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 .
DETAILED ACTION
Claims 1-5 are currently pending and have been examined.
Claim Objections
Claim 1 and 2 are objected to because of the following informalities:
Re-claims 1 and 2, the steps of the method recite labels (e.g. Step S1, Step S2 … Step S12). The labels of the method steps are not necessary and should be removed. Appropriate correction is required.
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.
Claims 1-5 are rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter.
Step 1: Claims 1-5 are directed to a deep learning-based method; thus, this claim is directed to a process, which is one of the statutory categories of invention.
In order to evaluate the Step 2A inquiry “Is the claim directed to a law of nature, a natural phenomenon or an abstract idea?” we must determine, at Step 2A Prong 1, whether the claim recites a law of nature, a natural phenomenon or an abstract idea and further whether the claim recites additional elements that integrate the judicial exception into a practical application.
Step 2A Prong 1:
Regarding clam 1, the claim recites limitations of: “on the basis of a raw data set, predicting a future workload of a central processing unit by using a deep learning-based prediction algorithm for cloud workloads (L-PAW) integrating a top-sparse auto-encoder (TSA) and a gated recurrent unit (GRU)” and “determining, by the CSP, a resource allocation strategy according to the predicted result, such that the cloud data center achieves load balancing”, which as drafted, are mathematical calculations. For example “predicting” in the context of the claims is performed by using a deep learning-based prediction algorithm to predict a future workload, and therefore encompasses mathematical concepts. Further, for example, “determining” in the context of the claim maybe performed by an algorithm and therefore a encompasses mathematical concepts. Additionally, limitations “determining … a resource allocation strategy …” as drafted covers performance in the mind. For example, “performing”, in the context of the claim encompasses a person observing and evaluating the predicted results and determining a strategy for allotting resources based on the predicted results.
Therefore, Yes, claim 1 recites a judicial exception. The claims have been identified to recite judicial exceptions, Step 2A Prong 2 will evaluate whether the claims are directed to the judicial exception.
Step 2A Prong 2:
Claim 1: The judicial exception is not integrated into a practical application.
The claims recite the additional elements –“ obtaining historical workload data of a cloud data center, and carrying out preprocessing” which is merely a recitation of extra-solution data gathering activity (see MPEP § 2106.05(g)) which does not integrate a judicial exception into practical application. Further, recites the additional elements “transmitting a predicted result to a cloud service provider (CSP)”, which is merely a recitation of extra-solution data output activity.
After having evaluating the inquires set forth in Steps 2A Prong 1 and 2, it has been concluded that the claim 1 not only recites a judicial exception but that the claim is directed to the judicial exception as the judicial exception has not been integrated into practical application.
Step 2B:
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements amount to no more than insignificant extra-solution activity without imposing meaningful limits on practicing the abstract idea and thus cannot provide an inventive concept. Further, the identified insignificant extra-solution activity comprises data gathering activity and data output activity, which is further Well-Understood, Routine and Conventional, see MPEP § 2106.05(d)(II) “The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) … Receiving or transmitting data over a network … which are data input and output similar to the instant claims which have been identified to be Well-Understood, Routine and Conventional.
Therefore, “Do the claims recite additional elements that amount to significantly more than the judicial exception? No, these additional elements, alone or in combination, do not amount to significantly more than the judicial exception. Having concluded analysis within the provided framework, Claim 1 does not recite patent eligible subject matter under 35 U.S.C. § 101.
With regard to claim 2, further recites additional elements of “obtaining historical workload data … extracting CPU utilization as raw workload data and normalizing …”, which merely further provide further details of the extra solution activity of gathering data, which does not integrate the judicial exception into practical application nor provide significantly more.
With regard to claim 3, further recites the additional elements “replacing a hidden layer of the RNN … setting a learning rate decay to control a learning rate y in stages … which merely further describes in more detail the abstract idea of predicting a future workload by specifying that the integrating further comprises replacing a hidden layer of the RNN with a GRU and setting a learning rate decay using an algorithm, which are mathematical calculation, and which does not integrate the judicial exception into practical application nor provide significantly more.
With regard to claim 4, further recites the additional elements “an input of the TSA being a vector X = (x1, x2 … xn) … during forward propagation, an average activation degree b of hidden units being computed as … next all the hidden units being sorted … computing a cost function … compressing workload data … executing backpropagation of the cost function …” which merely further describes in more detail the abstract idea of predicting a future workload using an algorithm which are mathematical calculations, and which does not integrate the judicial exception into practical application nor provide significantly more.
With regard to claim 4, further recites the additional elements “wherein the k hidden units with the highest activation degree are selected to reconstruct input data” which mere further describes the abstract idea by specifying selecting a hidden value k having highest activation degree, which does not integrate the judicial exception into practical application nor provide significantly more.
Claim Rejections - 35 USC § 102
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)(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.
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claims 1-5 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Chen et al. “Towards Accurate Prediction for High-Dimensional and Highly-Variable Cloud Workloads with Deep Learning”
As per claim 1, Chen teaches the invention as claimed including a deep learning-based method for predicting a high-dimensional and highly-variable cloud workload, comprising the following steps:
Step S1: obtaining historical workload data of a cloud data center, and carrying out preprocessing (page 926, right column, lines 13-17 Workload Processor. Historical workload data from the cloud data center is used in the proposed prediction model. After the preprocessing and compression of historical workloads, workload data is regarded as the input of the prediction processor; Fig 2, describe a workload prediction model comprising a first step of receiving Historical Workload data);
Step S2: on the basis of a raw data set, predicting a future workload of a central processing unit by using a deep learning based prediction algorithm for cloud workloads (L-PAW) integrating a top-sparse auto-encoder (TSA) and a gated recurrent unit (GRU), and transmitting a predicted result to a cloud service provider (CSP) (page 927, left column, lines 28-31 Prediction Processor. Using the normalized and compressed historical workload data from workload processor, the future workloads are predicted by prediction processor and transferred to CSPs; Fig. 2, a fourth step performs a Gated RNN-based workload prediction and transfers result to CSP); and
Step S3: determining, by the CSP, a resource allocation strategy according to the predicted result, such that the cloud data center achieves load balancing (page 927, left column, lines 31-33 CSPs, who will use the predictions to determine the suitable resource provisioning strategies for load balancing in a cloud data center; Fig. 2 describes a fifth step of determining by the CSPs resource allocation strategy for loadbalancing).
As per claim 2, Chen further teaches: wherein the Step S1 specifically comprises: Step S11: obtaining the historical workload data of the cloud data center, and extracting central processing unit (CPU) utilization as raw workload data, denoted as X= (X1,X2,.+,Xp,), wherein n€R, and xn is CPU utilization at that time; and Step S12: normalizing the raw workload data (page 926, right column, lines 17-19 Generally, the historical workload data includes various metrics about the system running status (e.g., CPU usage, memory usage, and disk I/O time); page 927, right column, lines 24-25 the input of the TSA is a vector of workload examples X~ (x1, x2 … xn), where n€R and xn is the CPU usage at time n).
As per claim 3, Chen further teaches: replacing a hidden layer of the RNN with a GRU block on the basis of a basic feature representation of the workload extracted by the TSA (page 928, right column, lines 2-5 . Based on the essential feature representations of workloads extracted by the proposed TSA, we replace the hidden layers of classic RNN with GRU blocks); and after the TSA is called to obtain a compressed workload, setting a learning rate decay A to control a learning rate y in stages (page 28, right column, lines 6-10 After calling the TSA to obtain the compressed workloads, we set a learning rate decay to control the learning rate g segmentally, which aims to achieve more efficient learning at different stages for training the neural networks); wherein the GRU comprises two gates, namely, an update gate zt and a reset gate rt and an update mode of the two gates is on the basis of a current input xct and a previous hidden status yt-1; new memory content yt is regarded as new information of current time t, and the reset gate rt is configured to control whether previous memory needs to be retained; and the update gate z, is configured to control the previous memory content yt-1 and the new memory content y, to be forgotten or added (page 928, right column, line 28 – page 29, left column 6 Thus, GRU consists of two gates, which are the update gate zt and the reset gate rt. As shown in Fig. 4, we illustrate the structure of GRU block in the proposed L-PAW algorithm. Similar to LSTM, the update mode of these two gates is based on the current input xc t and the previous hidden status yt-1).
As per claim 4, Chen further teaches: an input of the TSA being a vector X= (x4,x2 … xn) of a workload example, wherein n € R, and xn is CPU utilization at a time n (page 927, right column, lines 24-26 As shown in Fig. 3, the input of the TSA is a vector of workload examples X= (x1, x2 … xn), where n € R and xn is the CPU usage at time n);
during forward propagation, an average activation degree O of hidden units being computed as follows: p= 1/n Sum [a(h) (xi)] wherein a) is an activation function of the hidden layer (page 927, right column, lines 35-37 During the forward propagation, the average activation degree of each hidden unit, r^, is calculated by Eq. (7));
next, all the hidden units being sorted according to respective p values, and the first k hidden units being recognized, which are denoted as a vector t = topk (p) (page 927, right column, lines 39-41, Next, all hidden units are sorted by their respective values of r^ and the top k hidden units can be identified, which is denoted as a vector t=topk(p)); computing a cost function JTSA,(W, b) =J(W,b) + B sumk KL(p II pj) of the TSA (page 927, right column, lines 17-18 Next, all hidden units are sorted by their respective values of r^ and the top k hidden units can be identified, which is denoted as a vector KL(p II pj), and
compressed workload data xc = Wxn +b, wherein W is a weight and b is a bias (page 927, right column, lines 21-22 Output the compressed workloads) and
executing backpropagation of the cost function JTSA,(W,b) through t = topk(p) (page 927, right column, lines 21-22 Execute the backpropagation of cost JTSAðW; bÞ through the definition of t=topk(p)).
As per claim 5, Chen further teaches: wherein the k hidden units with the highest activation degree are selected to reconstruct input data (page 927, right column, lines 33-35 where top k hidden units with the highest activation degree are selected for reconstructing the input data).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
U.S. Pub. No. 20230035451 A1 teaches resource usage prediction for deep learning model.
U.S. Pub. No. 20150371244 A1 teaches forecasting information technology workload demand.
U.S. Pub. No. 20250030610 A1 teaches a server load prediction method based on deep learning.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Willy W. Huaracha whose telephone number is (571)270-5510. The examiner can normally be reached on M-F 8:30-5:00pm.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Aimee Li can be reached on (571) 272-4169. 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 http://pair-direct.uspto.gov. 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.
/WH/
Examiner, Art Unit 2195
/BING ZHAO/Primary Examiner, Art Unit 2151