Haggai Maron
I am a senior research scientist at NVIDIA Research and a member of NVIDIA's TLV lab. I will be joining the Faculty of Electrical and Computer Engineering at the Technion as an Assistant Professor in 2023. My main field of interest is machine learning. In particular, I am working on applying deep learning to structured domains (e.g., sets, graphs, point clouds, and surfaces), usually by leveraging their symmetry structure. I completed my Ph.D. in 2019 at the Weizmann Institute of Science under the supervision of Prof. Yaron Lipman. My CV can be found here.
Prospective students:
I am seeking highly motivated Ph.D. students to join my new lab at the Technion. My research combines machine learning, and some aspects of algebra and graph theory, so knowledge and/or interest in these fields are a big plus. Coding skills are also required. You can get an idea of what I am working on by taking a look at these two recent talks I gave:- Subgraph-based networks for expressive, efficient, and domain-independent graph learning
- Leveraging Permutation Group Symmetries for Equivariant Neural Networks
Email: hmaron (at) nvidia.com, Google scholar page, GitHub page
News
- I will serve as an area chair at NeurIPS 2023.
- A recent podcast with me on graph neural networks (hebrew)
- See our recent tutorial on Exploring the practical and theoretical landscape of expressive graph neural networks given at the Learning on Graphs Conference . With With Fabrizio Frasca and Beatrice Bevilacqua.
- See my recent talk on Subgraph-based networks for expressive, efficient, and domain-independent graph learning at the Signal Processing and Machine Learning on Graphs hosted by CIRM .
- Our paper Understanding and Extending Subgraph GNNs by Rethinking Their Symmetries was accepted as an oral presentation to NeurIPS2022
- Our paper Equivariant Subgraph Aggregation Networks was accepted as a spotlight presentation to ICLR2022
- A new blog post on Subggraph GNNs on Towards Data Science.
- The Second Israeli Workshop on Geometric Deep Learning took place virtually on October 11th 2021. See full video recording here.
- Our paper Deep Permutation Equivariant Structure from Motion was accepted to ICCV 2021 as an oral presentation!
- Two papers accepted to ICML 2021: From Local Structures to Size Generalization in Graph Neural Networks and Controlling Graph Dynamics with Reinforcement Learning and Graph Neural Networks.
- Two papers accepted to ICLR 2021: On the Universality of Rotation Equivariant Point Cloud Networks and Auxiliary Learning by Implicit Differentiation.
- See my talk on Leveraging Permutation Group Symmetries for Equivariant Neural Networks at the Workshop on Equivariance and Data Augmentation hosted by the University of Pennsylvania.
- Our ICML 2020 paper On Learning Sets of Symmetric Elements received the Outstanding paper award (best paper award). See this interview for more details.
- July 2020: I am organizing the first Israeli Workshop on Geometric Deep Learning that will take place virtually on August 2nd 2020. See full video recording here.
- June 2020: Two papers accepted to ICML 2020: On Learning Sets of Symmetric Elements and Learning Algebraic Multigrid Using Graph Neural Networks.
- Check out our new blog on deep learning of irregular data! Our first two blog posts explain Invariant Graph Networks (IGNs) and their power. These posts summarize our ICLR 2019, ICML 2019 and NeurIPS 2019 papers.
Teaching
- 2019/spring (WIS): Geometric and Algebraic Methods in Deep Learning
- 2018/winter (WIS): Geometry and Deep Learning
