] for each current subsequent sub-sequence, based on (K,Q,V) of current sub-sequence and output of previous sub-sequence, determine current block output, Szegedy [82, 106, 117, 118] each sub-sequence output is performed using Norm+Softmax operation).
Regarding claim 29, Szegedy and Li teach the invention as claimed in claim 21 above.
Szegedy further teaches wherein determining a next block output (i) for a next (Xi) of the N blocks, comprises performing a first Norm + Softmax operation using a key (Ki) for the next block and a query (Qi) for the next block, and output from a previous block (Xii); and then performing a second Norm + Softmax operation using output from the first Norm + Softmax operation, a value (Vi) of the next (Xi) of the N blocks, and the output from the previous block (Xi1) (Szegedy [58-62, 101, 37, 38, 89, 90] output determined for a first sub-sequence based on (K,Q,V) of first sub-sequence, Szegedy [37, 38, 58-62, 101] for each current subsequent sub-sequence, based on (K,Q,V) of current sub-sequence and output of previous sub-sequence, determine current block output, Szegedy [82, 106, 117, 118] each sub-sequence output is performed using Norm+Softmax operation).
Regarding claim 30, Szegedy and Li teach the invention as claimed in claim 21 above.
Szegedy further teaches wherein the N blocks are split into key, query, and/or value using separate linear layers of the neural network (Szegedy [58, 62, 76] sub-sequences are split into key, query, and/or value and supplied to different linear layers).
Regarding claim 31, Szegedy and Li teach the invention as claimed in claim 21 above.
Szegedy does not specifically teach wherein the input comprises the sequence length and an embedding size (C).
However Li teaches wherein the input comprises the sequence length and an embedding size (C) (Li Sec 3 input may have sequence length L and embeddings).
Regarding claim 33, Szegedy and Li teach the invention as claimed in claims 21 above.
Szegedy further teaches wherein a block output coming from previous sequence blocks comprises collective information from all previous sequence blocks (Szegedy [10] each current sub-sequence output is based on previous sub-sequence output, therefore current block uses information based on all previous sub-sequences).
Regarding claim 34, Szegedy and Li teach the invention as claimed in claim 21 above.
Szegedy further teaches a block output is fed back to a self-attention module as a hidden state for a next sequence block (Szegedy [41] sub-sequence output may be an intermediary (hidden) state that is fed as input for next sub-sequence processing).
Szegedy does not specifically teach instead of feeding a sequence of blocks together, to separate self-attention modules, parameters from one self-attention module are shared for all sequence blocks.
However Li teaches instead of feeding a sequence of blocks together, to separate self-attention modules, parameters from one self-attention module are shared for all sequence blocks (Li Sec 3 parameters are same for all self-attention mechanism for all sub-sequences).
Regarding claim 35, Szegedy and Li teach the invention as claimed in claim 21 above.
Szegedy further teaches providing a memory line (Szegedy [54-56, 76] previous sub-sequence outputs are stored and then used for final output).
Regarding claim 36, Szegedy and Li teach the invention as claimed in claim 35 above.
Szegedy further teaches wherein the memory line is configured to ensure sustained influence of one or more initial blocks of the N blocks, and to establish equal importance for each block in a sequence of N blocks (Szegedy [54-56, 76] previous sub-sequence outputs are stored and then used for final output by averaging them).
Regarding claim 37, Szegedy and Li teach the invention as claimed in claim 35 above.
Szegedy further teaches wherein the memory line is configured to assimilate information from a current output through a self-attention module, and wherein the information is appended to the input of subsequent blocks (Xi) (Szegedy [54-56, 76] previous sub-sequence outputs are stored and then used for final output by concatenating them).
Regarding claim 38, Szegedy and Li teach the invention as claimed in claim 35 above.
Szegedy further teaches wherein the memory line is configured to obtain information from a present self-attention module and subsequently determine which information should be transmitted to a next self-attention module (Szegedy [54-56, 62, 76] previous sub-sequence outputs are stored and then used for subsequent sub-sequences, select information may be used for a current sub-sequences based on similarity).
Regarding claim 40, Szegedy and Li teach the invention as claimed in claim 35 above.
Szegedy further teaches wherein a memory line computation is independent of a number of heads in a multi-head attention layer (Szegedy [10, 11, 12] size of the memory can be adjusted to maximize the memory budget).
Claim 41 is directed towards a method performing instructions similar in scope to the instructions stored by the medium of claim 21 and is rejected under the same rationale.
Claim(s) 42-51, 53-58 and 60 is/are dependent on claim 41 above, is/are directed towards a method performing instructions similar in scope to the instructions stored by the medium of claim(s) 22-31, 33-38 and 40 respectively, and is/are rejected under the same rationale.
Claims 32, 52, are rejected under 35 U.S.C. 103 as being unpatentable over Szegedy (US 20240412054 A1) in view of Li et al “Sequence Parallelism: Long Sequence Training from System Perspective”, and further in view of Chuang (US 20250045122 A1) and applicant’s disclosure.
Regarding claim 32, Szegedy and Li teach the invention as claimed in claim 31 above.
Szegedy does not specifically teach wherein operations (a)-(f) reduce K, Q, V multiply and accumulate operations from 3(ξ×C2)+2(ξ2×C) to 3(ξ×C2)+2(ξ2/N×C) and reduce multiply and accumulate operations because of matrix multiplications in “Norm+SoftMax” blocks by a factor of N.
However Chuang teaches reduce K, Q, V multiply and accumulate operations and reduce multiply and accumulate operations because of matrix multiplications in “Norm+SoftMax” blocks (Chuang [6, 33, 34, 35, 54, 67, 78] MAC operations used for Large Language model operations, using multi-head operations to process subsequences of input, using K, Q, V multiply and accumulate (MAC) operations and linear transformation matrix multiplication MAC operations, reduction in MAC operations).
It would have been obvious to one of an ordinary skill in the art before the effective filing date of the claimed invention, to have incorporated the concept taught by Chuang of reduce K, Q, V multiply and accumulate operations and reduce multiply and accumulate operations because of matrix multiplications in “Norm+SoftMax” blocks, into the invention suggested by Szegedy and Li; since both inventions are directed towards using multi-head operations to process subsequences of input, and incorporating the teaching of Chuang into the invention suggested by Szegedy and Li would provide the added advantage of reduction in MAC operations, and the combination would perform with a reasonable expectation of success (Chuang [6, 33, 34, 35, 54, 67, 78]).
Claim 31 discloses performing the operations for K, Q, V and “Norm+SoftMax” blocks.
Applicant’s disclosure in paragraphs [0073-0075] asserts that the operations of claim 31 would result in reduce K, Q, V multiply and accumulate operations from 3(ξ×C2)+2(ξ2×C) to 3(ξ×C2)+2(ξ2/N×C) and reduce multiply and accumulate operations because of matrix multiplications in “Norm+SoftMax” blocks by a factor of N (Applicant’s own disclosure assertion).
Claim(s) 52 is/are dependent on claim 41 above, is/are directed towards a method performing instructions similar in scope to the instructions stored by the medium of claim(s) 32 respectively, and is/are rejected under the same rationale.
Claims 39, 59, are rejected under 35 U.S.C. 103 as being unpatentable over Szegedy (US 20240412054 A1) in view of Li et al “Sequence Parallelism: Long Sequence Training from System Perspective”, and further in view of Deng (US 20250030610 A1).
Regarding claim 39, Szegedy and Li teach the invention as claimed in claim 35 above.
Szegedy does not specifically teach wherein the memory line comprises an adaptation of a Long Short- Term Memory (LSTM) neural network, and/or Gated Recurrent Units (GRUs)
However Deng teaches wherein the memory line comprises an adaptation of a Long Short- Term Memory (LSTM) neural network ... (Deng [15-18, 30, 33, 34] process sub-sequences of a sequence using self-attention modules, memory line may be adapted LSTM layer, improves the accuracy of neural network prediction by combining periodic information, long-term trend information, and short-term time series information).
It would have been obvious to one of an ordinary skill in the art before the effective filing date of the claimed invention, to have incorporated the concept taught by Deng of wherein the memory line comprises an adaptation of a Long Short- Term Memory (LSTM) neural network, into the invention suggested by Szegedy and Li; since both inventions are directed towards processing sub-sequences of a sequence using self-attention modules, and incorporating the teaching of **** into the invention suggested by Szegedy and Li would provide the added advantage of improving the accuracy of neural network prediction by combining periodic information, long-term trend information, and short-term time series information, and the combination would perform with a reasonable expectation of success (Deng [15-18, 30, 33, 34]).
Claim(s) 59 is/are dependent on claim 41 above, is/are directed towards a method performing instructions similar in scope to the instructions stored by the medium of claim(s) **** respectively, and is/are rejected under the same rationale.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SANCHITA ROY whose telephone number is (571)272-5310. The examiner can normally be reached Monday-Friday 12-8.
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, Usmaan Saeed can be reached at (571) 272-4046. 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.
SANCHITA ROY
Primary Examiner
Art Unit 2146
/SANCHITA ROY/Primary Examiner, Art Unit 2146
Prosecution Insights