Risi Kondor

Risi Kondor homepage

Risi Kondor
Associate Professor
Department of Computer Science
Department of Statistics,
Computational and Applied Mathematics Initiative (CAMI)
The University of Chicago

email address




Currently on leave at the Flatiron Institute (July 2019 - June 2021)

Current    Research    Generalized covariant neural network architectures
Multiresolution/multiscale matrix factorizations
Machine learning for many-particle physics
Learning (on) graphs and other combinatorial structures
Permutation problems and Fourier analysis on the symmetric group

Students    Hy Truong Son (CS PhD, 4th year)
Horace Pan (CS PhD, 5th year)

Postdoc    Erik Thiede

Former   
group   
members   
Maia Fraser (University of Ottawa)
Nedelina Teneva (now at Amazon)
Yi Ding (continuing CS PhD)
Brandon Anderson (now at Atomwise)
Jonathan Eskreis-Winkler (now at Etsy)
Pramod K Mudrakarta (now at Google)

Recent papers    A. Bogatskiy, B. Anderson, J. Offermann, M. Roussi, D. Miller and R. Kondor:  Lorentz group equivariant neural network for particle physics (ICML 2020) [preprint] [bib]

E. H. Thiede, T. Son Hy and R. Kondor:  The general theory of permutation equivarant neural networks and higher order graph variational encoders (April 2020) [bib]

B. Anderson, T. Son Hy and R. Kondor:  Cormorant: Covariant molecular neural networks (NeurIPS 2019) [bib]

K. Swanson, S. Trivedi, J. Lequieu, K. Swanson and R. Kondor:  Deep learning for automated classification and characterization of amorphous materials (Soft Matter, Royal Society of Chemistry, 2020) [bib]

R. Kondor, Z. Lin and S. Trivedi:  Clebsch-Gordan nets: a fully Fourier space spherical convolutional neural network (NeurIPS 2018)

T. Son Hy, S. Trivedi, H. Pan, B. M. Anderson and R. Kondor:  Predicting molecular properties with covariant compositional networks (JCP special issue on data enabled theoretical chemistry, June 2018) [pdf]

R. Kondor:  N-body networks: a covariant hierarchical neural network architecture for learning atomic potentials (March 2018)

R. Kondor and S. Trivedi:  On the generalization of equivariance and convolution in neural networks to the action of compact groups (ICML 2018) [poster] [slides] [preprint]

R. Kondor, Hy Truong Son, Horace Pan, Brandon Anderson, Shubhendu Trivedi:  Covariant compositional networks for learning graphs (preprint, January 2018) [PyTorch code] [GraphFlow] [video]

Y. Ding, R. Kondor and J. Eskreis-Winkler:  Multiresolution kernel approximation for Gaussian process regression (NIPS 2017)

P. K. Mudrakarta and R. Kondor:  A generic multiresolution preconditioner for sparse symmetric systems (preprint 7/7/17)

K. Rajendran, A. A. Kattis, A. Holiday, R. Kondor and I. G. Kevrekidis:  Data mining when each data point is a network (Patterns of Dynamics, Springer Proceedings) [preprint]

V. Ithapu, R. Kondor and V. Singh:  The incremental multiresolution matrix factorization algorithm (CVPR 2017)

R. Kondor and H. Pan:  The Multiscale Laplacian Graph Kernel (NIPS 2016 oral presentation)  [slides] [code] [preprint]

N. Teneva, P. K. Mudrakarta and R. Kondor:  Multiresolution Matrix Compression (AISTATS 2016, winner of notable student paper award)

R. Kondor, N. Teneva and P. K. Mudrakarta:  Parallel MMF:  a Multiresolution Approach to Matrix Computation (arXiv preprint, July 2015) [pMMF library]

G. Plumb, D. Pachauri, R. Kondor and V. Singh:  SnFFT:  a Julia Toolkit for Harmonic Analysis on the Symmetric Group (JMLR 2015)

D. Pachauri, R. Kondor, G. Sargur and V. Singh:  Permutation diffusion maps with application to the image association problem in computer vision (NIPS 2014)

R. Kondor, N. Teneva and V. Garg:  Multiresolution Matrix Factorization (ICML 2014) [supplement] [video]

Recent events    Multiresolution Methods for Large Scale Learning workshop at NIPS 2015 [videos]

IMA Summer School on Modern Applications of Representation Theory Chicago, July 21 - August 6, 2014  (co-organized with Jason Morton and Lek-Heng Lim) [videos]

Software    SnOB: a C++ library for computing fast Fourier transforms on the symmetric group

pMMF: a high performance C++ library for parallel Multiresolution Matrix Factorization

Mondrian:  C++ parallel blocked matrix library

GraphFlow: a C++ deep learning library with support for covariant compositional architectures (Hy Truong Son)

Links    All papers by topic
Curriculum Vitae
Detached threads blog