Devanshu Arya

I am happy to release our work "Patch Gradient Descent: Training Neural Networks on Very Large Images", a method that allows to train existing neural networks on images of very large size without the need to struggle with GPU memory bottleneck. See here: arxiv.org/abs/2301.13817

Congratulations Dr. @devanshuarya !

Adaptive Neural Message Passing for Inductive Learning on Hypergraphs. arxiv.org/abs/2109.10683

Attending @iclr_conf 2021 and interested to know more about our budget-aware structured pruning work? Drop by at Spot A3, Room 10C on May 6 between 10:00 am - 12:00 noon CET to hear more about 'ChipNet' openreview.net/forum?id=xCxXw…. Meet our researchers to know more about @TransmuteAI

Our recent work on improving the performance of solar cells using CNN reparameterization was accepted at @iclr_conf ICLR2021 Deep Learning for Simulation (SimDL) workshop. Paper: arxiv.org/abs/2104.04017 Code: github.com/transmuteAI/So… #DeepLearning #solarenergy #ai4physics

Glad to share that our recent paper "Rotation Equivariant Siamese Networks for Tracking", a work together with @devanshuarya and @egavves, is provisionally accepted at #CVPR2021 . Preprint: lnkd.in/dmK_Fpt #CVPR2021 #CVPR #deeplearning #computervision

Group equivariance improves object tracking with Siamese networks. - Scale changes: arxiv.org/pdf/2007.09115… by @artemmoskalev, @arnoldsmeulders and me - Rotations: arxiv.org/pdf/2012.13078… by @dk_gup, @devanshua, @egavves

HyperSAGE: Generalizing Inductive Representation Learning on Hypergraphs. arxiv.org/abs/2010.04558

How do you relate to your #Manager ? We are conducting a study on #workplace relationships. You can contribute to this research by using the link below. It would take less than 10 minutes! bit.ly/2ELHJ8h Thank you! @lvanderwerff1 @mary_kinahan

HyperLearn: A Distributed Approach for Representation Learning in Datasets With Many Moda... arxiv.org/abs/1909.09252

HyperLearn: A Distributed Approach for Representation Learning in Datasets With Many Modalities deepai.org/publication/hy… by Devanshu Arya et al. #ComputerScience #Learning