O pesquisador russo tratará dos temas “Pattern structures for data analysis” e “Explainable knowledge discovery in numerical data with interval pattern structures“
Nos dias 12 e 13 de novembro de 2024, serão apresentados dois seminários pelo professor Sergei Kuznetsov, do National Research University Higher School of Economics, Moscow, Rússia. Os eventos ocorrerão no Anfiteatro do Centro de Informática (CIn) da UFPE e terão os seguintes temas: no dia 12, às 9h, “Pattern structures for data analysis”, e no dia 13, às 14h, “Explainable knowledge discovery in numerical data with interval pattern structures”. O evento é aberto à comunidade do Centro e não é necessário realizar inscrição prévia.
Confira mais informações sobre o palestrante e o resumo dos seminários abaixo:
Seminário 1
Data: 12/11/2024
Horário: 9h
Local: Anfiteatro do CIn
Título: Pattern structures for data analysis
Resumo: Pattern Structures is an extension of Formal Concept Analysis to data where descriptions of objects make a semilattice based on similarity operation which takes object descriptions to the maximal common description . We’ll consider pattern structures defined by sets of graphs and tuples of numerical intervals. Pattern structures allow one to avoid binarization and work directly with data in their original form, by generating clusters of objects, taxonomies of objects, functional dependencies and implicational dependencies in data of different kinds. We’ll consider applications of pattern structures from bioinformatics to NLP.
Seminário 2
Data: 13/11/2024
Horário: 14h
Local: Anfiteatro do CIn
Título: Explainable knowledge discovery in numerical data with interval pattern structures
Resumo: Interval pattern structures allow for direct processing of numerical data by constructing clusters, taxonomies of objects, implicational dependencies, biclusters of similar values while avoiding binarization. Models, applications, and algorithmic problems related to interval pattern structures will be discussed. We show that interval pattern structures propose explainable methods of knowledge discovery in numerical data. Several applications in various domains will be discussed.
Sobre o palestrante – Prof. Sergei O. Kuznetsov graduated from Moscow Institute for Physics and Technology and defended Doctor of Science thesis on Machine Learning Models Based on Concept Lattices in 2002 at the Computing Center of Russian Academy of Science. He is now full professor at the HSE University in Moscow, being the head of School for Data Analysis and Artificial Intelligence, International Laboratory for Intelligent Systems, and academic supervisor of the Data Science master program. His main topic of research are the methods of knowledge discovery based on ordered sets.
Comentários desativados