Gautam Pai

Assistant Professor @ University of Twente

[Gautam Pai @gautampai89]

Personal Update - I am very happy to start as an Assistant Professor in the Department of Applied Mathematics @ University of Twente! ⭐️ Do reach out for all things: Numerics, Geometry and Machine Learning https://t.co/TjV9k5DfyT

[Gabriel Peyré @gabrielpeyre]

Oldies but goldies: Richard Sinkhorn, A relationship between arbitrary positive matrices and doubly stochastic matrices, 1964. The first proof of the convergence of the diagonal scaling algorithm (aka IPFP, RAS, etc) toward a uniquely defined bistochastic matrix. https://t.co/k0U1bU0KiK https://t.co/C5cUCeUvZw

[Gautam Pai @gautampai89]

@GThoppe Thank you Gugan!

[Gautam Pai @gautampai89]

Personal Update - I am very happy to start as an Assistant Professor in the Department of Applied Mathematics @ University of Twente! ⭐️ Do reach out for all things: Numerics, Geometry and Machine Learning https://t.co/TjV9k5DfyT

[Gautam Pai @gautampai89]

@sankethvedula Thank you @sankethvedula

[Gautam Pai @gautampai89]

@ZorahLaehner Thank You!

[Gautam Pai @gautampai89]

@sharbat_c Thank you so much @sharbat_c ! Hope to catch up soon sometime🙂

[Gautam Pai @gautampai89]

Also had the pleasure to talk about this in the Paris Shape Analysis Seminar https://t.co/SgvdbpryFp

[Gautam Pai @gautampai89]

Happy to share that: *Optimal Transport on the Lie Group of Roto-translations* will appear soon in SIAM-Journal on Imaging Sciences. Preprint: https://t.co/tkra7kKPJ4

[Gautam Pai @gautampai89]

An amazing collaboration with Math PhD students Daan Bon and @GijsBellaard together with Olga Mula and Remco Duits from @TUeindhoven

[Gautam Pai @gautampai89]

One of my favorites! Spectral geometry comes up in different forms in image filtering, shape analysis, unsupervised learning and probably many more …

[Gabriel Peyré @gabrielpeyre]

Oldies but goldies: M. Belkin, P. Niyogi, Laplacian Eigenmaps for Dimensionality Reduction and Data Representation, 2003. Non-linear dimensionality reduction by embedding data points using the eigenvectors of a graph Laplacian as coordinates. https://t.co/fE1B9r1O2Q

[Information Geometry @SN_INGE]

Gautam Pai, Gijs Bellaard, Bart M. N. Smets, Remco Duits @TUeindhoven🇳🇱
"Functional Properties of PDE-Based Group Equivariant Convolutional Neural Networks"
#GSI2023 #LNCS 14071

#NeurogeometryMeetsGeometricDeepLearning
Free view 🔗 https://t.co/l9jMCgVUD0
https://t.co/LKBWBcMQMj

[International Conference on 3D Vision (3DV) @3DVconf]

Best Paper Award at #3DV2021

DPFM: Deep Partial Functional Maps

Souhaib Attaiki, Gautam Pai and Maks Ovsjanikov

#TBThursday #3DV2024

[Gautam Pai @gautampai89]

PDE-based neural architectures with inf/sup convolutions like this have nice practical properties: lesser parameters and trainable with fewer data. Ongoing work with @GijsBellaard, Bart Smets, and Remco Duits at TU/e: https://t.co/cXIfWPUxFk

[Gautam Pai @gautampai89]

Morphological Operations like Erosion & Dilation are computed using inf/sup convolutions. Extendable to homogeneous spaces of Lie Groups, like SE(2) or the group of roto-translations by max/min pooling over geodesic Riemannian balls. See PDE-G-CNN's https://t.co/iuwyOBvT5X

[Gabriel Peyré @gabrielpeyre]

Fourier is to convolution what Legendre is to inf-convolution. https://t.co/Nd06D1k4g0 https://t.co/0yoS0KApj3

[Gautam Pai @gautampai89]

My research experience 🤣, still fun tho 😁

[ErrantScience @ErrantScience]

Somedays I wish I ran experiments that were a bit more…everything

[Gautam Pai @gautampai89]

@anirudhpammi Your reply:

[Gautam Pai @gautampai89]

@AviMohan21 @overleaf @IJCAIconf 🤣🤣, I felt this way when I accidentally re-read some of my old paper drafts.