M
Michalis Vazirgiannis
Researcher at École Polytechnique
Publications - 355
Citations - 15006
Michalis Vazirgiannis is an academic researcher from École Polytechnique. The author has contributed to research in topics: Computer science & Cluster analysis. The author has an hindex of 49, co-authored 326 publications receiving 13390 citations. Previous affiliations of Michalis Vazirgiannis include French Institute for Research in Computer Science and Automation & Télécom ParisTech.
Papers
More filters
Proceedings ArticleDOI
Rank Prediction in Graphs with Locally Weighted Polynomial Regression and EM of Polynomial Mixture Models
TL;DR: A learning framework enabling ranking predictions for graph nodes based solely on individual local historical data and the similarity between the predicted and the actual rankings and compared to alternative baseline predictor is described.
Posted Content
FastGAE: Fast, Scalable and Effective Graph Autoencoders with Stochastic Subgraph Decoding.
TL;DR: This paper introduces FastGAE, a general framework to scale graph AE and VAE to large graphs with millions of nodes and edges, based on node sampling and subgraph decoding, which significantly speeds up the training of graphAE andVAE while preserving or even improving performances.
Proceedings ArticleDOI
Speaker-change Aware CRF for Dialogue Act Classification
TL;DR: Experiments on the SwDA corpus show that the modified CRF layer outperforms the original one, with very wide margins for some DA labels, and visualizations demonstrate that the layer can learn meaningful, sophisticated transition patterns between DA label pairs conditioned on speaker-change in an end-to-end way.
Book ChapterDOI
Distributed knowledge discovery with non linear dimensionality reduction
TL;DR: D-Isomap is introduced, a novel distributed non linear dimensionality reduction algorithm, particularly applicable in large scale, structured peer-to-peer networks and capable of approximate reconstruction of the global dataset at peer level.
Posted Content
Unsupervised Word Polysemy Quantification with Multiresolution Grids of Contextual Embeddings
TL;DR: This work proposes a novel method to estimate polysemy based on simple geometry in the contextual embedding space, which is fully unsupervised and purely data-driven, and applicable to any language.