@article{
 title = {Semi-supervised clustering with inaccurate pairwise annotations},
 type = {article},
 year = {2022},
 pages = {441-457},
 volume = {607},
 websites = {https://arxiv.org/pdf/2104.02146.pdf},
 id = {0330d47e-9f29-373e-ab2f-5783b08a9ae5},
 created = {2021-04-07T13:22:58.049Z},
 file_attached = {true},
 profile_id = {5e3d1dc4-cb58-3af5-aff1-4d943d2eaf6a},
 last_modified = {2022-09-09T13:57:33.625Z},
 read = {true},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 citation_key = {Gribel2021},
 private_publication = {false},
 abstract = {Pairwise relational information is a useful way of providing partial supervision in domains where class labels are difficult to acquire. This work presents a clustering model that incorporates pairwise annotations in the form of must-link and cannot-link relations and considers possible annotation inaccuracies (i.e., a common setting when experts provide pairwise supervision). We propose a generative model that assumes Gaussian-distributed data samples along with must-link and cannot-link relations generated by stochastic block models. We adopt a maximum-likelihood approach and demonstrate that, even when supervision is weak and inaccurate, accounting for relational information significantly improves clustering performance. Relational information also helps to detect meaningful groups in real-world datasets that do not fit the original data-distribution assumptions. Additionally, we extend the model to integrate prior knowledge of experts' accuracy and discuss circumstances in which the use of this knowledge is beneficial.},
 bibtype = {article},
 author = {Gribel, D. and Gendreau, M. and Vidal, T.},
 journal = {Information Sciences},
 number = {C}
}

{"_id":"RNQTYEvirYinRZ2aE","bibbaseid":"gribel-gendreau-vidal-semisupervisedclusteringwithinaccuratepairwiseannotations-2022","author_short":["Gribel, D.","Gendreau, M.","Vidal, T."],"bibdata":{"title":"Semi-supervised clustering with inaccurate pairwise annotations","type":"article","year":"2022","pages":"441-457","volume":"607","websites":"https://arxiv.org/pdf/2104.02146.pdf","id":"0330d47e-9f29-373e-ab2f-5783b08a9ae5","created":"2021-04-07T13:22:58.049Z","file_attached":"true","profile_id":"5e3d1dc4-cb58-3af5-aff1-4d943d2eaf6a","last_modified":"2022-09-09T13:57:33.625Z","read":"true","starred":false,"authored":"true","confirmed":"true","hidden":false,"citation_key":"Gribel2021","private_publication":false,"abstract":"Pairwise relational information is a useful way of providing partial supervision in domains where class labels are difficult to acquire. This work presents a clustering model that incorporates pairwise annotations in the form of must-link and cannot-link relations and considers possible annotation inaccuracies (i.e., a common setting when experts provide pairwise supervision). We propose a generative model that assumes Gaussian-distributed data samples along with must-link and cannot-link relations generated by stochastic block models. We adopt a maximum-likelihood approach and demonstrate that, even when supervision is weak and inaccurate, accounting for relational information significantly improves clustering performance. Relational information also helps to detect meaningful groups in real-world datasets that do not fit the original data-distribution assumptions. Additionally, we extend the model to integrate prior knowledge of experts' accuracy and discuss circumstances in which the use of this knowledge is beneficial.","bibtype":"article","author":"Gribel, D. and Gendreau, M. and Vidal, T.","journal":"Information Sciences","number":"C","bibtex":"@article{\n title = {Semi-supervised clustering with inaccurate pairwise annotations},\n type = {article},\n year = {2022},\n pages = {441-457},\n volume = {607},\n websites = {https://arxiv.org/pdf/2104.02146.pdf},\n id = {0330d47e-9f29-373e-ab2f-5783b08a9ae5},\n created = {2021-04-07T13:22:58.049Z},\n file_attached = {true},\n profile_id = {5e3d1dc4-cb58-3af5-aff1-4d943d2eaf6a},\n last_modified = {2022-09-09T13:57:33.625Z},\n read = {true},\n starred = {false},\n authored = {true},\n confirmed = {true},\n hidden = {false},\n citation_key = {Gribel2021},\n private_publication = {false},\n abstract = {Pairwise relational information is a useful way of providing partial supervision in domains where class labels are difficult to acquire. This work presents a clustering model that incorporates pairwise annotations in the form of must-link and cannot-link relations and considers possible annotation inaccuracies (i.e., a common setting when experts provide pairwise supervision). We propose a generative model that assumes Gaussian-distributed data samples along with must-link and cannot-link relations generated by stochastic block models. We adopt a maximum-likelihood approach and demonstrate that, even when supervision is weak and inaccurate, accounting for relational information significantly improves clustering performance. Relational information also helps to detect meaningful groups in real-world datasets that do not fit the original data-distribution assumptions. Additionally, we extend the model to integrate prior knowledge of experts' accuracy and discuss circumstances in which the use of this knowledge is beneficial.},\n bibtype = {article},\n author = {Gribel, D. and Gendreau, M. and Vidal, T.},\n journal = {Information Sciences},\n number = {C}\n}","author_short":["Gribel, D.","Gendreau, M.","Vidal, T."],"urls":{"Website":"https://arxiv.org/pdf/2104.02146.pdf"},"biburl":"https://bibbase.org/service/mendeley/1465671","bibbaseid":"gribel-gendreau-vidal-semisupervisedclusteringwithinaccuratepairwiseannotations-2022","role":"author","metadata":{"authorlinks":{}}},"bibtype":"article","biburl":"https://bibbase.org/service/mendeley/1465671","dataSources":["yinfondEAJRbDM9sJ","sempRA6PhmAdGk3yG","2252seNhipfTmjEBQ"],"keywords":[],"search_terms":["semi","supervised","clustering","inaccurate","pairwise","annotations","gribel","gendreau","vidal"],"title":"Semi-supervised clustering with inaccurate pairwise annotations","year":2022}