Michalis Vazirgiannis
Affiliated Professor of Machine Learning
Research interests
Professor Vazirgiannis is working in the areas of machine and deep learning models and methods for large scale heterogeneous data (including graphs and text). Part of his work is devoted to graph structured data present in many domains including biology, social networks, power/communication/transport facilities. He focuses especially on Graph Neural networks with interesting methodological contributions. This area is advancing rapidly with promising potential for many real-life applications. His latest interests are on multimodal graph generative models.
He has extensive experience in the area of NLP where he contributed Graph of Words – a novel graph-based document representation. In the past few years, he supervised efforts for the data collection and training of multilingual and domain specific language models. Most recently, he has been working in the area of multi-modal generative AI with emphasis on graph modality with applications in synthetic data for biomedical cases.
Email
Prior to joining MBZUAI, Professor Vazirgiannis began his academic career at Athens Economic University from 1997 to 2013, laying the foundation for his enduring commitment to education and research as Professor of Informatics. From 2013, Professor Vazirgiannis affiliation has been at Ecole Polytechnique in France where, since 2016, he holds a position of Distinguished Professor.
He embarked on his academic journey with post-doctoral fellowships from the TMR-Chorochronos research network, enabling visits to INRIA in Paris and Fernuni-Hagen in Germany from 1997 to 1998. Following that, he was awarded a Marie Curie Fellowship for research work in INRIA Paris from 2006 to 2008, with a thematic focus on distributed/P2P web search.
His academic pursuits continued as he became the DIGITEO Chair grant holder at LIX, Ecole Polytechnique, Paris, France, from 2010 to 2013. In 2015, he earned an AXA-funded industrial chair in "Data Science for Insurance Data", which he held until 2018. Most recently, in 2020, he was appointed as the ANR Chair for "AML-HELAS – Advanced Machine/Deep Learning for Heterogeneous Large-scale Data," a position he is expected to hold until 2026.
- Ph.D. in computer science, NKUA, Athens, Greece
- M.Sc. in knowledge-based systems, Herriot Watt University, Edinburgh, UK
- Marie Curie Intra European Fellowship (2006–8) hosted at INRIA Paris
- "Rhino-Bird International Academic Expert Award" in recognition of his academic/professional work @ Tencent (2017) – details here
- Distinguished paper award – IJCAI 2018 (https://www.ijcai-18.org/distinguished-papers/index.html)
- His D-core metric measuring jointly authority and collaborativeness has been adopted the Aminer (a main Chinese scientific portal). See an example here. (click “ D-core”).
- "GNNs and Graph generative models for biomedical applications", Keynote speaker in WEBCONF2023, Texas USA, May 2023
- “Path neural networks: Expressive and accurate graph neural networks”, G Michel, G Nikolentzos, JF Lutzeyer, M Vazirgiannis, International Conference on Machine Learning, 24737-24755, 2023.
- “FrugalScore: Learning Cheaper, Lighter and Faster Evaluation Metrics for Automatic Text Generation”, Moussa Kamal Eddine1, Guokan Shang2*, Antoine J.-P. Tixier1*, Michalis Vazirgiannis, ACL (1) 2022: 1305-1318
- “BARThez: a Skilled Pretrained French Sequence-to-Sequence Model”. EMNLP (1) 2021: 9369-9390
- “GraKeL: A Graph Kernel Library in Python”, Giannis Siglidis, Giannis Nikolentzos, Stratis Limnios, Christos Giatsidis, Konstantinos Skianis, Michalis Vazirgiannis; JMLR 21(54):1−5, 2020.
- “Synthetic electronic health records generated with variational graph autoencoders”, G Nikolentzos, M Vazirgiannis, C Xypolopoulos, M Lingman, EG Brandt, Nature - Digital Medicine 6 (1), 83, 2023
- “A Degeneracy Framework for Graph Similarity", G Nikolentzos, P Meladianos, S Limnios, M Vazirgiannis, 2595-2601, IJCAI 2018 (best paper award)