DETAILED ACTION
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 1-6, 13, 17-21, 23 have been amended. Claims 7, 16 and 22 have been canceled. Claims 24-28 have been added. Claims 1-6, 8-15, 17-21 and 23-28 have been examined.
Response to Arguments
Applicant's arguments filed 7/2/2026 have been fully considered but they are not persuasive.
On p. 10 of the 7/2/2026 remarks, Applicant argues with respect to cited art of record Heidarpur, that “None of the designs selects between time steps at runtime in response to an input current.” In response to applicant's argument that the references fail to show certain features of the invention, it is noted that the features upon which applicant relies (i.e., “selects between time steps at runtime”) are not recited in the rejected claim(s). Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims. See In re Van Geuns, 988 F.2d 1181, 26 USPQ2d 1057 (Fed. Cir. 1993).
Applicant’s remaining arguments, see p. 10, filed 7/2/2026, with respect to the rejection(s) of claim(s) 13 under 35 USC 102(a)(1) have been fully considered and are persuasive. Therefore, the rejection has been withdrawn. However, upon further consideration, a new ground(s) of rejection is made in view of U.S. Patent Application Publication 20140101420 by Wu and U.S. Patent Application Publication 20020038294 by Matsugu.
On p. 10-11 of the remarks, Applicant argues that cited art of record Wu fails to teach the limitations of claim 1 because Wu’s input “is not a neural input current … does not have a neuron, nor spikes, and it does not have a membrane potential.” In response to applicant's arguments against the references individually, one cannot show nonobviousness by attacking references individually where the rejections are based on combinations of references. See In re Keller, 642 F.2d 413, 208 USPQ 871 (CCPA 1981); In re Merck & Co., 800 F.2d 1091, 231 USPQ 375 (Fed. Cir. 1986). In this case, Wu is not relied upon to teach elements of neuromorphic hardware. Instead, the rejection relies upon Heidarpur to teach limitations related to an artificial neural network. The rejection is not based upon Wu alone, but instead relies upon the teaching of Heidarpur in combination with Wu. Therefore, this argument is not persuasive.
At the bottom of p. 10 to the top of p. 11 of the remarks, Applicant argues that Heidarpur and Wu do not teach or suggest “a timer …” However, Wu teaches a timer as cited in the rejection below. This argument is not persuasive.
Applicant’s arguments, see the bottom of p. 23, filed 7/2/2026, with respect to claim 23, have been fully considered and are persuasive. The rejection of claim 23 has been withdrawn.
At the top of p. 12 filed 7/2/2026, Applicant suggests that claim 23 has been amended to avoid interpretation under 35 USC 112(f). It is noted that the current claims include multiple limitations that could be subject to interpretation under 35 USC 112(f) (e.g. “a comparator configured to compare …”). However, one of ordinary skill would understand these claimed limitations in view of Applicant’s disclosure as referring to hardware elements and therefore will not be interpreted under 35 USC 112(f).
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.
Claims 1-6, 8-9, 12-13, 17-21 and 24-25 are rejected under 35 U.S.C. 103 as being unpatentable over "Time Step Impact on Performance and Accuracy of Izhikevich Neuron: Software Simulation and Hardware Implementation" by Heidarpur et al. (“Heidarpur”) in view of U.S. Patent Application Publication 20140101420 by Wu ("Wu") and U.S. Patent Application Publication 20020038294 by Matsugu (“Matsugu”).
In regard to claim 1, Heidarpur discloses:
1. A method for operating an artificial neural network, the method comprising: See p. 4, section IV, “For instance, in a network by knowing the number of the inputs and the maximum weight of each one, the maximum input to the neuron and accordingly the appropriate time step could be determined using Fig. 3.”
receiving an input signal; [analyzing] … a sampling frequency of an artificial neuron based on the input signal; See Heidarpur, bottom right on p. 2, “By increasing the input current, the spike frequency of the neuron increases. Therefore, for a specific time step (sampling time), higher input currents result in a larger error. In addition, neurons simulated with higher input current tend to become unstable for a smaller time steps compared to those with lower currents.” Also see bottom right on p. 4, “the maximum input to the neuron and accordingly the appropriate time step could be determined using Fig. 3.”
While Heidarpur discloses analysis of various sample frequencies, Heidarpur does not expressly disclose: modulating. This is taught by Wu. See ¶ 0026-0027, “… determines a supply voltage and a clock rate (e.g. frequency) of the processor being controlled (according to lookup table 116, in an embodiment). … During runtime, an adaptive adjustment module 120 adaptively adjusts a sampling frequency and gain of the control loop (e.g., controller 108, DVFS HW/SW 106, and tracker 110) coupled to processor 104 in response to characteristics of the input workload. … According to an embodiment, lookup table 116 may include a plurality of entries of the form <C, S, f, V>, where f and V are desired frequency and supply voltage values.” It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use Wu’s sample rate modulation with Heidarpur’s analysis in order to achieve a performance-workload match even when workload characteristics change substantially as suggested by Wu (see ¶ 0022).
starting a timer; Heidarpur, section II-A, bottom left column on p. 2, e.g. “First, Izhikevich neuron was simulated with small time step …” Note that implementation using a particular “time step” requires the use of some form of timer. However, even if Heidarpur failed to fully disclose a timer, this is taught by Wu. See Wu ¶ 0061, “The monitoring and/or adaptive adjustment may be initiated by a trigger based upon, for example, a configured duration (e.g., periodically triggered by a timer).” It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use Wu’s timer with Heidarpur’s time step in order to utilize a configurable trigger as suggested by Wu (see ¶ 0061).
Heidarpur also discloses:
forwarding the input signal or a further input signal obtained from the input signal to the artificial neuron at the sampling frequency … See Heidarpur, Fig. 1 depicts various input signals that are provided to a neuron at a variety of sample frequencies. Also see section II(A), bottom left column on p. 3, “A rectangular current pulse with maximum level of 4 was applied to the neurons as depicted in this figure with a red line. First, Izhikevich neuron was simulated with small time step of 0.001, and subsequently, the simulation repeated with larger time steps.”
… when the timer reaches a value determined by the sampling frequency, … Note that Wu teaches use of a timer as cited above.
… for a time period selected for the artificial neuron to perform a single evaluation, Heidarpur section 1, p. 1, bottom left to bottom right column, e.g. “Euler’s method [17] is one of the most common methods used in previous works to numerically solve Ordinary Differential Equations (ODEs) describing neurons. A factor that has major impact on speed and computational cost of this method is time step. Smaller time steps preserve accuracy with the cost of higher computational cost. … Section II reviews Izhikevich neuron and presents the accuracy analysis for different time steps.”
Heidarpur also discloses:
wherein said modulating the sampling frequency comprises one or both of increasing the sampling frequency with an increased input signal and reducing the sampling frequency with a decreased input signal, and … See Heidarpur, bottom right on p. 2 and bottom right on p. 4 as cited above.
Heidarpur does not expressly disclose: … wherein the artificial neuron is disabled outside the time period selected for the artificial neuron to perform the single evaluation. However, this is taught by Matsugu. See Matsugu ¶ 0185, “The refractory period shown in FIG. 3B is a time period from the point immediately following the firing of a feature detection cell to the next time window starting time. obviously, the duration of the refractory period and the width of the time window can be set to arbitrary values.” It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use Matsugu’s refractory period with Heidarpur’s neuron in order to faithfully model a neuron as suggested by Matsugu (see at least ¶ 0007).
In regard to claim 2, Heidarpur also discloses:
2. The method of claim 1, wherein said modulating the sampling frequency comprises modulating the sampling frequency based on an amplitude of the input signal and one or more reference sampling frequencies, wherein the reference sampling frequencies are determined at a plurality of preselected input signal amplitudes, whereby an output or output signal of the artificial neuron from each said preselected input signal amplitude is stable at or above the associated reference sampling frequency. Heidarpur Fig. 3, depicting preselected input amplitudes and associated time steps/frequencies associated with stability thresholds. Also see p. 2, bottom right column, “The threshold time step that triggers dumped oscillation and the threshold that neuron becomes unstable are plotted against the input current in Fig. 3.” Also p. 4, lower right column, “… the maximum input to the neuron and accordingly the appropriate time step could be determined using Fig. 3.” Also see p. 2, lower right column, “A function approximation to the measured data in Fig. 3 (a) is: tsth = 5.76 I-0.22 − 1.26 (4) Where tsth is the stability threshold and I is the input current.” Also see Wu, ¶ 0026-0027, as cited above, e.g. “According to an embodiment, lookup table 116 may include a plurality of entries of the form <C, S, f, V>, where f and V are desired frequency and supply voltage values.”
In regard to claim 3, Heidarpur also discloses:
3. The method of claim 1, wherein said modulating the sampling frequency comprises selecting one of two or more predefined sampling frequencies based on an amplitude of the input signal, each said predefined sampling frequency being selected for a range of the amplitude of the input signal, whereby an output or output signal of the artificial neuron from the amplitude of the input signal in at least a portion of the range is stable at the associated predefined sampling frequency. See Heidarpur Fig. 3 and pp. 2 and 4 as cited above with respect to claim 2.
In regard to claim 4, Heidarpur also discloses:
4. The method of claim 2, wherein said modulating the sampling frequency comprises selecting one of two or more predefined sampling frequencies obtained from the reference sampling frequencies, each said predefined sampling frequency being selected for a range of the amplitude of the input signal, whereby the output or output signal of the artificial neuron from the amplitude of the input signal in at least a portion of the range is stable at the associated predefined sampling frequency. See Heidarpur Fig. 3 and pp. 2 and 4 as cited above with respect to claim 2.
In regard to claim 5, Heidarpur also discloses:
5. The method of claim 4, wherein each said predefined sampling frequency is equal to at least one said reference sampling frequency determined at the associated preselected input signal amplitude falling within the associated range. Heidarpur Fig. 3 and pp. 2 and 4 as cited above. Also p. 2, bottom right column, “Where tsth is the stability threshold.” Also see section IV, p. 4, “For instance, in a network by knowing the number of the inputs and the maximum weight of each one, the maximum input to the neuron and accordingly the appropriate time step could be determined using Fig. 3.”
In regard to claim 6, Heidarpur also discloses:
6. The method of claim 3, wherein the output or output signal of the artificial neuron from the amplitude of the input signal in the range is stable at or about the associated predefined sampling frequency. Heidarpur section V, p. 5, top left column, “In this work, the relation between the threshold time step that neuron produces stable output and its input current was uncovered.” Also section IV, p. 4, “the appropriate time step could be determined using Fig. 3.”
In regard to claim 8, Heidarpur discloses:
8. The method of claim 1, wherein the input signal is an input current. See Heidarpur, bottom right on p. 2, “By increasing the input current, …”
In regard to claim 9, Heidarpur discloses:
9. The method of claim 1, wherein the artificial neural network is a spiking neural network. See Heidarpur, section I, “Spiking neural networks.”
In regard to claim 12, Heidarpur discloses:
12. The method of claim 1, wherein the method is operable to reduce power consumption in the artificial neural network. Heidarpur, section 1, p. 1, right column, “Results indicated that larger time steps than those that used in the previous works could be employed which significantly improves performance and cost of the design.” Also see p. 4, top right column, “smaller time steps consume marginally less power. However, energy consumption is a function of the total time taken to complete a task. Table IV presents the number of clock cycles required for each design to generate 5 spikes. According to this table, generating 5 spikes with a time step of 1/1024 consumes 64 x (90=85) = 67:8 times more power compared to those with time step of 1/16.”
In regard to claim 13, Heidarpur discloses:
13. A method for operating a spiking neural network, the method comprising: See p. 4, section IV, “For instance, in a network by knowing the number of the inputs and the maximum weight of each one, the maximum input to the neuron and accordingly the appropriate time step could be determined using Fig. 3.”
receiving an input current; determining a time step based on the input current by comparing the input current to at least one threshold value; See Heidarpur, bottom right on p. 2, “By increasing the input current, the spike frequency of the neuron increases. Therefore, for a specific time step (sampling time), higher input currents result in a larger error. In addition, neurons simulated with higher input current tend to become unstable for a smaller time steps compared to those with lower currents.” Also see bottom right on p. 4, “the maximum input to the neuron and accordingly the appropriate time step could be determined using Fig. 3.” Heidarpur Fig. 3 provides a graph for determining a time step by comparing input currents to threshold time steps.
starting a timer; and Heidarpur, section II-A, bottom left column on p. 2, e.g. “First, Izhikevich neuron was simulated with small time step …” Note that implementation using a particular “time step” requires the use of some form of timer. However, even if Heidarpur failed to fully disclose a timer, this is taught by Wu. See Wu ¶ 0061, “The monitoring and/or adaptive adjustment may be initiated by a trigger based upon, for example, a configured duration (e.g., periodically triggered by a timer).” It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use Wu’s timer with Heidarpur’s time step in order to utilize a configurable trigger as suggested by Wu (see ¶ 0061).
forwarding the input current to an artificial neuron … See Heidarpur, Fig. 1 depicts various input signals that are provided to a neuron at a variety of sample frequencies. Also see section II(A), bottom left column on p. 3, “A rectangular current pulse with maximum level of 4 was applied to the neurons as depicted in this figure with a red line. First, Izhikevich neuron was simulated with small time step of 0.001, and subsequently, the simulation repeated with larger time steps.”
… when the timer reaches the time step … Note that Wu teaches use of a timer as cited above.
… for a time period selected for the artificial neuron to perform a single evaluation, and Heidarpur section 1, p. 1, bottom left to bottom right column, e.g. “Euler’s method [17] is one of the most common methods used in previous works to numerically solve Ordinary Differential Equations (ODEs) describing neurons. A factor that has major impact on speed and computational cost of this method is time step. Smaller time steps preserve accuracy with the cost of higher computational cost. … Section II reviews Izhikevich neuron and presents the accuracy analysis for different time steps.”
Heidarpur does not expressly disclose: disabling the artificial neuron after the single evaluation until the timer is restarted. However, this is taught by Matsugu. See Matsugu ¶ 0185, “The refractory period shown in FIG. 3B is a time period from the point immediately following the firing of a feature detection cell to the next time window starting time. obviously, the duration of the refractory period and the width of the time window can be set to arbitrary values.” It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use Matsugu’s refractory period with Heidarpur’s neuron in order to faithfully model a neuron as suggested by Matsugu (see at least ¶ 0007).
In regard to claim 17, Heidarpur also discloses:
17. The method of claim 13, wherein said determining the time step comprises determining the time step based on the input current and one or more reference time steps, wherein the reference time steps are determined at a plurality of preselected input currents, whereby an output or output signal of the artificial neuron from each said preselected input current is stable at or lower than the associated time step. Heidarpur Fig. 3, depicting preselected input amplitudes and associated time steps/frequencies associated with stability thresholds. Also see p. 2, bottom right column, “The threshold time step that triggers dumped oscillation and the threshold that neuron becomes unstable are plotted against the input current in Fig. 3.” Also p. 4, lower right column, “… the maximum input to the neuron and accordingly the appropriate time step could be determined using Fig. 3.” Also see p. 2, lower right column, “A function approximation to the measured data in Fig. 3 (a) is: tsth = 5.76 I-0.22 − 1.26 (4) Where tsth is the stability threshold and I is the input current.”
In regard to claim 18, Heidarpur also discloses:
18. The method of claim 13, wherein said determining the time step comprises selecting one of two or more predefined time steps based on the input current, each said predefined time step being selected for a range of the input current, whereby the output or output signal of the artificial neuron from the input current in at least a portion of the range is stable at or about the associated predefined time step. See Heidarpur Fig. 3 and pp. 2 and 4 as cited above with respect to claim 17.
In regard to claim 19, Heidarpur also discloses:
19. The method of claim 17, wherein said determining the time step comprises selecting one of two or more predefined time steps obtained from the reference time steps, each said predefined time step being selected for a range of the input current, whereby the output or output signal of the artificial neuron from the input current in at least a portion of the range is stable at the associated predefined time step. See Heidarpur Fig. 3 and pp. 2 and 4 as cited above with respect to claim 2.
In regard to claim 20, Heidarpur also discloses:
20. The method of claim 19, wherein each said predefined time step is equal to at least one said reference time step determined at the associated preselected input current falling within the associated range. Heidarpur Fig. 3 and pp. 2 and 4 as cited above. Also p. 2, bottom right column, “Where tsth is the stability threshold.” Also see section IV, p. 4, “For instance, in a network by knowing the number of the inputs and the maximum weight of each one, the maximum input to the neuron and accordingly the appropriate time step could be determined using Fig. 3.”
In regard to claim 21, Heidarpur also discloses:
21. The method of claim 19, wherein the output or output signal of the artificial neuron from the input current in the range is stable at or about the associated predefined time step. Heidarpur section V, p. 5, top left column, “In this work, the relation between the threshold time step that neuron produces stable output and its input current was uncovered.” Also section IV, p. 4, “the appropriate time step could be determined using Fig. 3.”
In regard to claim 24, Heidarpur and Matsugu also teach:
24. The method of claim 13, wherein said disabling the artificial neuron comprises one or both of disabling said forwarding the input current and disabling the timer from counting or forwarding a clock. See Matsugu, ¶ 0125, “The start of a time window provides the reference for the phase of a neuron output …” Also ¶ 0184, “More specifically, there is a similarity to the actual neuron in that no output is provided from a neuron during the refractory period, i.e., the time range other than the time window, whatever input it may receive, whereas firing based on an input level takes place in the time window other than the refractory period.”
In regard to claim 25, Heidarpur discloses:
25. The method of claim 13, wherein said determining the time step comprises selecting one of three predefined time steps by comparing the input current to two said at least one threshold values. See Heidarpur, Fig. 3(a), depicting a threshold graph providing threshold current values of e.g. 40 and 60 for associating input currents <40 with e.g. time step 1.5, currents 40-60 with e.g. time step 1, and current >60 with e.g. time step 0.5. Alternate values are depicted at any point along the curve.
Claims 10-11 are rejected under 35 U.S.C. 103 as being unpatentable over Heidarpur, Wu and Matsugu as applied above, and further in view of “A Survey of Neuromorphic Computing and Neural Networks in Hardware” by Schuman et al. (“Schuman”) and "Spike-based strategies for rapid processing" by Thorpe et al. (“Thorpe”).
In regard to claim 10, Heidarpur discloses:
10. The method of claim 1, wherein the artificial neural network comprises a digital neuron model selected from the group consisting of … Izhikevich model, See Heidarpur, p. 1, right column, “… Izhikevich model.”
Heidarpur does not expressly disclose the remaining models.
Schuman teaches: Adaptive-Exponential Integrate-and-Fire model, Hodgkin-Huxley model, Leaky Integrate-and-Fire model, … probabilistic and stochastic spiking neuron model and probabilistic neurogenetic model. See Schuman, p. 5, section III(A)(1), “Hodgkin-Huxley.” Also p. 6, section III(A)(4), “leaky integrate-and-fire … adaptive exponential integrate-and-fire.” Also section III(C), p. 7, “spiking probabilistic networks.” Also section III (C), p. 8, “Probabilistic neural networks … Stochastic neural networks.” Also section IV(A), p. 10, “Off-chip nature based implementations include differential evolution [1321]-[1324], evolutionary or genetic algorithms.” In practicality there are only a finite number of neural models. It would have been obvious to one of ordinary skill in the art to try the various models of Schuman in the neural network of Heidarpur in an attempt to provide better neural computation, as a person with ordinary skill has a good reason to pursue the known replacement algorithms within his or her technical grasp. In turn, because neural models when used in the system of Heidarpur has the predicted properties of the neural behavior, it would have been obvious.
Schuman does not expressly disclose: Thorpe's model. This is taught by Thorpe. See Thorpe, p. 715, right column, “… a coding scheme that encodes information in the relative timing of spikes across a population of neurons, or more specifically, in the order in which neurons fire.” It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use Thorpe’s model with Heidarpur’s spiking neural network in order to utilize efficient computation as suggested by Thorpe (see section 1, p. 715, right column).
In regard to claim 11, Heidarpur discloses:
11. The method of claim 10, wherein the digital neuron model is Adaptive-Exponential Integrate-and-Fire model or Izhikevich model, and said modulating the sampling frequency comprises selecting the sampling frequency between dt= 1/4096 ms and dt=8 ms. See Heidarpur, p. 1, right column, “… Izhikevich model.” Also see Fig. 1 depicting time steps between 0.001 and 3 ms.
Claims 14-15 are rejected under 35 U.S.C. 103 as being unpatentable over Heidarpur, Wu and Matsugu as applied above, and further in view of Schuman and Thorpe.
In regard to claims 14-15, parent claim 13 is addressed above.
All further limitations of claims 14-15 have been addressed in the above rejection of claims 10-11, respectively.
Allowable Subject Matter
Claims 23 and 26-28 are allowed.
The following is a statement of reasons for the indication of allowable subject matter:
As cited with respect to claims 1-6, 8-15, 17-21 and 24-25 above, cited art of record Heidarpur provides a detailed analysis of variable time steps with respect to implementation of a digital neuron. However, the cited prior art taken alone or in combination fail to teach, in combination with the other claimed limitations, a digital neuron module and an input-dependent variable sampling module coupled to the digital neuron module, in the manner provided in claim 23 including arrangement of a comparator and a timer coupled to the comparator and to a clock input, as well as a forwarding circuit coupled to the enable signal, wherein the digital neuron module is configured to perform a single evaluation while the enable signal is asserted and to output a sampled membrane potential. These limitations are present in independent claim 23. The distinctions provided by the independent claims apply equally to all dependent claims. Thus claims 23 and 26-28 are allowed.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
U.S. Patent 6492929 to Coffey et al. See col. 4 lines 56-63, “In this embodiment, the comparison means may comprise two comparators, each receiving the input signal and a different one of two threshold levels, UPPER or LOWER, forming an amplitude window about the current input signal, whereby if the input signal crosses one of the threshold levels the corresponding comparator generates a primary digital output signal to a subsequent logic stage for generating the UP/DOWN digital output signal.”
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). 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 James D Rutten whose telephone number is (571)272-3703. The examiner can normally be reached M-F 9:00-5:30 ET.
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, Li B Zhen can be reached at (571)272-3768. 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.
/James D. Rutten/Primary Examiner, Art Unit 2121