Supporting Collaborative Modeling via Natural Language Processing - PowerPoint PPT Presentation

. . . . . . . . . . . . . . . . Supporting Collaborative Modeling via Natural Language Processing Fatma Başak Aydemir 1 Fabiano Dalpiaz 2 1 Boğaziçi University 2 Utrecht University . . . . . . . . . . . . . . . . . . . . . . . . 39 th International Conference on Conceptual Modeling

. . . . . . . . . . . . . . . . Collaborative Modeling Various F.B. Aydemir Supporting Collaborative Modeling via Natural Language Processing ER 2020 2 . . . . . . . . . . . . . . . . . . . . . . . . 19 • Experts • Modeling Languages • Locations • Time Zones

. . . . . . . . . . . . . . Concept Suggestion Service . A web service integrated to the modeling environments concepts modelers F.B. Aydemir Supporting Collaborative Modeling via Natural Language Processing ER 2020 3 . 19 . . . . . . . . . . . . . . . . . . . . . . . . Concept Concept Concept Concept Modeler 2 Modeler 2 Modeler 2 Modeler 2 Modeler 1 Modeler 1 Modeler 1 Modeler 1 Suggester suggester suggester suggester O1. commitModel(m1) • Ignores the meta-models O2. analyzeCommit(m1) O1. commitModel(m2) • Analyzes the labels O3. requestSuggestions(m1) O2. analyzeCommit(m2) • Keeps track of the modelled O4. analyzeModels(m1, {m2}) O3. requestSuggestions(m2) O5. suggest(s1, m1) • Suggests missing concepts to O6. feedback(f1, s1) O4. analyzeModels(m2, {m1}) O5. suggest(s2, m2) O6. feedback(f2, s2) • Hides model details • Completeness • Common vocabulary

. . . . . . . . . . . . . . . . Service Operations . F.B. Aydemir Supporting Collaborative Modeling via Natural Language Processing ER 2020 4 . 19 . . . . . . . . . . . . . . . . . . . . . . {pt 1 , ..., pt q } {c 1 , ..., c n } Domain Project Model 1 terms model O2. Extract noun O4a. Identify O4b. Match terms phrases (terms) missing terms with concepts O1. Model 1 O3. Suggestions Candidate committed requested concepts {t 1 , ..., t n } {..., t j , ... } Model 1 Missing terms terms Legend Messages Control flow Data Activity object Data flow Standard Catching End

. . . . . . . . . . . . . . Matching Heuristics Possible matches for “thesis” Thesis Thesis Project Graduation project Coordinator F.B. Aydemir Supporting Collaborative Modeling via Natural Language Processing ER 2020 5 . . . . . . . . . . . . . . . . . . . . . . . . . . 19 • Exact match • Sub-string match • Similarity • Relatedness

. . . . . . . . . . . . . . Matching Heuristics Possible matches for “thesis” Thesis Project Graduation project Coordinator F.B. Aydemir Supporting Collaborative Modeling via Natural Language Processing ER 2020 5 . . . . . . . . . . . . . . . . . . . . . . . 19 . . . • Exact match • Thesis • Sub-string match • Similarity • Relatedness

. . . . . . . . . . . . . . . Matching Heuristics Possible matches for “thesis” Graduation project Coordinator F.B. Aydemir Supporting Collaborative Modeling via Natural Language Processing ER 2020 5 . . . . . . . . . . . . . . . . . . . . . . 19 . . . • Exact match • Thesis • Sub-string match • Thesis Project • Similarity • Relatedness

. . . . . . . . . . . . . . . Matching Heuristics Possible matches for “thesis” Coordinator F.B. Aydemir Supporting Collaborative Modeling via Natural Language Processing ER 2020 5 . . . . . . . . . . . . . . . . . . . . . . . . . 19 • Exact match • Thesis • Sub-string match • Thesis Project • Similarity • Graduation project • Relatedness

. . . . . . . . . . . . . . . . Matching Heuristics Possible matches for “thesis” F.B. Aydemir Supporting Collaborative Modeling via Natural Language Processing ER 2020 5 . . . . . . . . . . . . . . . . . . . . . . . . 19 • Exact match • Thesis • Sub-string match • Thesis Project • Similarity • Graduation project • Relatedness • Coordinator

. . . . . . . . . . . . . . . . Suggestion Heuristics F.B. Aydemir Supporting Collaborative Modeling via Natural Language Processing ER 2020 6 . . . . . . . . . . . . . . . . . . . . . . . . 19 • Parent • Child • Sibling

. . . . . . . . . . . . . . . . Suggestion Heuristics F.B. Aydemir Supporting Collaborative Modeling via Natural Language Processing ER 2020 7 . . . . . . . . . . . . . . . . . . . . . . . . 19

. . . . . . . . . . . . . . . . Filtering Heuristics F.B. Aydemir Supporting Collaborative Modeling via Natural Language Processing ER 2020 8 . . . . . . . . . . . . . . . . . . . . . . . . 19 • Fixed number • User feedback • Frequency • Limiting matches per missing item

. . . . . . . . . . . . . . . . Similarity of Compound Nouns Domain model based F.B. Aydemir Supporting Collaborative Modeling via Natural Language Processing ER 2020 9 . . . . . . . . . . . . . . . . . . . . . 19 . . . • Compound nouns for many concepts • Used in heuristics • Add detail to the models • Similarity check for common terminology • Explore the domain model • Two algorithms to calculate the similarity of a pair of compound nouns • WordNet and Word2Vecbased

. . . . . . . . . . . . . . . . Similarity of Compound Nouns F.B. Aydemir Supporting Collaborative Modeling via Natural Language Processing ER 2020 9 . . . . . . . . . . . . . . . . . . . . 19 . . . . • Compound nouns for many concepts • Used in heuristics • Add detail to the models • Similarity check for common terminology • Explore the domain model • Two algorithms to calculate the similarity of a pair of compound nouns • WordNet and Word2Vecbased • Domain model based

. . . . . . . . . . . . . . . WordNet and Word2Vec based Similarity Thesis Project, Graduation Project (1) F.B. Aydemir Supporting Collaborative Modeling via Natural Language Processing ER 2020 10 . . . . . . . . . . . . . . . . . . . . . . . . . 19 • Get the Word2Vec similarity scores of noun pairs of compound words • If the score is higher than a threshold check if they are synonyms using WordNet • Set the score to 1 for synonyms, leave as is otherwise • The similarity of the compounds is a weighted average of the pairs γ · sim(thesis, graduation) + δ · sim(project, project) + ϵ · sim(thesis, project) + κ · sim(project, graduation)

. . . . . . . . . . . . . . . . Domain Model based Similarity F.B. Aydemir Supporting Collaborative Modeling via Natural Language Processing ER 2020 11 . . . . . . . . . . . . . . . . . . . . . . . . 19 • Based on how well the compounds are matched in the domain model • Individual matching score for each compound is calculated • The similarity of the two term is the average of their scores

. . . . . . . . . . . . . . . . Experimental Setup to Detect Similarity scale F.B. Aydemir Supporting Collaborative Modeling via Natural Language Processing ER 2020 12 . . . . . . . . . . . . . . . . . . . . . . . . 19 • Used 20 pairs of 2-word compound nouns • Gold standard: surveyed people to assess the similarity on a 5-point Likert type • Compared the results with Bert-web and spaCy similarity

. P15 second phase P7 4 offjcial ceremony scientifjc paper P16 2 fjrst supervisor fjrst phase P6 3 graduation ceremony offjcial ceremony 1 2 relevant literature literature review P5 3 short proposal project proposal P14 1 project idea graduation project P4 3 information science computing science second presentation P17 1 department member 13 ER 2020 Supporting Collaborative Modeling via Natural Language Processing F.B. Aydemir 4 literature review research question P20 2 thesis report thesis topic P10 4 participation token P19 Google calendar 2 MBI colloquium MBI thesis P9 4 project facilitator company supervisor P18 2 graduation supervisor graduation ceremony P8 4 MBI colloquium P13 thesis topic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . MBI thesis project facilitator P3 3 second supervisor fjrst supervisor P12 1 student administration MBI student P2 3 second supervisor company supervisor P11 1 thesis project . P1 Cat. Second compound First compound Pair ID Cat. Second compound First compound Pair ID Pairs Used . . . . 19