MG

Michael Galkin

AI Research Scientist

Contact Information


  • Full Name: Michael Galkin

  • Position: AI Research Scientist

  • Company: Intel Labs



  • Location: San Diego, CA, USA




Professional Summary


Michael Galkin is currently an AI Research Scientist at Intel AI Lab. His expertise encompasses Graph Machine Learning (Graph ML), Geometric Deep Learning (DL), Knowledge Graphs, and Graph Neural Networks (GNNs). He is deeply involved in advancing machine learning applications in materials science.

Career Highlights


Current Role


  • Intel Labs:

  • Role: AI Research Scientist

  • Focus: Graph Machine Learning and Geometric Deep Learning applications.

  • Projects and Contributions:

  • Michael has organized numerous discussions and panels, bringing together experts to explore advancements in Graph and Geometric Machine Learning.

  • Recently posted on LinkedIn about Intel Labs' ULTRA project and the Long Range Graph Benchmark (LRGB).


Previous Experience


  • Postdoctoral Fellow at Mila and McGill University (2021 - 2022)

  • Location: Montreal, Quebec, Canada

  • Supervisors: Jian Tang, Reihaneh Rabbany, William L. Hamilton

  • Research Scientist at Fraunhofer IAIS, Dresden, Germany (2018 - 2020)

  • Member of the Smart Data Analytics group.

  • Worked on conversational AI guided by Prof. Dr. Jens Lehmann.

  • Research Assistant at Fraunhofer IAIS, Bonn, Germany (2015 - 2018)

  • Member of the Smart Data Analytics group at the University of Bonn.

  • Supervisor: Prof. Dr. Sören Auer.


Academic Background


  • Ph.D. in Computer Science (Knowledge Graphs), University of Bonn, 2018

  • M.Sc. in Computer Science, ITMO University, Saint Petersburg, 2014


Community Involvement and Teaching


  • Michael co-authored an open Knowledge Graph course aimed at Russian-speaking audiences.

  • Actively runs the Mila Graph Representation Learning reading group.

  • Chair of the KG+NLP track at the Knowledge Graph Conference 2022.

  • Facilitated a panel discussion at Connected Data World 2021 focused on Graph ML and its applications in research and industry.


Selected Talks and Presentations


  • Graph Foundation Models for Knowledge Graph Reasoning and Beyond - Talk at National University of Singapore (May 15, 2024)

  • Foundation Models for Knowledge Graph Reasoning: Is Inductive Reasoning Solved? - Talk at Universidad San Sebastian, Santiago, Chile (March 04, 2024)

  • Neural Graph Reasoning - Online presentation (January 22, 2024)

  • Graph Foundation Models - Dagstuhl Seminar, Germany (December 05, 2023)

  • Towards Foundation Models for Graph Reasoning and AI4Science - UC San Diego, La Jolla, San Diego (October 11, 2023)


Publications


Michael has an extensive portfolio of publications, notable among them include:
  • Scaling Computational Performance of Spherical Harmonics Kernels with Triton

  • Position: Graph Foundation Models Are Already Here

  • TGB 2.0: A Benchmark for Learning on Temporal Knowledge Graphs and Heterogeneous Graphs

  • GraphAny: A Foundation Model for Node Classification on Any Graph

  • Zero-shot Logical Query Reasoning on any Knowledge Graph


Conclusion


Michael Galkin is a significant figure in the AI research community, primarily focusing on Graph Machine Learning and its applications. His role at Intel Labs positions him at the forefront of industry innovation. His academic and professional experiences, coupled with his involvement in community education and extensive publications, make him a key contact for advancing technical partnerships and collaborative research initiatives.