Семинар за вештачку интелигенцију, 22. новембар 2023.
Детаљније: Наредни састанак Семинара биће одржан онлајн у среду, 22. новембра 2023. са почетком у 19 часова.
Предавач: Милош Радовановић, Faculty of Sciences, University of Novi Sad
Наслов предавања: GRAPHS IN SPACE AND TIME: GRAPH EMBEDDINGS FOR MACHINE LEARNING IN COMPLEX DYNAMICAL SYSTEMS
Апстракт: In this talk, we will present our newly approved project “Graphs in Space and Time: Graph Embeddings for Machine Learning in Complex Dynamical Systems (TIGRA)” supported by the Serbian Science Fund under the PRISMA program, which is a continuation of our previous work on the project “Graphs in Space: Graph Embeddings for Machine Learning on Complex Data (GRASP)”, also supported by the Serbian Science Fund. After introducing the basic concepts and overviewing notable results produced by the GRASP project for embedding static graphs, we will focus on their possible extensions to dynamically evolving graphs, which are a useful abstraction for representing many complex systems of high technological, social, and scientific importance. Existing dynamic graph embedding algorithms provide various mechanisms to incorporate temporality into the main principles of static graph embedding learning, but they do not take higher-order structural properties of graphs into account. The TIGRA project will investigate the impact of hubness (high connectedness) and LID (local intrinsic dimensionality) to various quality aspects of dynamic graph embeddings. Afterwards, we plan to design novel hub-aware and LID-aware dynamic graph embedding methods. Additionally, we will propose the first algorithms that are able to embed graphs with time-series attributes into Euclidean spaces. We expect the project to produce novel dynamic graph embedding methods substantially more accurate than the state-of-the-art in three aspects: (1) preserving higher-order graph structural properties, (2) evolutionary stability, and (3) the accuracy of derived machine learning models for classification, clustering, event prediction, anomaly detection and diffusion prediction.
Напомена: Регистрациона форма за учешће на Семинару је доступна на линку:
https://miteam.mi.sanu.ac.rs/asset/CW5nJWDSEZDj7p32p
Уколико желите само да пратите предавање без могућности активног учешћа, пренос је доступан на линку:
https://miteam.mi.sanu.ac.rs/asset/4LNW8WtML7rLKojoz