Knowledge Representation and Reasoning

1. Knowledge Representation and Reasoning Jie Tang Department of Computer Science and Technology Tsinghua University • The slides can be downloaded at http://keg.cs.tsinghua.edu.cn/jietang

2. Knowledge Graph • “Knowledge graph” was used by Google 2012 • Knowledge engineering, expert system • CYC: the world's longest-lived AI project (1985)

3. Knowledge & AI * Two keys for AI: Knowledge Base+Intelligent Algorithm TimBernersLee Father of WWW Turing Award EdwardFeigenbaum Father of KB Turing Award Big Data Knowledge Intelligence [1] J. Tang, J. Zhang, L. Yao, J. Li, L. Zhang, and Z. Su. ArnetMiner: Extraction and Mining of Academic Social Networks. KDD’08. pp.990-998.

4. Example: Multi-hopQA Quality Café LosAngeles AlessandroMoschitti Alessandro Moschitti is a professor of the CS Department of the University of Trento, Italy. He is currently a Principal Research Scientist of the Qatar Computing Research Institute (QCRI) The Quality Cafe is a now-defunct diner in Los Angeles, California. The restaurant has appeared as a location featured in a number of Hollywood films, including Old School, Gone in 60 Seconds, ... Los Angeles is the most populous city in California, the second most populous city in the United States, after New York City, and the third most populous city in North America. ToddPhillips OldSchool TsinghuaUniversity Todd Phillips is an American director, producer, screenwriter, and actor. He is best known for writing and directing films, including Road Trip (2000), Old School (2003), Starsky & Hutch (2004), and The Hangover Trilogy. Tsinghua University is a major research university in Beijing and dedicated to academic excellenceand global development. Tsinghua is perennially ranked as one of the top academic institutions in China, Asia, and worldwide... Old School is a 2003 American comedy film released by DreamWorks Pictures and The Montecito Picture Company and directed by Todd Phillips.

5. Relatedwork:DrQAframework • DrQA[1]proposedapopulartwo-stageframeworkforopen-domainQA. • DocumentRetriever: • Retrieve 5 relatedWikipedia articles given any question. • DocumentReader: • Predictthespanofanswerwith(usuallydeeplearning)models. Chen, Danqi, et al. "Reading wikipedia to answer open-domain questions." arXiv preprint arXiv:1704.00051 (2017).

6. Challenge1:MyopicRetrievalProblem DocumentA Old School is a 2003American comedy film released by DreamWorks Pictures and The Montecito Picture Company… DocumentB ManydirectorsshootscenesinHollywood,Los Angeles,whichisnotable as the home of theU.S.filmindustry. document B share more common words with the question.

7. Relatedwork:ReadingComprehensionmodels • TheDocumentReadersinDrQAframeworkaretargetedatsingle-paragraphreadingcomprehensionQA,whosebenchmarkisSQuAD[1]dataset. BiDAF BERT XLNet Rajpurkar, Pranav, et al. SQuAD: 100,000+ Questions for Machine Comprehension of Text. EMNLP’’16.

8. Challenge2:Explainability • MostRCmodelscanonlybeseenasblack-boxes: • For verification,theuser needs: • reasoningpath(graph) • supportingfacts(sentences)ateachhop • Otherpossibleanswersandtheirreasoningpathsforcomparison

9. Ideal Results

10. Cognitive Psychology Framework System 1 Intuitive System 2 Analytic Dual Process Theory Jonathan St BT Evans. 1984. Heuristic and analytic processes in reasoning. British Journal of Psychology, 75(4):451–468.

11. How to do the reasoning?

12. Cognitive Graph • System1: • Knowledge expansionbyassociationintextwhenreading • System2: • Decision making w/ all the information System 1 Intuitive System 2 Analytic

13. CogQA: Cognitive Graph for QA • Aniterativeframeworkcorrespondingtodualprocesstheory • System1 • extractentitiestobuildthecognitivegraph • generatesemanticvectorsforeachnode • System2 • Doreasoningbasedonsemanticvectorsandgraph • FeedcluestoSystem1toextractnext-hopentities Question System1 extract CognitiveGraph input Qualitycafé LosAngeles …location featured in a number of Hollywood films, including Old School, Gone in 60 Seconds… clues Gonein60seconds Oldschool input System2 DominicSena ToddPhillips predict

14. Cognitive Graph: DL + Dual Process Theory explicit decision implicit knowledge expansion • M. Ding, C. Zhou, Q. Chen, H. Yang, and J. Tang. Cognitive Graph for Multi-Hop Reading Comprehension at Scale. ACL’19.

15. System1:BERTImplementation • Extracttop-knext-hopentitiesandanswercandidatesrespectively • Predictthestartandendprobabilitiesofeachposition • Generatesemanticvectorsforentitiesbasedontheirdocuments • Takethe0-thprobabilityasnegativethreshold • Ignorethespanswhosestartprobabilitiesaresmallthanthenegativethreshold

16. System2:theGNNimplementation Ateachstep,hiddenrepresentationsXfornodesareupdatedaccordingtothepropagationrules: PredictorFisatwo-layerMLP,whichpredictsthefinalanswerbasedonhiddenrepresentationsX:

17. CogQAAlgorithmandTraining • Preprocess • Markthespansofnext-hopentitiesandanswerbyfuzzymatching • Goldencognitivegraphandnegative(hopandspan)nodes • TrainTask#1(System1)： • Next-hopspansandanswerspansextraction • TrainTask#2(System1&2)： • Graphconvolutionandanswerprediction

18. Performance • HotpotQAisadatasetwithleaderboardsimilartoSQuAD • CogQAranked1stfrom21,Febto15,May(nearly3month) ** Code available at https://github.com/THUDM/CogQA

19. Reasoning Power CogQAPerformsmuchbetteronquestionwithmorehops！

20. CaseStudy • Tree-shapeCognitiveGraph • Userscanverifytheanswerbycomparingitwithanotherpossiblereasoningchain. • “UpperHouse”inthequestionissimilarto“Senate”not“HouseofRepresentative”

21. CaseStudy • DAG-shapeCognitiveGraph • Multiplesupportingfactsprovidesricherinformation,increasingthecredibilityoftheanswer.

22. CaseStudy • Inthiscase,CogQAgivestheanswer“MarijusAdomaitis” while the annotated ground truth is “Ten Walls”. • Byexaminingthecognitivegraph,wefindthatTen WallsisjustthestagenameofMarijusAdomaitis! • Withoutcognitivegraphs,black-boxmodelscannotachieveit.

23. CogDL—A Toolkit for ​Deep Learning on Graphs ** Code available at https://keg.cs.tsinghua.edu.cn/cogdl/

24. CogDL—A Toolkit for ​Deep Learning on Graphs

25. CogDLfeatures • CogDL supports these features: • Sparsification: fast network embedding on large-scale networks with tens of millions of nodes • Arbitrary: dealing with different graph structures: attributed, multiplex and heterogeneous networks • Distributed: multi-GPU training on one machine or across multiple machines • Extensible: easily register new datasets, models, criterions and tasks • Homepage:https://keg.cs.tsinghua.edu.cn/cogdl/

26. CogDL Leaderboards Multi-label Node Classification Thefollowingleaderboardisbuilt from unsupervised learningwithmulti-label node classification setting. http://keg.cs.tsinghua.edu.cn/cogdl/node-classification.html

27. CogDL Leaderboards Node Classification with Attributes. http://keg.cs.tsinghua.edu.cn/cogdl/node-classification.html

28. Leaderboards: Link Prediction http://keg.cs.tsinghua.edu.cn/cogdl/link-prediction.html

29. Joinus • Feelfreetojoinuswiththethreefollowingways: • addyourdataintotheleaderboard • addyouralgorithmintotheleaderboard • addyouralgorithmintothetoolkit • Notethat,inordertocollaborate withus, youmayreferto the documentation in our homepage.

30. Cognitive Graph: DL + Dual Process Theory explicit decision implicit knowledge expansion • M. Ding, C. Zhou, Q. Chen, H. Yang, and J. Tang. Cognitive Graph for Multi-Hop Reading Comprehension at Scale. ACL’19.

31. Related Publications • Ming Ding, Chang Zhou, Qibin Chen, Hongxia Yang, and Jie Tang. Cognitive Graph for Multi-Hop Reading Comprehension at Scale. ACL’19. • Jie Zhang, Yuxiao Dong, Yan Wang, Jie Tang, and Ming Ding. ProNE: Fast and Scalable Network Representation Learning. IJCAI’19. • Yukuo Cen, Xu Zou, Jianwei Zhang, Hongxia Yang, Jingren Zhou and Jie Tang. Representation Learning for Attributed Multiplex Heterogeneous Network. KDD’19. • Fanjin Zhang, Xiao Liu, Jie Tang, Yuxiao Dong, Peiran Yao, Jie Zhang, Xiaotao Gu, Yan Wang, Bin Shao, Rui Li, and Kuansan Wang. OAG: Toward Linking Large-scale Heterogeneous Entity Graphs. KDD’19. • Yifeng Zhao, Xiangwei Wang, Hongxia Yang, Le Song, and Jie Tang. Large Scale Evolving Graphs with Burst Detection. IJCAI’19. • Yu Han, Jie Tang, and Qian Chen. Network Embedding under Partial Monitoring for Evolving Networks. IJCAI’19. • JiezhongQiu, Yuxiao Dong, Hao Ma, Jian Li, Chi Wang, Kuansan Wang, and Jie Tang. NetSMF: Large-Scale Network Embedding as Sparse Matrix Factorization. WWW'19. • JiezhongQiu, Jian Tang, Hao Ma, Yuxiao Dong, Kuansan Wang, and Jie Tang. DeepInf: Modeling Influence Locality in Large Social Networks. KDD’18. • JiezhongQiu, Yuxiao Dong, Hao Ma, Jian Li, Kuansan Wang, and Jie Tang. Network Embedding as Matrix Factorization: Unifying DeepWalk, LINE, PTE, and node2vec. WSDM’18. • Jie Tang, Jing Zhang, Limin Yao, Juanzi Li, Li Zhang, and Zhong Su. ArnetMiner: Extraction and Mining of Academic Social Networks. KDD’08. • For more, please check here http://keg.cs.tsinghua.edu.cn/jietang

32. Thank you！ Collaborators: Ming Ding, Qibin Chen, et al. (THU) Hongxiao Yang, Chang Zhou (Alibaba) • Jie Tang, KEG, Tsinghua U http://keg.cs.tsinghua.edu.cn/jietang • Download all data & Codeshttps://keg.cs.tsinghua.edu.cn/cogdl/