IP Library › Granted Patent US 12,737,530
Granted Patent B2
US 12,737,530 · App. 18/885,775 · Granted Sep 15, 2026

System and method using global and local-aware denoising framework for commonsense knowledge graph noise detection

Inventors: Zheye Deng (Hong Kong, CN); Weiqi Wang (Hong Kong, CN); Zhaowei Wang (Hong Kong, CN); Yangqiu Song (Hong Kong, CN)
Assignee: The Hong Kong University of Science and Technology
G06F40/157G06F40/30
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 12,737,530
App. No.
18/885,775
Granted
Sep 15, 2026
Kind
B2
Abstract

A system using global and local-aware denoising framework for commonsense knowledge graph noise detection is provided. The system includes an input receiver, a triple encoder, a global noise detector, a local noise detector, and a comprehensive evaluation scorer. The triple encoder receives the CSKG information from the input receiver and extract semantic information from free-text formatted nodes in the CSKG information, so as to obtain structured triples. The global noise detector identifies high-frequency patterns through rule mining and uses a rule encoder to generalize learned rules and guide noise detection. The local noise detector utilizes the structured triples output by the triple encoder to aggregate neighboring triple information around both the head and tail nodes via graph networks. The comprehensive evaluation scorer provides a comprehensive score based on global noise detection and local noise detection, so as to indicate a noise level for the CSKG information.

Claims (35)

1 . A system using global and local-aware denoising framework for commonsense knowledge graph noise detection, comprising:

an input receiver serving as a user interface responsible for receiving user inputs and configured to process the user inputs to obtain commonsense knowledge graph (CSKG) information;

a triple encoder configured to receive the CSKG information from the input receiver and extract semantic information from free-text formatted nodes in the CSKG information, so as to obtain structured triples;

a global noise detector receiving the structured triples output from the triple encoder and configured to analyze and detect noise at a global level from the CSKG information, wherein the global noise detector identifies high-frequency patterns through rule mining and uses a rule encoder to generalize learned rules and guide noise detection;

a local noise detector receiving the structured triples output from the triple encoder and configured to analyze and detect noise at a local level from the CSKG information, wherein the local noise detector utilizes the structured triples output by the triple encoder to aggregate neighboring triple information around both the head and tail nodes via graph networks; and

a comprehensive evaluation scorer configured to provide a comprehensive score based on global noise detection and local noise detection from the global noise detector and the local noise detector, respectively, to indicate a noise level for the CSKG information.

2 . The system according to claim 1 , wherein the triple encoder utilizes a transformer-based pretrained language model to encode natural language descriptions of nodes and relations within the CSKG information into sentence embeddings.

3 . The system according to claim 2 , wherein the triple encoder is further configured to transform original input data of the user inputs into the structured triples using the sentence embeddings, and each of the structured triples consists of a head entity, a relation, and a tail entity.

4 . The system according to claim 1 , wherein the global noise detector comprises a recurrent neural network (RNN) employed as the neutralized rule encoder.

5 . The system according to claim 1 , wherein the local noise detector is further configured to adopt a graph neural network to measure similarity of aggregated semantic information of the neighboring triples of the head and tail nodes using a graph attention network.

6 . The system according to claim 1 , wherein the comprehensive evaluation scorer is configured to take global energy information output from the global noise detector and local energy information output from the local noise detector as its inputs.

7 . The system according to claim 6 , further comprising:

a comprehensive score processing module configured to receive final comprehensive score information from the comprehensive evaluation scorer and perform one or more functions in response to the final comprehensive score information.

8 . The system according to claim 7 , wherein the comprehensive score processing module comprises an automatic filtering submodule used to configure thresholds and automatically delete noise data with scores below the threshold.

9 . A method using global and local-aware denoising framework for commonsense knowledge graph noise detection, comprising:

receiving, by an input receiver serving as a user interface, user inputs to process the user inputs, so as to obtain commonsense knowledge graph (CSKG) information;

receiving, by a triple encoder, the CSKG information from the input receiver;

extracting, by the triple encoder, semantic information from free-text formatted nodes in the CSKG information, so as to obtain structured triples;

receiving, by a global noise detector, the structured triples output from the triple encoder;

analyzing and detecting, by the global noise detector, noise at a global level from the CSKG information, wherein the global noise detector identifies high-frequency patterns through rule mining and uses a rule encoder to generalize learned rules and guide noise detection;

receiving, by a local noise detector, the structured triples output from the triple encoder;

analyzing and detecting, by the local noise detector, noise at a local level from the CSKG information, wherein the local noise detector utilizes the structured triples output by the triple encoder to aggregate neighboring triple information around both the head and tail nodes via graph networks;

providing, by a comprehensive evaluation scorer, a comprehensive score based on global noise detection and local noise detection from the global noise detector and the local noise detector, respectively, so as to indicate a noise level for the CSKG information.

10 . The method according to claim 9 , wherein the triple encoder utilizes a transformer-based pretrained language model to encode natural language descriptions of nodes and relations within the CSKG information into sentence embeddings.

11 . The method according to claim 10 , further comprising:

transforming, by the triple encoder, original input data of the user inputs into the structured triples using the sentence embeddings, wherein each of the structured triples consists of a head entity, a relation, and a tail entity.

12 . The method according to claim 9 , wherein the global noise detector comprises a recurrent neural network (RNN) employed as the neutralized rule encoder.

13 . The method according to claim 9 , further comprising:

adopting, by the local noise detector, a graph neural network to measure similarity of aggregated semantic information of the neighboring triples of the head and tail nodes using a graph attention network.

14 . The method according to claim 9 , further comprising:

taking, by the comprehensive evaluation scorer, global energy information output from the global noise detector and local energy information output from the local noise detector as its inputs.

15 . The method according to claim 14 , further comprising:

receiving, by a comprehensive score processing module, final comprehensive score information from the comprehensive evaluation scorer; and

performing, by the comprehensive score processing module, one or more functions in response to the final comprehensive score information.

16 . The method according to claim 15 , wherein the comprehensive score processing module comprises an automatic filtering submodule used to configure thresholds and automatically delete noise data with scores below the threshold.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Sep 18, 2024
From: DENG, ZHEYE; WANG, WEIQI; WANG, ZHAOWEI; SONG, YANGQIU
To: THE HONG KONG UNIVERSITY OF SCIENCE AND TECHNOLOGY
Reel/Frame 068616/0588 →
Continuity (2)
Provisional Application 63588299 · Oct 6, 2023
Related Publication 20250117572A1 · Apr 10, 2025
References Cited (84)
US 9251228B1 · Iyer · 2016 [cited by examiner]
US 11693896B2 · Mihindukulasooriya et al. · 2023 [cited by applicant]
US 11714832B2 · Cervantes et al. · 2023 [cited by applicant]
US 20200097861A1 · Mihindukulasooriya · 2020 [cited by examiner]
US 20220179857A1 · Kompella et al. · 2022 [cited by applicant]
CN 114860945A · 2022 [cited by applicant]
CN 115587583A · 2023 [cited by applicant]
Bo An et al., “Accurate Text-Enhanced Knowledge Graph Representation Learning,” Proceedings of NAACL-HLT 2018, Association for Computational Linguistics, 2018, p. 745-755. [cited by applicant]
Molood Barati et al., “Mining semantic association rules from RDF data,” Knowledge-Based Systems, ScienceDirect, 2017, vol. 133, p. 183-196. [cited by applicant]
Caleb Belth et al., “What is Normal, What is Strange, and What is Missing in a Knowledge Graph: Unified Characterization via Inductive Summarization,” WWW '20, IW3C2, ACM, 2020. [cited by applicant]
Chandra Bhagavatula et al., “Abductive Commonsense Reasoning,” ICLR 2020, arXiv: 1908.05739v2 2020, p. 1-18. [cited by applicant]
Yonatan Bisk et al., “PIQA: Reasoning about Physical Commonsense in Natural Language,” The Thirty-Fourth AAAI Conference on Artificial Intelligence (AAAI-20), Association for the Advancement of Artificial Intelligence, … [cited by applicant]
Antoine Bordes et al., “Translating Embeddings for Modeling Multi-relational Data,” NIPS'13: Proceedings of the 26th International Conference on Neural Information Processing Systems, ACM Digital Library, 2013, vol. 2, … [cited by applicant]
Antoine Bosselut et al., “Dynamic Neuro-Symbolic Knowledge Graph Construction for Zero-Shot Commonsense Question Answering,” The Thirty-Fifth AAAI Conference on Artificial Intelligence (AAAI-21), Association for the Adv… [cited by applicant]
Antoine Bosselut et al., “COMET: Commonsense Transformers for Automatic Knowledge Graph Construction,” Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, Association for Computation… [cited by applicant]
Tom B. Brown et al., “Language Models are Few-Shot Learners,” arXiv: 2005.14165v4, 2020, p. 1-75. [cited by applicant]
Jiangjie Chen et al., “Say What You Mean! Large Language Models Speak Too Positively about Negative Commonsense Knowledge,” Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics, Associ… [cited by applicant]
Kewei Cheng et al., “Neural Compositional Rule Learning for Knowledge Graph Reasoning,” ICLR 2023, arXiv: 2303.03581v1, 2023, p. 1-18. [cited by applicant]
Kewei Cheng et al., “RLogic: Recursive Logical Rule Learning from Knowledge Graphs,” KDD '22, ACM, 2022. [cited by applicant]
Ernest Davis et al., “Commonsense Reasoning and Commonsense Knowledge in Artificial Intelligence,” Communications of the ACM, 2015, vol. 58(9), p. 92-103. [cited by applicant]
Tianqing Fang et al., “Benchmarking Commonsense Knowledge Base Population with an Effective Evaluation Dataset,” Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, Association for Co… [cited by applicant]
Tianqing Fang et al., “DISCOS: Bridging the Gap between Discourse Knowledge and Commonsense Knowledge,” WWW'21, IW3C2, ACM, 2021. [cited by applicant]
Luis Galárraga et al., “Fast rule mining in ontological knowledge bases with AMIE+,” The VLDB Journal, 2015, vol. 24, p. 707-730. [cited by applicant]
Luis Galárraga et al., “AMIE: Association Rule Mining under Incomplete Evidence in Ontological Knowledge Bases,” WWW'13, IW3C2, ACM, 2013. [cited by applicant]
Congcong Ge et al., “KGClean: An Embedding Powered Knowledge Graph Cleaning Framework,” arXiv: 2004.14478v1, 2020. [cited by applicant]
Shu Guo et al., “Jointly Embedding Knowledge Graphs and Logical Rules,” Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, Association for Computational Linguistics, 2016, p. 192-202. [cited by applicant]
Shu Guo et al., “Knowledge Graph Embedding with Iterative Guidance from Soft Rules,” The Thirty-Second AAAI Conference on Artificial Intelligence (AAAI-18), Association for the Advancement of Artificial Intelligence, 20… [cited by applicant]
Xu Han et al., “OpenKE: An Open Toolkit for Knowledge Embedding,” Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing (System Demonstrations), Association for Computational Linguistics… [cited by applicant]
Mutian He et al., “Acquiring and Modelling Abstract Commonsense Knowledge via Conceptualization,” Artificial Intelligence, ScienceDirect, 2024, vol. 333, 104149, p. 1-36. [cited by applicant]
Pengcheng He et al., “DEBERTAV3: Improving Deberta Using Electra-Style Pre-Training With Gradientdisentangled Embedding Sharing,” ICLR 2023, arXiv: 2111.09543v4, 2023, p. 1-16. [cited by applicant]
Shengbin Jia et al., “Triple Trustworthiness Measurement for Knowledge Graph.” WWW '19, IW3C2, ACM, 2019, p. 2865-2871. [cited by applicant]
Yu Jin Kim et al., “Modularized Transfer Learning with Multiple Knowledge Graphs for Zero-shot Commonsense Reasoning,” Proceedings of the 2022 Conference of the North American Chapter of the Association for Computationa… [cited by applicant]
Diederik P. Kingma et al., “ADAM: a Method for Stochastic Optimization,” ICLR 2015, arXiv: 1412.6980v5, 2015, p. 1-13. [cited by applicant]
Jonathan Lajus et al., “Fast and Exact Rule Mining with AMIE 3,” ESWC 2020, LNCS, 2020, vol. 12123, p. 36-52. [cited by applicant]
Xiang Li et al., “Commonsense Knowledge Base Completion,” Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics, Association for Computational Linguistics, 2016, p. 1445-1455. [cited by applicant]
Yankai Lin et al., “Modeling Relation Paths for Representation Learning of Knowledge Bases,” Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, Association for Computational Linguist… [cited by applicant]
H. Liu et al., “ConceptNet—a practical commonsense reasoning tool-kit,” BT Technology Journal, 2004, vol. 22(4), p. 211-226. [cited by applicant]
Rui Liu et al., “Enhancing Zero-shot and Few-shot Stance Detection with Commonsense Knowledge Graph,” Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, Association for Computational Linguistics… [cited by applicant]
Yinhan Liu et al., “ROBERTa: A Robustly Optimized BERT Pretraining Approach,” arXiv: 1907.11692v1, 2019. [cited by applicant]
Kaixin Ma et al., “Knowledge-driven Data Construction for Zero-shot Evaluation in Commonsense Question Answering,” The Thirty-Fifth AAAI Conference on Artificial Intelligence (AAAI-21), Association for the Advancement o… [cited by applicant]
Chaitanya Malaviya et al., “Commonsense Knowledge Base Completion with Structural and Semantic Context,” The Thirty-Fourth AAAI Conference on Artificial Intelligence (AAAI-20), Association for the Advancement of Artific… [cited by applicant]
Elan Markowitz et al., “StATIK: Structure and Text for Inductive Knowledge Graph Completion,” Findings of the Association for Computational Linguistics: NAACL 2022, Association for Computational Linguistics, 2022, p. 60… [cited by applicant]
Nasrin Mostafazadeh et al., “GLUCOSE: Generalized and Contextualized Story Explanations,” Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, Association for Computational Linguistics… [cited by applicant]
Stephen Muggleton et al., “Inductive Logic Programming: Theory and Methods,” The Journal of Logic Programming, 1994, vol. 19-20, p. 629-679. [cited by applicant]
Jianmo Ni et al., “Sentence-T5 (ST5): Scalable Sentence Encoders from Pre-trained Text-to-Text Models,” Findings of the Association for Computational Linguistics: ACL 2022, Association for Computational Linguistics, 202… [cited by applicant]
Openai, “ChatGPT: Optimizing Language Models for Dialogue” URL: https://chatgpt.r4wand.eu.org/ Downloaded on Nov. 11, 2024. [cited by applicant]
Long Ouyang et al., “Training language models to follow instructions with human feedback,” NIPS'22: Proceedings of the 36th International Conference on Neural Information Processing Systems, ACM Digital Library, 2022, N… [cited by applicant]
Meng Qu et al., “Rnnlogic: Learning Logic Rules for Reasoning on Knowledge Graphs,” ICLR 2021, arXiv: 2010.04029v2, 2021, p. 1-21. [cited by applicant]
Julien Romero et al., “Mapping and Cleaning Open Commonsense Knowledge Bases with Generative Translation,” International Semantic Web Conference 2023 (ISWC), HAL open science, 2023, p. 368-387. [cited by applicant]
Ali Sadeghian et al., “DRUM: End-To-End Differentiable Rule Mining on Knowledge Graphs,” Proceedings of the 33rd International Conference on Neural Information Processing Systems, ACM Digital Library, 2019, No. 1375, p.… [cited by applicant]
Tara Safavi et al., “NEGATER: Unsupervised Discovery of Negatives in Commonsense Knowledge Bases,” Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, Association for Computational Li… [cited by applicant]
Keisuke Sakaguchi et al., “WinoGrande: An Adversarial Winograd Schema Challenge at Scale,” Communications of the ACM, 2021, vol. 64(9), p. 99-106. [cited by applicant]
Maarten Sap et al., “ATOMIC: An Atlas of Machine Commonsense for If-Then Reasoning,” The Thirty-Third AAAI Conference on Artificial Intelligence (AAAI-19), Association for the Advancement of Artificial Intelligence, 201… [cited by applicant]
Maarten Sap et al., “SOCIAL IQA: Commonsense Reasoning about Social Interactions,” Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Nat… [cited by applicant]
Mike Schuster et al., “Bidirectional Recurrent Neural Networks,” IEEE Transactions on Signal Processing, IEEE, 1997, vol. 45(11), p. 2673-2681. [cited by applicant]
Jianhao Shen et al., “Joint Language Semantic and Structure Embedding for Knowledge Graph Completion,” Proceedings of the 29th International Conference on Computational Linguistics, 2022, p. 1965-1978. [cited by applicant]
Vered Shwartz et al., “Unsupervised Commonsense Question Answering with Self-Talk,” Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, Association for Computational Linguistics, 2020… [cited by applicant]
Robyn Speer et al., “ConceptNet 5.5: An Open Multilingual Graph of General Knowledge,” Proceedings of the Thirty-First AAA! Conference on Artificial Intelligence (AAAI-17), Association for the Advancement of Artificial … [cited by applicant]
Ying Su et al., “MICO: A Multi-alternative Contrastive Learning Framework for Commonsense Knowledge Representation,” Findings of the Association for Computational Linguistics: EMNLP 2022, Association for Computational L… [cited by applicant]
Zhiqing Sun et al., “Rotate: Knowledge Graph Embedding by Relational Rotation in Complex Space,” ICLR 2019, arXiv: 1902.10197v1, 2019. P. 1-18. [cited by applicant]
Alon Talmor et al., “Commonsenseqa: A Question Answering Challenge Targeting Commonsense Knowledge,” Proceedings of NAACL-HLT 2019, Association for Computational Linguistics, 2019, p. 4149-4158. [cited by applicant]
Niket Tandon et al., “Acquiring Comparative Commonsense Knowledge from the Web,” Proceedings of the Twenty-Eighth AAAI Conference on Artificial Intelligence, Association for the Advancement of Artificial Intelligence, 2… [cited by applicant]
Niket Tandon et al., “WebChild 2.0: Fine-Grained Commonsense Knowledge Distillation,” Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics-System Demonstrations, Association for Comput… [cited by applicant]
Théo Trouillon et al., “Complex Embeddings for Simple Link Prediction,” Proceedings of the 33rd International Conference on Machine Learning, JMLR: W&CP, 2016, vol. 48. [cited by applicant]
Petar Veličković et al., “Graph Attention Networks,” ICLR 2018, arXiv: 1710.10903v3, 2018, p. 1-12. [cited by applicant]
Bin Wang et al., “Inductive Learning on Commonsense Knowledge Graph Completion,” 2021 International Joint Conference on Neural Networks (IJCNN), IEEE, 2021. [cited by applicant]
Bo Wang et al., “Structure-Augmented Text Representation Learning for Efficient Knowledge Graph Completion,” WWWW'21, IW3C2, ACM, 2021, p. 1737-1748. [cited by applicant]
Weiqi Wang et al., “CAR: Conceptualization-Augmented Reasoner for Zero-Shot Commonsense Question Answering,” Findings of the Association for Computational Linguistics: EMNLP 2023, Association for Computational Linguisti… [cited by applicant]
Zhigang Wang et al., “Text-Enhanced Representation Learning for Knowledge Graph,” Proceedings of the Twenty-Fifth International Joint Conference on Artificial Intelligence (IJCAI-16), ACM Digital Library, 2016, p. 1293-… [cited by applicant]
Jason Wei et al., “Chain of Thought Prompting Elicits Reasoning in Large Language Models,” arXiv: 2201.11903v1, 2022. [cited by applicant]
Peter West et al., “Symbolic Knowledge Distillation: from General Language Models to Commonsense Models,” Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics… [cited by applicant]
Thomas Wolf et al., “Transformers: State-of-the-Art Natural Language Processing,” Proceedings of the 2020 EMNLP (Systems Demonstrations), Association for Computational Linguistics, 2020, p. 38-45. [cited by applicant]
Ruobing Xie et al., “Does William Shakespeare REALLY Write Hamlet? Knowledge Representation Learning with Confidence,” The Thirty-Second AAAI Conference on Artificial Intelligence (AAAI-18), Association for the Advancem… [cited by applicant]
Bishan Yang et al., “Embedding Entities and Relations for Learning and Inference in Knowledge Bases,” ICLR 2015, arXiv: 1412.6575v4, 2015, p. 1-12. [cited by applicant]
Fan Yang et al., “Differentiable Learning of Logical Rules for Knowledge Base Reasoning,” NIPS'17: Proceeding of the 31st Conference on Neural Information Processing Systems, ACM Digital Library, 2017, p. 2316-2325. [cited by applicant]
Liang Yao et al., “KG-BERT: BERT for Knowledge Graph Completion,” Association for the Advancement of Artificial Intelligence, arXiv: 1909.03193v2, 2019. [cited by applicant]
Michihiro Yasunaga et al., “Deep Bidirectional Language-Knowledge Graph Pretraining,” 36th Conference on Neural Information Processing Systems (NeurIPS 2022), arXiv: 2210.09338v2, 2022, p. 1-19. [cited by applicant]
Changlong Yu et al., “FolkScope: Intention Knowledge Graph Construction for Discovering E-commerce Commonsense,” arXiv: 2211.08316v1, 2022. [cited by applicant]
Hengtong Zhang et al., “Data Poisoning Attack against Knowledge Graph Embedding.” Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence (IJCAI-19), 2019, p. 4853-4859. [cited by applicant]
Hongming Zhang et al., “ASER: Towards Large-scale Commonsense Knowledge Acquisition via Higher-order Selectional Preference over Eventualities,” Artificial Intelligence, ScienceDirect, 2022, vol. 309, 103740, p. 1-43. [cited by applicant]
Houyu Zhang et al., “Grounded Conversation Generation as Guided Traverses in Commonsense Knowledge Graphs,” Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, Association for Comput… [cited by applicant]
Qinggang Zhang et al., “Contrastive Knowledge Graph Error Detection,” CIKM '22, Association for Computing Machinery, ACM, 2022. [cited by applicant]
Wen Zhang et al., “Iteratively Learning Embeddings and Rules for Knowledge Graph Reasoning,” WWW '19, IW3C2, ACM, 2019. [cited by applicant]
Jiangtao Ma et al., “PTrustE: A high-accuracy knowledge graph noise detection method based on path trustworthiness and triple embedding,” Knowledge-Based Systems, ScienceDirect, 2022, vol. 256, 109688, p. 1-14. [cited by applicant]