IP Library › Granted Patent US 12,737,645
Granted Patent B2
US 12,737,645 · App. 18/178,047 · Granted Sep 15, 2026

Graph-based machine learning model with feature-agnostic training and holistic feature metrics

Inventors: Premnath Kandhasamy Narayanan (Athione, IE); David S. Monaghan (Dublin, IE); Brian Carter (Dublin, IE); Amirhossein Yazdavar (Scottsdale, AZ); Triet Pham (Eden Prairie, MN)
Assignee: Optum Services (Ireland) Limited
G06N5/022G06N3/045G06N3/08G06N20/00
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,737,645
App. No.
18/178,047
Granted
Sep 15, 2026
Kind
B2
Abstract

Various embodiments of the present disclosure describe data evaluation techniques that leverage a graph-based machine learning model to evaluate a knowledge graph. The techniques include using a target graph model to generate a predictive representation for a graph node of a graph training dataset. The techniques include using a feature prediction model to generate predicted feature values for the graph node based on the predictive representation. The techniques include generating a data evaluation score for the graph training dataset based on the predicted feature values. The techniques include using the target graph model to generate a predictive output for the graph node based on the predictive representation and then generating an evaluation output for the target graph model based on the evaluation score and the predictive output.

Claims (46)

1 . A computer-implemented method comprising:

generating, by one or more processors and using a target graph model, a predictive representation for a graph node of a graph training dataset;

generating, by the one or more processors and using a feature prediction model, one or more predicted feature values for the graph node based on the predictive representation;

generating, by the one or more processors, a data evaluation score for the graph training dataset based on the one or more predicted feature values, wherein the data evaluation score is based on a graph feature confidence score indicative of a predicted accuracy of the one or more predicted feature values;

generating, by the one or more processors and using the target graph model, a predictive output for the graph node based on the predictive representation, wherein (i) the predictive output comprises a node classification for the graph node, (ii) the one or more predicted feature values correspond to one or more evaluation features of the graph training dataset, and (iii) the target graph model is previously trained to generate an evaluation feature-agnostic predictive representation that at least partially prevents the target graph model from generating the predictive output based on the one or more evaluation features; and

generating, by the one or more processors, an evaluation output for the target graph model based on the data evaluation score and the predictive output.

2 . The computer-implemented method of claim 1 , wherein the target graph model comprises a first graph neural network (GNN) and the feature prediction model comprises a second GNN.

3 . The computer-implemented method of claim 1 , wherein the target graph model and the feature prediction model are jointly trained using a joint objective function.

4 . The computer-implemented method of claim 3 , wherein:

the joint objective function comprises a first objective function and a second objective function,

the first objective function comprises a first optimization function for the target graph model, and

the second objective function comprises a second optimization function for the feature prediction model.

5 . The computer-implemented method of claim 4 , wherein the first objective function comprises a comparison between the predictive output of the target graph model and a ground truth label for the graph node.

6 . The computer-implemented method of claim 1 , wherein the predictive representation comprises a feature embedding that encodes one or more features of the graph node and one or more adjacent features of one or more neighboring nodes of the graph node in the graph training dataset.

7 . The computer-implemented method of claim 6 , wherein the one or more neighboring nodes of the graph node comprise one or more nodes of the graph training dataset that are connected to the graph node by one or more edges.

8 . The computer-implemented method of claim 1 , wherein the graph feature confidence score is based on a predicted feature confidence score indicative of a confidence level for the one or more predicted feature values.

9 . The computer-implemented method of claim 8 , wherein the predicted feature confidence score is generated by the feature prediction model.

10 . A system comprising:

one or more processors; and

at least one memory storing processor-executable instructions that, when executed by the one or more processors, cause the one or more processors to:

generate, using a target graph model, a predictive representation for a graph node of a graph training dataset;

generate, using a feature prediction model, one or more predicted feature values for the graph node based on the predictive representation;

generate data evaluation score for the graph training dataset based on the one or more predicted feature values, wherein the data evaluation score is based on a graph feature confidence score indicative of a predicted accuracy of the one or more predicted feature values;

generate, using the target graph model, a predictive output for the graph node based on the predictive representation, wherein the predictive output comprises a node classification for the graph node, wherein the one or more predicted feature values correspond to one or more evaluation features of the graph training dataset, and wherein the target graph model is previously trained to generate an evaluation feature-agnostic predictive representation that at least partially prevents the target graph model from generating the predictive output based on the one or more evaluation features; and

generate an evaluation output for the target graph model based on the data evaluation score and the predictive output.

11 . The system of claim 10 , wherein the target graph model comprises a first graph neural network (GNN) and the feature prediction model comprises a second GNN.

12 . The system of claim 10 , wherein the target graph model and the feature prediction model are jointly trained using a joint objective function.

13 . The system of claim 12 , wherein:

the joint objective function comprises a first objective function and a second objective function,

the first objective function comprises a first optimization function for the target graph model, and

the second objective function comprises a second optimization function for the feature prediction model.

14 . The system of claim 13 , wherein the first objective function comprises a comparison between the predictive output generated by the target graph model and a ground truth label for the graph node.

15 . The system of claim 10 , wherein the graph feature confidence score is based on a predicted feature confidence score indicative of a confidence level for the one or more predicted feature values.

16 . One or more non-transitory computer-readable storage media including instructions that, when executed by one or more processors, cause the one or more processors to:

generate, using a target graph model, a predictive representation for a graph node of a graph training dataset;

generate, using a feature prediction model, one or more predicted feature values for the graph node based on the predictive representation;

generate data evaluation score for the graph training dataset based on the one or more predicted feature values, wherein the data evaluation score is based on a graph feature confidence score indicative of a predicted accuracy of the one or more predicted feature values;

generate, using the target graph model, a predictive output for the graph node based on the predictive representation, wherein the predictive output comprises a node classification for the graph node, wherein the one or more predicted feature values correspond to one or more evaluation features of the graph training dataset, and wherein the target graph model is previously trained to generate an evaluation feature-agnostic predictive representation that at least partially prevents the target graph model from generating the predictive output based on the one or more evaluation features; and

generate an evaluation output for the target graph model based on the data evaluation score and the predictive output.

17 . The one or more non-transitory computer-readable storage media of claim 16 , wherein the predictive representation comprises a feature embedding that encodes one or more features of the graph node and one or more adjacent features of one or more neighboring nodes of the graph node in the graph training dataset.

18 . The one or more non-transitory computer-readable storage media of claim 16 , wherein the target graph model comprises a first graph neural network (GNN) and the feature prediction model comprises a second GNN.

19 . The one or more non-transitory computer-readable storage media of claim 16 , wherein the target graph model and the feature prediction model are jointly trained using a joint objective function.

20 . The one or more non-transitory computer-readable storage media of claim 19 , wherein:

the joint objective function comprises a first objective function and a second objective function,

the first objective function comprises a first optimization function for the target graph model, and

the second objective function comprises a second optimization function for the feature prediction model.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Mar 3, 2023
From: NARAYANAN, PREMNATH KANDHASAMY; MONAGHAN, DAVID S.; CARTER, BRIAN; YAZDAVAR, AMIRHOSSEIN; PHAM, TRIET
To: OPTUM SERVICES (IRELAND) LIMITED
Reel/Frame 062875/0305 →
Continuity (2)
Provisional Application 63479874 · Jan 13, 2023
Related Publication 20240256832A1 · Aug 1, 2024
References Cited (73)
US 11256989B2 · Dalli et al. · 2022 [cited by applicant]
US 11263550B2 · Vasconcelos et al. · 2022 [cited by applicant]
US 20200364859A1 · Najarian et al. · 2020 [cited by applicant]
US 20210073627A1 · Sarferaz · 2021 [cited by applicant]
US 20210357316A1 · Chang et al. · 2021 [cited by applicant]
US 20210406727A1 · Wang et al. · 2021 [cited by applicant]
US 20220114399A1 · Castiglione et al. · 2022 [cited by applicant]
US 20220129794A1 · McGrath et al. · 2022 [cited by applicant]
US 20220318639A1 · Upadhyay et al. · 2022 [cited by applicant]
US 20220318640A1 · Malur Srinivasan et al. · 2022 [cited by applicant]
US 20230074606A1 · Jesus et al. · 2023 [cited by applicant]
Dai et al., Say No to the Discrimination: Learning Fair Graph Neural Networks with Limited Sensitive Attribute Information, WSDM' 21, Mar. 8-12, 2021, Virtual Event, Israel; pp. 680-688 (Year: 2021). [cited by examiner]
Fan et al., Debiasing Graph Neural Networks via Learning Disentangled Causal Substructure; arXiv:2209.14107v1 [cs. LG] Sep. 28, 2022; Total p. 18 (Year: 2022). [cited by examiner]
Dong et al., EDITS: Modeling and Mitigating Data Bias for Graph Neural Networks; arXiv:2108.05233v2 [cs.LG] Feb. 19, 2022; Total p. 11 (Year: 2022). [cited by examiner]
Linardatos et al., Explainable AI: A Review of Machine Learning Interpretability Methods; Entropy 2021, 23, 18. https://dx.doi.org/10.3390/e23010018; Total p. 45 (Year: 45). [cited by examiner]
“MDClone—The World's Most Powerful Healthcare Data Platform,” (Year: 2023), (6 pages), [Retrieved from the Internet May 12, 2023] <URL: https://www.mdclone.com/>. [cited by applicant]
“Safely Incorporate Generative AI Into Your Data.” Gretel, Aug. 27, 2018, (6 pages), [Retrieved from the Internet May 12, 2023] <URL: https://gretel.ai/>. [cited by applicant]
“Synthetic Data For Machine Learning: Its Nature, Types, and Means of Generation,” Altexsoft, Mar. 22, 2022, (18 pages), [Retrieved from the Internet May 12, 2023] <URL: https://www.altexsoft.com/blog/synthetic-data-gen… [cited by applicant]
“Text Generation API,” DeepAI, (10 pages), (online), [Retrieved from the Internet May 12, 2023] <URL: https://deepai.org/machine-learning-model/text-generator>. [cited by applicant]
“What is an EDI 837?” TrueCommerce, May 4, 1997, (6 pages), [Retrieved from the Internet May 12, 2023] <URL: https://www.truecommerce.com/edi-transaction-codes/edi-837>. [cited by applicant]
“What is X12?” X12, Oct. 19, 1998, (8 pages), [Retrieved from the Internet May 12, 2023] <URL: https://x12.org/>. [cited by applicant]
Anderson, Lydia et al. “Census Bureau Survey Explores Sexual Orientation and Gender Identity,” United States Census Bureau, Nov. 4, 2021, (8 pages), available online: https://www.census.gov/library/stories/2021/11/censu… [cited by applicant]
Barry, Laurence et al. “The Fairness of Machine Learning In Insurance: New Rags for an Old Man?” arXiv:2205.08112v1 [econ.GN], May 17, 2022, pp. 1-18, available online: https://arxiv.org/abs/2205.08112. [cited by applicant]
Bommasani, Rishi et al. “On The Opportunities and Risks of Foundation Models,” Center for Research on Foundation Models (CRFM), arXiv:2108.07258v3 [cs.LG], Jul. 12, 2022, pp. 1-214, available online: https://arxiv.org/p… [cited by applicant]
Centers for Medicate & Medicaid Services. “CMS 2008-2010 Data Entrepreneurs' Synthetic Public Use File (DE-SynPUF),” Jun. 24, 2022, (4 pages), [Retrieved from the Internet May 12, 2023] <URL: https://www.cms.gov/Researc… [cited by applicant]
Cheng, James et al. “GraphGen—A Synthetic Graph Data Generator,” (2 pages), [Retrieved from the Internet May 12, 2023] <URL: https://cse.hkust.edu.hk/graphgen/>. [cited by applicant]
CitiusTech. “Improving Claims Auto Adjudication Rate Using AI/ML,” May 13, 2020, (4 pages), YouTube, available online: https://www.youtube.com/watch?v=KiKMKnMJyi4. [cited by applicant]
Dandl, Susanne et al. “9.3 Counterfactual Explanations,” Interpretable Machine Learning: a Guide for Making Black Box Models Explainable, Feb. 28, 2022, (14 pages), [Retrieved from the Internet May 12, 2023] <URL: https… [cited by applicant]
Datacebo. “SDV—The Synthetic Data Vault,” (Year: 2023), (8 pages), [Retrieved from the Internet May 12, 2023] <URL: https://sdv.dev/>. [cited by applicant]
Devaux, Elise et al. “An Overview of Synthetic Data Types and Generation Methods,” KD Nuggets, (Year: 2023), (6 pages), [Retrieved from the Internet May 12, 2023] <URL: https://www.kdnuggets.com/2021/02/overview-synthet… [cited by applicant]
Documentation. “Model-Assisted Labeling,” Jul. 1, 2022, (2 pages), [Retrieved from the Internet May 12, 2023] <URL: https://docs.segments.ai/tutorials/model-assisted-labeling>. [cited by applicant]
European Commission. “A European Approach to Artificial Intelligence” Shaping Europe's Digital Future, Jan. 26, 2023, (10 pages), [Retrieved from the Internet May 12, 2023] <URL: https://digital-strategy.ec.europa.eu/en… [cited by applicant]
European Commission. “Proposal for a Regulation Laying Down Harmonised Rules on Artificial Intelligence,” Shaping Europe's Digital Future, Apr. 21, 2021, (2 pages), [Retrieved from the Internet May 12, 2023] <URL: https… [cited by applicant]
European Commission. “Regulatory Framework Proposal on Artificial Intelligence,” Shaping Europe's Digital Future, Sep. 29, 2022, (7 pages), [Retrieved from the Internet May 13, 2023] <URL: https://digital-strategy.ec.eu… [cited by applicant]
Federico, Arenas L. “Generate Synthetic Data With Blender and Python,” GitHub, Feb. 13, 2021, (5 pages), [Retrieved from the Internet May 12, 2023] <URL: https://github.com/federicoarenasl/Data-Generation-with-Blender>. [cited by applicant]
Friedman, Jerome H. “Greedy Function Approximation: A Gradient Boosting Machine.” Annals of Statistics, vol. 29, No. 5, (Year:2001), pp. 1189-1232. [cited by applicant]
Goodfellow, Ian J. et al. “Generative Adversarial Nets,” arXiv:1406.2661v1 [stat.ML], Jun. 10, 2014, pp. 1-9, available online: https://arxiv.org/pdf/1406.2661.pdf. [cited by applicant]
Google AI. “Responsible AI Practices,” (Year: 2022), (12 pages), (online), [Retrieved from the Internet May 12, 2023] <URL: https://ai.google/responsibility/responsible-ai-practices/>. [cited by applicant]
Hugging Face. “Mustang/Bert_responsible_AI,” Jan. 26, 2022, (2 pages), [Retrieved from the Internet May 12, 2023] <URL: https://huggingface.co/Mustang/BERT_responsible_AI>. [cited by applicant]
Kingma, Diederik P. et al. “Auto-Encoding Variational Bayes,” arXiv:1312.6114v11 [stat.ML], Dec. 10, 2022, pp. 1-14, available online: https://arxiv.org/pdf/1312.6114.pdf. [cited by applicant]
Kusner, Matt et al. “Counterfactual Fairness,” 31st Conference on Neural Information Processing System (NIPS 2017), Long Beach, CA, USA, (11 pages), available online: https://proceedings.neurips.cc/paper/2017/file/a486c… [cited by applicant]
Marina, Gretchen. “GPT-3: Language Models Are Few-Shot Learners,” GitHub, Sep. 18, 2020, (3 pages), [Retrieved from the Internet May 12, 2023] <URL: https://github.com/openai/gpt-3>. [cited by applicant]
MLOps. “MLOps Principles,” Jan. 22, 2023, (23 pages), [Retrieved from the Internet May 12, 2023] <URL: https://ml-ops.org/content/mlops-principles. [cited by applicant]
Morocho-Cayamcela, Manuel Eugenio et al. “Machine Learning For 5G/B5G Mobile and Wireless Communications: Potential, Limitations, and Future Directions,” IEEE Access, vol. 7, Sep. 19, 2019, pp. 137184-137206, DOI: 10.11… [cited by applicant]
Mostly AI. “bn-learn—An R Package for Bayesian Network Learning and Inference,” Dec. 15, 2017, (3 pages), [Retrieved from the Internet May 12, 2023] <URL: https://mostly.ai/synthetic-data/fair-synthetic-data>. [cited by applicant]
Mostly AI. “Fair Synthetic Data for Bias Free Machine Learning,” Dec. 15, 2017, (7 pages), [Retrieved from the Internet May 12, 2023] <URL: https://mostly.ai/synthetic-data/fair-synthetic-data>. [cited by applicant]
Narayanan, Premnath K. et al. “Sequencing of Autonomous Network Functions Using eXplainable AI Methods,” 2022 3rd International Conference on Intelligent Engineering and Management (ICIEM), London, United Kingdom, Apr. … [cited by applicant]
Obermeyer, Ziad et al. “Dissecting Racial Bias in an Algorithm Used to Manage the Health of Populations,” Science, vol. 366, pp. 447-453, Oct. 25, 2019, DOI: 10.1126/science.aax2342. [cited by applicant]
Pandas. “pandas.cut,” (3 pages), [Retrieved from the Internet May 12, 2023] <URL: https://pandas.pydata.org/docs/reference/api/pandas.cut.html>. [cited by applicant]
Pawelczyk, Martin et al. “Carla: A Python Library to Benchmark Algorithmic Recourse and Counterfactual Explanation Algorithms,” arXiv:2108.00783v1 [cs.LG], Aug. 2, 2021, (22 pages), available online: https://arxiv.org/p… [cited by applicant]
Pessach, Dana et al. “Algorithmic Fairness,” arXiv:2001.09784v2 [cs.CY], Jan. 21, 2020, pp. 1-31 pages, available online: https://arxiv.org/pdf/2001.09784.pdf. [cited by applicant]
Python Package Index. “ethnicolr: Predict Race and Ethnicity From Name,” ethnicolr 0.9.6, Apr. 17, 2023, (20 pages), [Retrieved from the Internet May 12, 2023] <URL: https://pypi.org/project/ethnicolr/>. [cited by applicant]
Python Package Index. “gender-guesser 0.4.0,” Dec. 5, 2016, (4 pages), [Retrieved from the Internet May 12, 2023] <URL: https://pypi.org/project/gender-guesser/>. [cited by applicant]
Sarmento, David. “Chapter 22: Correlation Types and When to Use Them, ” Ademos, (14 pages), [Retrieved from the Internet May 12, 2023] <URL: https://ademos.people.uic.edu/Chapter22.html>. [cited by applicant]
Scarselli, Franco et al. “The Graph Neural Network Model,” IEEE Transactions on Neural Networks, vol. 20, No. 1, pp. 61-80, Jan. 1, 2009, available online: https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.1015.… [cited by applicant]
Settles, Burr. “Active Learning Literature Survey,” Computer Sciences Technical Report 1648, University of Wisconsin—Madison, Jan. 26, 2010, (67 pages), available online: https://burrsettles.com/pub/settles.activelearni… [cited by applicant]
Singh, Amitojdeep et al. “Explainable Deep Learning Models in Medical Image Analysis,” Journal of Imaging, vol. 6, No. 52, Jun. 20, 2020, pp. 1-19, DOI: 10.3390/jimaging6060052. [cited by applicant]
Solomou, Chris et al. “Utilizing Chest X-Rays for Age Prediction and Gender Classification,” In Proceedings of the 4th International Seminar on Research of Information Technology and Intelligent Systems (ISRITI), Intern… [cited by applicant]
Stanford University. “Introducing The Center for Research on Foundation Models (CRFM),” Aug. 18, 2021, (2 pages), (article), [Retrieved from the Internet May 12, 2023] <URL: https://hai.stanford.edu/news/introducing-cen… [cited by applicant]
The White House. “Blueprint for an AI Bill of Rights—Making Automated Systems Work for the American People,” (8 pages), [Retrieved from the Internet May 12, 2023] <URL: https://www.whitehouse.gov/ostp/ai-bill-of-rights/… [cited by applicant]
Thesmar, David et al. “Combining the Power of Artificial Intelligence With the Richness of Healthcare Claims Data—Opportunities and Challenges,” PharmacoEconomics, Vo. 37, pp. 745-752, Mar. 8, 2019, DOI: 10.1007/s40273-… [cited by applicant]
United States Census Bureau. “Why We Ask Questions About . . . Race,” American Community Survey, (Year: 2023), (2 pages), available online: https://www.census.gov/acs/www/about/why-we-ask-each-question/race/#:~:text=The… [cited by applicant]
West, Jeremy. “A Theoretical Foundation for Inductive Transfer—Abstract,” Spring Research Presentation—Brigham Young University, College of Physical and Mathematical Sciences, (Year: 2009), (1 page), available online: h… [cited by applicant]
Xu, Lei et al. “Modeling Tabular Data Using Conditional GAN,” arXiv:1907.005032v2 [cs.LG], Oct. 28, 2019, (15 pages), 33rd Conference on Neural Information Processing Systems (NeurIPS 2019), Vancouver Canada, available … [cited by applicant]
Xue, Zhiyun et al. “Gender Detection From Spine X-Ray Images Using Deep Learning,” 2018 IEEE 31st International Symposium on Computer-Based Medical Systems (CBMS), Jun. 18, 2018, pp. 54-58, DOI: 10.1109/CBMS.2018.00017. [cited by applicant]
ZumoLabs. “zpy-Synthetic Data for Computer Vision. An Open Source Toolkit Using Blender and Python,” GitHub, Aug. 4, 2021, (3 pages), [Retrieved from the Internet May 12, 2023] <URL: https://github.com/ZumoLabs/zpy>. [cited by applicant]
Mehrabi, Ninareh, et al., “A Survey on Bias and Fairness in Machine Learning” Jul. 13, 2021, 36 pages, ACM Computing Surveys, vol. 54, No. 6, article 115, Association of Computing Machinery, US. [cited by applicant]
Mothilal, Ramaravind, et al., “Explaining Machine Learning Classifiers through Diverse Counterfactual Explanations”, submitted Dec. 6, 2019 to Cornell University Olin Online Library Archive, available on the Internet at… [cited by applicant]
Non-Final Rejection Mailed on Jan. 12, 2026 for U.S. Appl. No. 18/178,043, 27 page(s). [cited by applicant]
Non-Final Rejection Mailed on Jan. 22, 2026 for U.S. Appl. No. 18/178,056, 61 page(s). [cited by applicant]
Wan, Cen, et al., “Protein function prediction is improved by creating synthetic feature samples with generative adversarial networks”, Nature Machine Intelligence, Aug. 31, 2020, 26 pages, vol. 2, No. 9, Springer Scien… [cited by applicant]
Final Rejection Mailed on Jul. 7, 2026 for U.S. Appl. No. 18/178,043, 32 page(s). [cited by applicant]
Final Rejection Mailed on Jun. 24, 2026 for U.S. Appl. No. 18/178,056, 74 page(s). [cited by applicant]