a lab on the cutting edge of theoretical deep learning research


At Landskape, we focus primarily on Deep Learning Theory by expanding our current understanding of correlating concepts from Mathematics and Physics. We employ methodologies to experiment and explore approximation, generalization and optimization processes in deep neural networks.


Deep Learning


Featured Research


Mish: A Self Regularized Non-Monotonic Neural Activation Function.

Triplet Attention

Rotate to Attend: Capturing Cross Dimension Interaction using Permuted Attention.


Let's go HAM -

Hyper Attention Masks


Diganta Misra


An avid mathematician and deep learning enthusiast, focused on deep learning theory and algorithms. With a profound interest in theorems and cosmology, I believe Mathematics is the holy grail of the known and the unknown universe.

Contact Us

Want to join the team or have an interesting research proposal in deep learning theory, drop us an email. 

If you're an organization/ company or a research lab, and want to sponsor our work or collaborate with us, we'd love to talk to you.

  • GitHub-Mark

©2019 by Landskape AI.