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
This action is responsive to the following communication: Non-Provisional Application filed Feb. 6, 2024.
Claims 1-20 are pending in the case. Claims 1, 16 and 20 are independent claims.
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 person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1-20 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Duan et al. (Hereinafter Duan) “SLADE: ASelf-Training Framework For Distance Metric Learning” 2021.
With respect to independent claim 1, Duan teaches a computer-implemented method comprising:
accessing, by a computing system including one or more processors, a first data set, the first data set comprising a plurality of unlabeled data elements (see e.g., Fig. 2 section 3 – “we use the fine-tuned teacher network to ex tract features and cluster the unlabeled data using k-means clustering.”);
training, by the computing system, one or more machine-learned encoder models for data encoding using each unlabeled data element in the first data set as input (see e.g., Fig. 2 section 3 – “we use self-supervised learning to initialize the teacher model. We use self-supervised pre-trained models ([5], [8], [4]) and fine-tune them on our data … In the pre-training stage, we fine tune the self-supervised model on the union of the labeled and un labeled images without using the label information to initialize the teacher model.”);
generating, by the computing system using the one or more machine-learned encoder models, an encoded version of each of a plurality of labeled data elements of a second data set (see e.g., Fig. 2 section 3 – “we fine tune the teacher network with a ranking loss (e.g., contrastive loss) on labeled data” corresponding to generating an encoded version of each of a plurality of labeled data elements);
training, by the computing system, a plurality of machine-learned decoder models for task-specific label generation using the encoded version of each of the plurality of labeled data elements of the second data set as input (see e.g., Fig. 2 figure 2: "Teacher", "Labeled data", and section "3.2. Pseudo Label Generation", discloses "we constrain pseudo label generation to linear combinations of a set of basis vectors. These basis vectors are trained in a supervised manner using labeled data and can be considered as class centers of labeled data", corresponding to training a machine-learned decoder model for task-specific label generation);
generating, by the computing system using the one or more machine-learned encoder models and the plurality of machine-learned decoder models, a plurality of associated provisional labels for the plurality of unlabeled data elements in the first data set, such that each unlabeled data element has an associated provisional label (figure 2: "2. Pseudo label generation (Sec. 3.2)", "Pseudo labels", and section "3.2. Pseudo Label Generation", discloses "we use teacher model to extract features, and cluster the unlabeled images using k-means [...] we constrain pseudo label generation to linear combinations in a supervised manner using labelled data", corresponding to using encoder and decoder models to generate provisional labels for the plurality of unlabeled data elements );
training, by the computing system, one or more student models using the plurality of unlabeled data elements from the first data set and the plurality of associated provisional labels (see e.g., figure 2: "3. Optimization of student network and basis vectors (Sec. 3.3)", "Student", "Unlabeled data", and section "3.3.3 Joint Training", discloses "we train the student network and basis vectors by minimizing a function L", corresponding to training a student model using unlabeled data and the associated provisional labels); and
deploying, by the computing system, the one or more student models onto one or more embedded computing devices (see e.g., section "2. Related work", discloses "self-training involves knowledge distillation from larger, more complex models or from ensembles of models (teachers) to less powerful, smaller students. Their end purpose is often reducing model size", corresponding to deploying a student model on an ).
With respect to dependent claim 2, Duan teaches a number of the unlabeled data elements in the first data set exceeds a number of the labeled data elements in the second data set (see e.g., section "4.1 Datasets", discloses "we use CUB-200-2011 as the labeled data and NABirds as the unlabeled data. CUB-200-2011 contains [...] 11,788 images [...] NABirds [...] has 48,000 images", corresponding to a number of unlabeled data elements exceeding a number of labeled data elements).
With respect to dependent claim 3, Duan teaches each model has a plurality of parameters (see e.g., abstract, discloses "teacher network" and "student network", and section Introduction", discloses "we propose an additional representation layer after the embedding layer", corresponding to each model having a plurality of parameters).
With respect to dependent claim 4, Duan teaches a number of parameters associated with the machine-learned encoder models and the machine-learned encoder models exceed a number of parameters associated with the student models (see e.g., section "2. Related work", discloses "self-training involves knowledge distillation from larger, more complex models or from ensembles of models (teachers) to less powerful, smaller students. Their end purpose is often reducing model size", corresponding to a number of parameters of the teacher model exceeding a number of parameters of the student model).
With respect to dependent claim 5, Duan teaches the one or more machine-learned encoder models are trained to be task agnostic (section "1. Introduction", discloses "self-supervised learning leverages unlabeled data to learn general features in a task-agnostic manner", corresponding to an encoder model trained to be task agnostic.).
With respect to dependent claim 6, Duan teaches training, by the computing system, one or more machine-learned encoder models for data encoding using each unlabeled data element in the first data set as input further comprises: for a respective machine-learned encoder model in the one or more machine-learned encoder models: initializing, by the computing system, values for a plurality of parameters associated with the respective machine-learned encoder model; generating, by the computing system and using the respective machine-learned encoder model, encoded data for a plurality of data elements in the first data set; evaluating, by the computing system, the encoded data using a task-agnostic algorithm; and updating, by the computing system, the values for the plurality of parameters associated with the respective machine-learned encoder model based on the evaluation of the encoded data using the task-agnostic algorithm (see e.g., Section "3.1. Self-Supervised Pre-Training and Fine-Tuning for Teacher Network", discloses "in the pre-training stage, we fine tune the self- supervised model on the union of the labeled and unlabeled images without using label information to initialize the teacher model [...] we fine tune the teacher-network using a ranking loss [...] on the labeled data", corresponding to initializing the parameters of the encoder model, evaluating the encoded data using a task-agnostic algorithm, and updating the encoder model based on the task-agnostic algorithm ).
With respect to dependent claim 7, Duan teaches the task-agnostic algorithm is a clustering algorithm (see e.g., section "3.2. Pseudo Label Generation", discloses "we use teacher model to extract features, and cluster the unlabeled images using k-means", corresponding to a task-agnostic clustering algorithm).
With respect to dependent claim 8, Duan teaches training, by the computing system, a plurality of machine-learned decoder models for task-specific label generation using the encoded version of each of the plurality of labeled data elements of the second data set as input further comprises: for a respective machine-learned decoder model in the one or more machine-learned encoder models: initializing, by the computing system, values for a plurality of parameters associated with the respective machine-learned decoder model; generating, by the computing system using the respective machine-learned decoder model, labels for a plurality of data elements in the second data set; comparing, by the computing system, the generated labels with the labels for the plurality of data elements in the second data set; and updating, by the computing system, the values for the plurality of parameters associated with the respective machine-learned decoder model based on comparing the generated labels with the labels for the plurality of data elements in the second data set (see e.g., section "3.1. Self-Supervised Pre-Training and Fine-Tuning for Teacher Network", discloses "we fine tune the teacher-network using a ranking loss [...] on the labeled data", corresponding to generating labels by a decoder model, comparing the generated labels with the labels in the labeled data set, and updating the parameters of the decoder model based on the comparison result.)
With respect to dependent claim 9, Duan teaches the labels are predetermined and domain specific labels associated with a specific task (see e.g., section "3.2. Pseudo Label Generation", discloses "cluster ids are then used as pseudo labels", corresponding to domain specific labels being associated with a specific task.)
With respect to dependent claim 10, Duan teaches aggregating, by the computing system for a particular data element included in the first data set, a plurality of distinct provisional labels generated by a plurality of machine-learned decoder models into an aggregated provisional label (see e.g., section "3.3.1 Feature Basis Learning", discloses "we optimize with a global similarity distribution loss on the unlabeled data. We use the pseudo labels to sample a set of pseudo positive pairs and pseudo negative pairs, where the pseudo positive pairs are sampled from the same pseudo class and the pseudo negative pairs are sampled from different pseudo classes", corresponding to aggregating a plurality of district provisional labels into an aggregated provisional label.)
With respect to dependent claim 11, Duan teaches the aggregated provisional label includes one or more potential labels, each potential label having an associated likelihood value (see e.g., section "3.3.1 Feature Basis Learning", discloses "we optimize with a global similarity distribution loss on the unlabeled data. We use the pseudo labels to sample a set of pseudo positive pairs and pseudo negative pairs, where the pseudo positive pairs are sampled from the same pseudo class and the pseudo negative pairs are sampled from different pseudo classes", corresponding to aggregating a plurality of district provisional labels into an aggregated provisional label, each potential label having an associated likelihood value.).
With respect to dependent claim 12, Duan teaches training, by the computing system, one or more student models using the plurality of unlabeled data elements from the first data set and the plurality of associated provisional labels further comprises: for a respective student model in the one or more machine-learned encoder models: initializing, by the computing system, values for a plurality of parameters associated with the respective student model; generating, by the computing system using the respective student model, labels for a plurality of data elements in the first data set; comparing, by the computing system, the generated labels with the aggregated provisional labels generated by the plurality of machine-learned decoder models for the plurality of data elements in the first data set; and updating, by the computing system, the values for the plurality of parameters associated with the respective student model based on comparing the generated labels and the aggregated provisional labels generated by the plurality of machine-learned decoder models for the plurality of data elements in the first data set (see e.g., section "3.3.2 Sample Mining", discloses "we use the basis vectors to select high-confidence sample pairs from the unlabeled images for training the student network [...] we set the confidence thresholds of T1 and T2", corresponding to initializing parameters of a student model, comparing generated labels with aggregated labels, and updating the model based on the comparison).
With respect to dependent claim 13, Duan teaches combining the one or more machine-learned encoder models and the plurality of machine-learned decoder models after training into a plurality of machine-learned teacher models that take data elements from the first data set as input and output provisional labels associated each data element in the first data set (see e.g., section "1 Introduction", discloses "once the teacher model is pre-trained and fine tuned, we use it to generate pseudo labels for unlabeled data [...] we propose an additional representation layer after the embedding layer. This new layer is only used for unlabeled data and aims at learning basis functions for the feature representation of unlabeled data [...] we use the learned basis vectors to compute the feature representation of each image and measure pairwise similarity for unlabeled data", corresponding to combining encoder models and a plurality of decoder models into a plurality of teacher models, that output provisional labels.).
With respect to dependent claim 14, Duan teaches removing, by the computing system, labels from the plurality of labeled data elements in the second data set to generate one or more unlabeled data elements and using the one or more unlabeled data elements from the second data set to train the student models (see e.g., "3.1. Self-Supervised Pre-Training and Fine-Tuning for Teacher Network", discloses "we fine-tune the self-supervised model on the union of the labeled and unlabeled images without using the label information", corresponding to removing labels for training.).
With respect to dependent claim 15, Duan teaches training, by the computing system, a plurality of machine-learned decoder models for task-specific label generation using the encoded version of each of the plurality of labeled data elements of the second data set as input further comprises: for a respective machine-learned decoder model: combining, by the computing system, the respective machine-learned decoder model with a machine-learned encoder model into an encoder decoder teacher model; generating, by the computing system using the encoder decoder teacher model, labels for a plurality of data elements in the second data set; comparing, by the computing system, the generated labels with the labels for the plurality of data elements in the second data set; and updating, by the computing system, parameter values associated with the respective machine-learned decoder model and the machine-learned encoder model included in the encoder decoder teacher model (see e.g., section "1 Introduction", discloses "once the teacher model is pre-trained and fine tuned, we use it to generate pseudo labels for unlabeled data [...] we propose an additional representation layer after the embedding layer. This new layer is only used for unlabeled data and aims at learning basis functions for the feature representation of unlabeled data [...] we use the learned basis vectors to compute the feature representation of each image and measure pairwise similarity for unlabeled data", corresponding to combining encoder models and a plurality of decoder models into a plurality of teacher models, that output provisional labels.).
Claim 16 is rejected for the similar reasons discussed above with respect to claim 1.
Claim 17 is rejected for the similar reasons discussed above with respect to claim 2.
Claim 18 is rejected for the similar reasons discussed above with respect to claim 3.
Claim 19 is rejected for the similar reasons discussed above with respect to claim 4.
Claim 20 is rejected for the similar reasons discussed above with respect to claim 1.
It is noted that any citation to specific pages, columns, lines, or figures in the prior art references and any interpretation of the references should not be considered to be limiting in any way. “The use of patents as references is not limited to what the patentees describe as their own inventions or to the problems with which they are concerned. They are part of the literature of the art, relevant for all they contain.” In re Heck, 699 F.2d 1331, 1332-33, 216 USPQ 1038, 1039 (Fed. Cir. 1983) (quoting In re Lemelson, 397 F.2d 1006, 1009, 158 USPQ 275, 277 (CCPA 1968)). Further, a reference may be relied upon for all that it would have reasonably suggested to one having ordinary skill the art, including nonpreferred embodiments. Merck & Co. v. Biocraft Laboratories, 874 F.2d 804, 10 USPQ2d 1843 (Fed. Cir.), cert. denied, 493 U.S. 975 (1989). See also Upsher-Smith Labs. v. Pamlab, LLC, 412 F.3d 1319, 1323, 75 USPQ2d 1213, 1215 (Fed. Cir. 2005); Celeritas Technologies Ltd. v. Rockwell International Corp., 150 F.3d 1354, 1361, 47 USPQ2d 1516, 1522-23 (Fed. Cir. 1998).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to PEIYONG WENG whose telephone number is (571)270-1660. The examiner can normally be reached on Mon.-Fri. 8 am to 5 pm.
If attempts to reach the examiner by telephone are unsuccessful, the examiner's supervisor, Matthew Ell, can be reached on (571) 270-3264. 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://portal.uspto.gov/external/portal. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free).
/PEI YONG WENG/Primary Examiner, Art Unit 2141
Prosecution Insights