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 .
Priority
Receipt is acknowledged of certified copies of papers required by 37 CFR 1.55.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 07/09/2024 has been considered by examiner.
Patent statutory analysis
The claims 1-15 appear to be patent eligible under the 2019 PEG. Even although it seems the claims reciting abstract idea involving mathematical concepts / ML computations—e.g., computing embeddings, computing distance metrics, training a graph, k-nearest-neighbor predictions, the claims as a whole give the idea to integrate those computations into a practical application. A specific improvement to graph-based machine learning, enabling inductive prediction for unseen graph entities when entities may lack numerical/categorical features and may have only name strings or limited/no graph connectivity.
Support for examiner’s analysis can be found in the specification as indicted below:
Improvement to graph-based inductive learning (paragraphs 19–24, and 35).
use of pretrained models to generate contextual entity embeddings from related corpora (paragraphs 26–31, and 47–54).
enabling unseen entities to be added at test time without retraining or prior graph relationships, (paragraphs 24, 35, and 61–62).
k-nearest-neighbor score aggregation and self-supervised graph update mechanisms, (paragraphs 58–69).
Therefore, claims 1–15 is eligible under § 101 at Step 2A, Prong 2, because the recited operations are integrated into a practical application that improves graph-based machine-learning functionality for inductive predictions on unseen entities.
Claim Rejections - 35 USC § 102
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 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.
Claim(s) 1-15 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by “scientific language model for biomedical knowledge base completion: an impirical study”, (hereinafter Study).
Regarding claims 1, and 14-15, study teaches a computer-implemented method, system and medium for inductive learning on graphs, wherein the graph includes a plurality of entities, wherein relationships exist between the plurality of entities, and wherein the plurality of entities and relationships have a name string, (section 2 discloses presenting the knowledge graph completion task and predicting missing links, and section 2.1 discloses the knowledge graph is consist of entities E, relation R, and triples T, and each of the entity is also associated with some text which can be the entity name) comprising:
(a) creating for each entity of the plurality of entities of the graph, a related text corpus, based on the respective name string of each entity's name string, a related text corpus (Name string; text created for use with an LM model; text corpus for each entity using its name and textual description. see page 2, and section 2.1 and 3.1);
(b) using a pretrained language model to compute, from the related text corpus of each entity, a respective contextual entity embedding for each entity of the graph (See section 2.2, 2.4, 3.2 and 4.3; LM used to compute, a contextual entity embedding for each entity; using pretrained LM model to generate contextualized embedding from entity text);
training a graph-based machine-learning; -ML)- model by using, for each entity of the graph, the computed entity embeddings; and repeating for unseen entities, steps (a) and (b) and using the trained ML model to perform inductive predictions for the unseen entities (see section 2.2, 2.4, 3.2 and 4.3. Train graph-based ML models using LM derived embeddings; the model is trained for inductive completion/learning).
Regarding claim 2, Study discloses the method according to claim 1, further comprising: computing a distance metric between the entity embeddings and identifying, based on the distance metric, the k-nearest neighbors set for each entity of the plurality of entities of the graph; and computing, by using the trained ML model, a set of k+1 predictions for each entity including a prediction for the respective entity itself and k predictions for the k-nearest neighbors (section 2.4, and 4.3. also see table 9. similarity between seen and unseen entities to replace each unseen with the closest trained embedding).
Regarding claim 3, study discloses the method according to claim 2, further comprising:
aggregating the computed k+1 predictions to obtain an aggregated prediction for each entity; and providing the aggregated predictions as a prediction output (aggregation of predictions from entity and neighbors and use it for evaluation tasks. Providing the output for further use. see section 4.2, and 5.1).
Regarding claim 4, study discloses the method further comprising: updating the
graph using triples obtained by substituting a test entity with its k- nearest neighbors of the test entity (updating the graph with new triples based on neighbor predictions. see section 4.4).
Regarding claim 5, study discloses the method further comprising: repeating the
steps of claims 2 until a desired or configurable end condition is reached (neighbor aggregation and graph updated is repeated; computation until a stopping condition. See section 4.5).
Regarding claim 6, study discloses the method of claim 1, wherein the related text corpus for each node entity of the plurality of entities of the graph is created by querying a database of textual data and/or by mining text from an external source (text is extracted; text mining and database querying for entity corpus. See section 3.1).
Regarding claims 7-13, study further discloses extracting and converting them to natural language (Claim 7, rule-based templates to convert extracted relations into natural language. See section 3.3), embedding aggregation (claims 8; mean pooling, max pooling, sum pooling. See section 3.2), specific biomedical relationship (claims 10-13. See section 2.2, 2.3, 2.4, and 5.3).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Nasser Goodarzi whose telephone number is (571)272-4195. The examiner can normally be reached on 8:00am -4:30pm. If attempts to reach the examiner by telephone are unsuccessful, the examiner's supervisor, Colleen Fauz can be reached on 571-272-1667. 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 USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/NASSER M GOODARZI/Supervisory Patent Examiner, Art Unit 2426
Prosecution Insights