Pub Date : 2023-03-03DOI: 10.48550/arXiv.2303.01844
Caglar Demir, A. N. Ngomo
Concept learning deals with learning description logic concepts from a background knowledge and input examples. The goal is to learn a concept that covers all positive examples, while not covering any negative examples. This non-trivial task is often formulated as a search problem within an infinite quasi-ordered concept space. Although state-of-the-art models have been successfully applied to tackle this problem, their large-scale applications have been severely hindered due to their excessive exploration incurring impractical runtimes. Here, we propose a remedy for this limitation. We reformulate the learning problem as a multi-label classification problem and propose a neural embedding model (NERO) that learns permutation-invariant embeddings for sets of examples tailored towards predicting $F_1$ scores of pre-selected description logic concepts. By ranking such concepts in descending order of predicted scores, a possible goal concept can be detected within few retrieval operations, i.e., no excessive exploration. Importantly, top-ranked concepts can be used to start the search procedure of state-of-the-art symbolic models in multiple advantageous regions of a concept space, rather than starting it in the most general concept $top$. Our experiments on 5 benchmark datasets with 770 learning problems firmly suggest that NERO significantly (p-value<1%) outperforms the state-of-the-art models in terms of $F_1$ score, the number of explored concepts, and the total runtime. We provide an open-source implementation of our approach.
{"title":"Learning Permutation-Invariant Embeddings for Description Logic Concepts","authors":"Caglar Demir, A. N. Ngomo","doi":"10.48550/arXiv.2303.01844","DOIUrl":"https://doi.org/10.48550/arXiv.2303.01844","url":null,"abstract":"Concept learning deals with learning description logic concepts from a background knowledge and input examples. The goal is to learn a concept that covers all positive examples, while not covering any negative examples. This non-trivial task is often formulated as a search problem within an infinite quasi-ordered concept space. Although state-of-the-art models have been successfully applied to tackle this problem, their large-scale applications have been severely hindered due to their excessive exploration incurring impractical runtimes. Here, we propose a remedy for this limitation. We reformulate the learning problem as a multi-label classification problem and propose a neural embedding model (NERO) that learns permutation-invariant embeddings for sets of examples tailored towards predicting $F_1$ scores of pre-selected description logic concepts. By ranking such concepts in descending order of predicted scores, a possible goal concept can be detected within few retrieval operations, i.e., no excessive exploration. Importantly, top-ranked concepts can be used to start the search procedure of state-of-the-art symbolic models in multiple advantageous regions of a concept space, rather than starting it in the most general concept $top$. Our experiments on 5 benchmark datasets with 770 learning problems firmly suggest that NERO significantly (p-value<1%) outperforms the state-of-the-art models in terms of $F_1$ score, the number of explored concepts, and the total runtime. We provide an open-source implementation of our approach.","PeriodicalId":91439,"journal":{"name":"Advances in intelligent data analysis. International Symposium on Intelligent Data Analysis","volume":"9 1","pages":"103-115"},"PeriodicalIF":0.0,"publicationDate":"2023-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84515096","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-02-14DOI: 10.48550/arXiv.2302.06967
Luis Galárraga
Knowledge graphs (KGs) are key tools in many AI-related tasks such as reasoning or question answering. This has, in turn, propelled research in link prediction in KGs, the task of predicting missing relationships from the available knowledge. Solutions based on KG embeddings have shown promising results in this matter. On the downside, these approaches are usually unable to explain their predictions. While some works have proposed to compute post-hoc rule explanations for embedding-based link predictors, these efforts have mostly resorted to rules with unbounded atoms, e.g., bornIn(x,y) =>residence(x,y), learned on a global scope, i.e., the entire KG. None of these works has considered the impact of rules with bounded atoms such as nationality(x,England) =>speaks(x, English), or the impact of learning from regions of the KG, i.e., local scopes. We therefore study the effects of these factors on the quality of rule-based explanations for embedding-based link predictors. Our results suggest that more specific rules and local scopes can improve the accuracy of the explanations. Moreover, these rules can provide further insights about the inner-workings of KG embeddings for link prediction.
{"title":"Effects of Locality and Rule Language on Explanations for Knowledge Graph Embeddings","authors":"Luis Galárraga","doi":"10.48550/arXiv.2302.06967","DOIUrl":"https://doi.org/10.48550/arXiv.2302.06967","url":null,"abstract":"Knowledge graphs (KGs) are key tools in many AI-related tasks such as reasoning or question answering. This has, in turn, propelled research in link prediction in KGs, the task of predicting missing relationships from the available knowledge. Solutions based on KG embeddings have shown promising results in this matter. On the downside, these approaches are usually unable to explain their predictions. While some works have proposed to compute post-hoc rule explanations for embedding-based link predictors, these efforts have mostly resorted to rules with unbounded atoms, e.g., bornIn(x,y) =>residence(x,y), learned on a global scope, i.e., the entire KG. None of these works has considered the impact of rules with bounded atoms such as nationality(x,England) =>speaks(x, English), or the impact of learning from regions of the KG, i.e., local scopes. We therefore study the effects of these factors on the quality of rule-based explanations for embedding-based link predictors. Our results suggest that more specific rules and local scopes can improve the accuracy of the explanations. Moreover, these rules can provide further insights about the inner-workings of KG embeddings for link prediction.","PeriodicalId":91439,"journal":{"name":"Advances in intelligent data analysis. International Symposium on Intelligent Data Analysis","volume":"415 1","pages":"143-155"},"PeriodicalIF":0.0,"publicationDate":"2023-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79429056","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-02-13DOI: 10.48550/arXiv.2302.06523
C. SimoAlami, Rim Kaddah, J. Read
Clustering in high dimension spaces is a difficult task; the usual distance metrics may no longer be appropriate under the curse of dimensionality. Indeed, the choice of the metric is crucial, and it is highly dependent on the dataset characteristics. However a single metric could be used to correctly perform clustering on multiple datasets of different domains. We propose to do so, providing a framework for learning a transferable metric. We show that we can learn a metric on a labelled dataset, then apply it to cluster a different dataset, using an embedding space that characterises a desired clustering in the generic sense. We learn and test such metrics on several datasets of variable complexity (synthetic, MNIST, SVHN, omniglot) and achieve results competitive with the state-of-the-art while using only a small number of labelled training datasets and shallow networks.
{"title":"Transferable Deep Metric Learning for Clustering","authors":"C. SimoAlami, Rim Kaddah, J. Read","doi":"10.48550/arXiv.2302.06523","DOIUrl":"https://doi.org/10.48550/arXiv.2302.06523","url":null,"abstract":"Clustering in high dimension spaces is a difficult task; the usual distance metrics may no longer be appropriate under the curse of dimensionality. Indeed, the choice of the metric is crucial, and it is highly dependent on the dataset characteristics. However a single metric could be used to correctly perform clustering on multiple datasets of different domains. We propose to do so, providing a framework for learning a transferable metric. We show that we can learn a metric on a labelled dataset, then apply it to cluster a different dataset, using an embedding space that characterises a desired clustering in the generic sense. We learn and test such metrics on several datasets of variable complexity (synthetic, MNIST, SVHN, omniglot) and achieve results competitive with the state-of-the-art while using only a small number of labelled training datasets and shallow networks.","PeriodicalId":91439,"journal":{"name":"Advances in intelligent data analysis. International Symposium on Intelligent Data Analysis","volume":"2 1","pages":"15-28"},"PeriodicalIF":0.0,"publicationDate":"2023-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82341871","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-02-07DOI: 10.1007/978-3-031-30047-9_14
Edith Heiter, Bo Kang, R. Seurinck, Jefrey Lijffijt
{"title":"Revised Conditional t-SNE: Looking Beyond the Nearest Neighbors","authors":"Edith Heiter, Bo Kang, R. Seurinck, Jefrey Lijffijt","doi":"10.1007/978-3-031-30047-9_14","DOIUrl":"https://doi.org/10.1007/978-3-031-30047-9_14","url":null,"abstract":"","PeriodicalId":91439,"journal":{"name":"Advances in intelligent data analysis. International Symposium on Intelligent Data Analysis","volume":"34 1","pages":"169-181"},"PeriodicalIF":0.0,"publicationDate":"2023-02-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88503901","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-01-19DOI: 10.1007/978-3-031-30047-9_13
R. Heese, Sascha Mücke, Matthias Jakobs, Thore Gerlach, N. Piatkowski
{"title":"Shapley Values with Uncertain Value Functions","authors":"R. Heese, Sascha Mücke, Matthias Jakobs, Thore Gerlach, N. Piatkowski","doi":"10.1007/978-3-031-30047-9_13","DOIUrl":"https://doi.org/10.1007/978-3-031-30047-9_13","url":null,"abstract":"","PeriodicalId":91439,"journal":{"name":"Advances in intelligent data analysis. International Symposium on Intelligent Data Analysis","volume":"176 1","pages":"156-168"},"PeriodicalIF":0.0,"publicationDate":"2023-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74708045","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-01-01DOI: 10.1007/978-3-031-30047-9_23
Maxime Masson, P. Roose, C. Sallaberry, R. Agerri, M. Bessagnet, A. L. Parc-Lacayrelle
{"title":"APs: A Proxemic Framework for Social Media Interactions Modeling and Analysis","authors":"Maxime Masson, P. Roose, C. Sallaberry, R. Agerri, M. Bessagnet, A. L. Parc-Lacayrelle","doi":"10.1007/978-3-031-30047-9_23","DOIUrl":"https://doi.org/10.1007/978-3-031-30047-9_23","url":null,"abstract":"","PeriodicalId":91439,"journal":{"name":"Advances in intelligent data analysis. International Symposium on Intelligent Data Analysis","volume":"29 1","pages":"287-299"},"PeriodicalIF":0.0,"publicationDate":"2023-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74052834","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-01-01DOI: 10.1007/978-3-031-30047-9_21
Antonio Lopez-Martinez-Carrasco, Hugo Manuel Proença, J. Juarez, M. Leeuwen, M. Campos
{"title":"Discovering Diverse Top-K Characteristic Lists","authors":"Antonio Lopez-Martinez-Carrasco, Hugo Manuel Proença, J. Juarez, M. Leeuwen, M. Campos","doi":"10.1007/978-3-031-30047-9_21","DOIUrl":"https://doi.org/10.1007/978-3-031-30047-9_21","url":null,"abstract":"","PeriodicalId":91439,"journal":{"name":"Advances in intelligent data analysis. International Symposium on Intelligent Data Analysis","volume":"95 1","pages":"262-273"},"PeriodicalIF":0.0,"publicationDate":"2023-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80878482","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-01-01DOI: 10.1007/978-3-031-30047-9_33
Andreas Theissler, Manuel Wengert, Felix Gerschner
{"title":"ROCKAD: Transferring ROCKET to Whole Time Series Anomaly Detection","authors":"Andreas Theissler, Manuel Wengert, Felix Gerschner","doi":"10.1007/978-3-031-30047-9_33","DOIUrl":"https://doi.org/10.1007/978-3-031-30047-9_33","url":null,"abstract":"","PeriodicalId":91439,"journal":{"name":"Advances in intelligent data analysis. International Symposium on Intelligent Data Analysis","volume":"31 1","pages":"419-432"},"PeriodicalIF":0.0,"publicationDate":"2023-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81223273","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-01-01DOI: 10.1007/978-3-031-30047-9_38
Zhaozhen Xu, Zhijin Guo, N. Cristianini
{"title":"On Compositionality in Data Embedding","authors":"Zhaozhen Xu, Zhijin Guo, N. Cristianini","doi":"10.1007/978-3-031-30047-9_38","DOIUrl":"https://doi.org/10.1007/978-3-031-30047-9_38","url":null,"abstract":"","PeriodicalId":91439,"journal":{"name":"Advances in intelligent data analysis. International Symposium on Intelligent Data Analysis","volume":"108 1","pages":"484-496"},"PeriodicalIF":0.0,"publicationDate":"2023-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84017422","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-01-01DOI: 10.1007/978-3-031-30047-9_24
J. Matthiesen, Hanne Hastedt, Ulf Brefeld
{"title":"User Authentication via Multifaceted Mouse Movements and Outlier Exposure","authors":"J. Matthiesen, Hanne Hastedt, Ulf Brefeld","doi":"10.1007/978-3-031-30047-9_24","DOIUrl":"https://doi.org/10.1007/978-3-031-30047-9_24","url":null,"abstract":"","PeriodicalId":91439,"journal":{"name":"Advances in intelligent data analysis. International Symposium on Intelligent Data Analysis","volume":"34 1","pages":"300-313"},"PeriodicalIF":0.0,"publicationDate":"2023-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78328974","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：