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 .
Specification
The abstract of the disclosure is objected to because it has more than 150 words. Appropriate correction is required.
Claim Interpretation
Observation: Claim 19 recites distributed system with nodes. The term “node” is interpreted for one with an ordinary skill in the art as a term with known hardware structure. Therefore, this would not invoke 101 rejection software per se.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Regarding claim 21: The instant recites “a non-volatile readable storage medium” the claimed invention is directed to non-statutory subject matter. The claim(s) does/do not fall within at least one of the four categories of patent eligible subject matter because the specification points to no hardware structure about these elements and these are no corresponding structure, paragraph [0334-0335] state that “the non-volatile readable storage medium may include but is not limited to various media that can store a computer program, such as a USB flash disk, a Read-Only Memory (referred to for short as ROM), a Random Access Memory (referred to for short as RAM), a mobile hard disk drive, a diskette and an optical disk.” One with an ordinary skill in the art would interpret this a signal per se with support evidence in the specification. Examiner recommends applicant to recite “non transitory readable storage medium” in order to overcome this specific rejection.
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 1-16, 18-19, 21-22 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as failing to set forth the subject matter which the inventor or a joint inventor, or for applications subject to pre-AIA 35 U.S.C. 112, the applicant regards as the invention.
Regarding Claims 1, 19, 21-22: The instant claims recite “the target controlling node refers to each of the controlling nodes”. The instant claims recite early “a target controlling node”. It is unclear how a target controlling node represents one node, and then it represents each of the controlling nodes. This would render the claims indefinite. For purpose of examination, examiner interprets this term as “target controlling node” represent the controlling node in the distributed system, and controlling node represents a controlling node in one of the node sets.
Regarding Claims 2-16: The dependent claims 2-16 are rejected as they depend on independent claim 1.
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 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.
The factual inquiries set forth in Graham v. John Deere Co., 383 U.S. 1, 148 USPQ 459 (1966), that are applied for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-2, 18-19, 21-22 are rejected under 35 U.S.C. 103 as being un-patentable by Mermoud et al. (“Mermoud”, US 20150332165 A1) hereinafter Mermoud, in view of Mota et al. (“Mota”, US 20150193695 A1) hereinafter Mota.
Regarding claim 1, Mermoud teaches a method of node training for a distributed system, wherein the method is applied in a target controlling node deployed in the distributed system, nodes comprised by the distributed system are grouped into node sets of a quantity M, wherein M is greater than 1 (Fig, 1, Fig, 5, Fig, 6, target controlling node is the centralized controller, Far1, Far2,… are the controlling nodes {root in the local area network} in each of the set of nodes 1, 2, 3,..),
a plurality of nodes that are comprised by each of the node sets are [[sequentially]] connected to obtain a logical chain having a direction, the node sets of the quantity M are [[sequentially]] connected to obtain a logical ring having a direction (Fig, 1, Fig, 5, Fig, 6, target controlling node is the centralized controller, Far1, Far2,… are the controlling nodes {root in the local area network} in each of the set of nodes 1, 2, 3,..),
a last one node of each of the logical chains is a controlling node, the target controlling node refers to each of the controlling nodes (Fig, 1, Fig, 5, Fig, 6, target controlling node is the centralized controller, Far1, Far2,… are the controlling nodes {root in the local area network} in each of the set of nodes 1, 2, 3,..)([0019] The illustrative root node, such as a field area router (FAR) of a FAN, may interconnect the local network with a WAN 130, which may house one or more other relevant devices such as management devices or servers 150)([0059] The edge devices 520 may represent one or more FARs (e.g., FAR 1, FAR 2, and FAR 3) or other root nodes connected to multiple other network-enable devices (e.g., sensors, routers, etc.) in an “edge” network. Any number of edge devices 520 may be connected in the network. As shown in FIG. 1, the edge devices 520 may be interconnected with the devices (e.g., devices 110) by various methods of communication (e.g., links 105), where certain devices 110, such as, e.g., routers, sensors, computers, etc., may be in communication with other devices 110, e.g., based on distance, signal strength, current operational status, location, etc. The illustrative edge devices 520 may interconnect the local network (e.g., LAN, WPAN, etc.) with a private/public network, as described above.), and the method comprises:
receiving a first data packet sent by a first controlling node, wherein the first controlling node refers to a preceding one controlling node of the target controlling node in the logical ring ([0022, 0042-0045] Fig. 2, transmit and/or receive data using a variety of different communication protocols. Note, further, that the nodes may have two different types of network connections 210,transmitting message between different network devices)([0020] Data packets 140 (e.g., traffic and/or messages) may be exchanged among the nodes/devices of the computer network 100);
by using the first data packet and a target data packet, executing a target training task of the target controlling node, to obtain a second data packet, wherein the target data packet carries training data of all of the nodes in a target node set where the target controlling node is located ([0065-0075] Figs, 5-8, As shown in FIG. 7, outcomes of the batch version 610 of the machine learning model and the incremental version 710 of the machine learning model are exchanged, and the batch version 610 and the incremental version 710 of the machine learning model can be cross-validated against each other. By doing so, the reliability and accuracy of predictions made by the machine learning model may be ensured, whereby a countermeasure can be performed should an outcome of the batch version 610 of the machine learning model substantially diverge from the outcomes of the incremental version 710 of the machine learning model (using the same data))([0077-0080] Fig. 10, at step 1020, an incremental version of the machine learning model that operates in incremental mode is pushed to an edge device of the plurality of edge devices, such that the incremental version of the machine learning model is hosted at the edge device. As a result, the batch version and the incremental version of the machine learning model run in parallel with one another)([0064] an incremental version 710 of the machine learning model (e.g., a regression model computed using an incremental algorithm) may be hosted at the edge devices 520. As a result, the batch version 610 of the machine learning model may be hosted at the centralized controller device 510 simultaneously with the incremental version 710 of the machine learning model being hosted at the edge devices 520. Thus, the batch version 610 of the machine learning model and the incremental version 710 of the machine learning model may run in parallel with one another in a distributed manner, thereby improving the convergence properties and overall reliability of these algorithms, as described further below. Indeed, this architecture intelligently distributes the computational load across the network, while taking into account the computational constraints and access to “fresh” data of each device in the network participating in the learning); and
sending the second data packet to a second controlling node, till the target controlling node has received training data of all of the controlling nodes of the distributed system, wherein the second controlling node refers to a next one controlling node of the target controlling node in the logical ring, and the training data of all of the controlling nodes of the distributed system and the target data packet are dispatched to the other nodes than the target controlling node in the target node set ([0065-0075] Figs, 5-8, As shown in FIG. 7, outcomes of the batch version 610 of the machine learning model and the incremental version 710 of the machine learning model are exchanged, and the batch version 610 and the incremental version 710 of the machine learning model can be cross-validated against each other. By doing so, the reliability and accuracy of predictions made by the machine learning model may be ensured, whereby a countermeasure can be performed should an outcome of the batch version 610 of the machine learning model substantially diverge from the outcomes of the incremental version 710 of the machine learning model (using the same data))([0077-0080] Fig. 10, at step 1020, an incremental version of the machine learning model that operates in incremental mode is pushed to an edge device of the plurality of edge devices, such that the incremental version of the machine learning model is hosted at the edge device. As a result, the batch version and the incremental version of the machine learning model run in parallel with one another)([0064] an incremental version 710 of the machine learning model (e.g., a regression model computed using an incremental algorithm) may be hosted at the edge devices 520. As a result, the batch version 610 of the machine learning model may be hosted at the centralized controller device 510 simultaneously with the incremental version 710 of the machine learning model being hosted at the edge devices 520. Thus, the batch version 610 of the machine learning model and the incremental version 710 of the machine learning model may run in parallel with one another in a distributed manner, thereby improving the convergence properties and overall reliability of these algorithms, as described further below. Indeed, this architecture intelligently distributes the computational load across the network, while taking into account the computational constraints and access to “fresh” data of each device in the network participating in the learning)([0025-0030] Fig. 2, nodes in each set has learning machine process element 248 performing a task).
Mermoud does not explicitly teach the sequentially connection between set of nodes, but Mota teaches
Sets of nodes are connected sequentially ([0129] FIG. 6C, FAR-1 may send the new ANN parameters to FAR-2 after FAR-1 trains the model using at least a portion of its own local dataset)([0130] FIG. 6D, FAR-N may send the final, trained model parameters back to the NMS.)
It would have been obvious to a person skilled in the art, before the effective filing date of the invention, to modify Mermoud in view of Mota in order to have client devices and sets of network nodes connected sequentially and consider different training sets collected by different training entities because it would improve significantly the quality of the distributed machine learning model and allow for potentially reducing the amount of traffic to be collected in each individual mesh (Mota, [0073]).
Regarding claim 2, Mermoud and Mota teaches the method according to claim 1,
Mermoud teaches wherein before the step of, by using the first data packet and the target data packet, executing the target training task of the target controlling node ([0065-0075] Figs, 5-8, As shown in FIG. 7, outcomes of the batch version 610 of the machine learning model and the incremental version 710 of the machine learning model are exchanged, and the batch version 610 and the incremental version 710 of the machine learning model can be cross-validated against each other. By doing so, the reliability and accuracy of predictions made by the machine learning model may be ensured, whereby a countermeasure can be performed should an outcome of the batch version 610 of the machine learning model substantially diverge from the outcomes of the incremental version 710 of the machine learning model (using the same data))([0077-0080] Fig. 10, at step 1020, an incremental version of the machine learning model that operates in incremental mode is pushed to an edge device of the plurality of edge devices, such that the incremental version of the machine learning model is hosted at the edge device. As a result, the batch version and the incremental version of the machine learning model run in parallel with one another)([0064] an incremental version 710 of the machine learning model (e.g., a regression model computed using an incremental algorithm) may be hosted at the edge devices 520. As a result, the batch version 610 of the machine learning model may be hosted at the centralized controller device 510 simultaneously with the incremental version 710 of the machine learning model being hosted at the edge devices 520. Thus, the batch version 610 of the machine learning model and the incremental version 710 of the machine learning model may run in parallel with one another in a distributed manner, thereby improving the convergence properties and overall reliability of these algorithms, as described further below. Indeed, this architecture intelligently distributes the computational load across the network, while taking into account the computational constraints and access to “fresh” data of each device in the network participating in the learning)([0025-0030] Fig. 2, nodes in each set has learning machine process element 248 performing a task), the method further comprises:
Mermoud does not explicitly teach, but Mota teaches
receiving a reference data packet sent by a preceding one node of the target controlling node in the target node set, wherein the reference data packet refers to a data packet that is formed by transmitting and operating training datum of each of the nodes in the target node set sequentially in a target logical chain of the target node set and is formed in the preceding one node (Fig. 6A, Fig. 6B, sending training data and model parameters from the controller to the FAR-1, FAR2,…) ([0129] FIG. 6C, FAR-1 may send the new ANN parameters to FAR-2 after FAR-1 trains the model using at least a portion of its own local dataset)([0130] FIG. 6D, FAR-N may send the final, trained model parameters back to the NMS)([0131] Fig. 7)([0132-0135] Fig. 8, step 820, the newly generated model parameters are then sent to another training device and procedure 800 ends at step 825.); and
by using the reference data packet and a target training datum of the target controlling node, executing the target training task, to obtain the target data packet (Fig. 6A, Fig. 6B, sending training data and model parameters from the controller to the FAR-1, FAR2,…) ([0129] FIG. 6C, FAR-1 may send the new ANN parameters to FAR-2 after FAR-1 trains the model using at least a portion of its own local dataset)([0130] FIG. 6D, FAR-N may send the final, trained model parameters back to the NMS)([0131] Fig. 7)([0132-0135] Fig. 8, step 820, the newly generated model parameters are then sent to another training device and procedure 800 ends at step 825.).
It would have been obvious to a person skilled in the art, before the effective filing date of the invention, to modify Mermoud in view of Mota in order to have client devices and sets of network nodes connected sequentially and consider different training sets collected by different training entities because it would improve significantly the quality of the distributed machine learning model and allow for potentially reducing the amount of traffic to be collected in each individual mesh (Mota, [0073]).
Regarding claim 18, Mermoud and Mota teaches the method according to claim 1,
Mermoud teaches wherein before the step of receiving the first data packet sent by the first controlling node, the method further comprises:
according to a quantity of the nodes comprised by the distributed system, determining M and quantities of the nodes comprised by the node sets ([0031] specifying a set of link and node metrics)([0040-0041] defines a host or set of hosts, such as a host serving as a data collection point)([0020] set of rules defining how the nodes interact with each other)([0059] identifying number of edge devices connected in the network, and number FAR);
according to M and the quantities of the nodes comprised by the node sets, dividing the nodes comprised by the distributed system, to obtain the node sets of the quantity M ([0031] specifying a set of link and node metrics)([0040-0041] defines a host or set of hosts, such as a host serving as a data collection point)([0020] set of rules defining how the nodes interact with each other)([0059] identifying number of edge devices connected in the network, and number FAR); and
building links between the node sets of the quantity M, to obtain the logical ring ([0031] specifying a set of link and node metrics)([0040-0041] defines a host or set of hosts, such as a host serving as a data collection point)([0020] set of rules defining how the nodes interact with each other)([0059] identifying number of edge devices connected in the network, and number FAR).
Regarding claim 19, claim 19 is rejected with the same reasoning as claim 1.
Regarding claim 21, claim 21 is rejected with the same reasoning as claim 1.
Regarding claim 22, claim 22 is rejected with the same reasoning as claim 1.
Allowable Subject Matter
Claims 3-6, 7-12, 13-16 are objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims and overcome the 112(b) rejections set forth in this office action.
The following is the reason for allowable subject matter in claim 3.
The prior arts (“Mermoud”, US 20150332165 A1) and (“Mota”, US 20150193695 A1) fail to teach or suggest fairly detecting operation states of all of the nodes in the target node set; and in response to a node the operation state of which is an abnormal operation state existing in the target node set, sending first indicating information to a third controlling node in the distributed system, wherein the first indicating information is for indicating that the target node set is in the abnormal operation state, and the third controlling node refers to a controlling node other than the target controlling node in the distributed system.
The following is the reason for allowable subject matter in claim 7.
The prior arts (“Mermoud”, US 20150332165 A1) and (“Mota”, US 20150193695 A1) fail to teach or suggest fairly receiving second indicating information sent by a fourth controlling node in the distributed system, wherein the second indicating information is for indicating that the node set where the fourth controlling node is located is in an abnormal operation state; and in response to the fourth controlling node comprising the second controlling node, building a link with a first instance of the nodes of the logical chain that has a lowest distance to the target controlling node in the logical ring and is in a normal operation state, to obtain a logical ring that has been updated, wherein the logical chains of the quantity M are sequentially connected to obtain the logical ring having the direction.
The following is the reason for allowable subject matter in claim 13.
The prior arts (“Mermoud”, US 20150332165 A1) and (“Mota”, US 20150193695 A1) fail to teach or suggest fairly by using an internal-memory protocol of a calculation quick-connection protocol, storing the second data packet into a server main-memory of the distributed system, to obtain a data copy; performing jumbo-frame packaging processing to the second data packet, to obtain a target jumbo frame; and transmitting the target jumbo frame to the second controlling node.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to FADI HAJ SAID whose telephone number is (571)272-2833. The examiner can normally be reached on 8:00 AM - 5:00 PM EST. 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, John Follansbee can be reached on 571-272-3964. 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 the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see https://ppair-my.uspto.gov/pair/PrivatePair. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/FADI HAJ SAID/Primary Examiner, Art Unit 2444
