← Search

Mohammad Mohaiminul Islam

2 accepted papers

2026

Platonic Transformers: A Solid Choice For Equivariance

ICML 2026poster

While widespread, Transformers lack inductive biases for geometric symmetries common in science and computer vision. Existing equivariant methods often sacrifice the efficiency and flexibility that make Transformers so effective through complex, computationally intensive designs. We introduce the Pl…

Cited by 0SourceScholar
2025

Probing Equivariance and Symmetry Breaking in Convolutional Networks

NeurIPS 2025poster

In this work, we explore the trade-offs of explicit structural priors, particularly group-equivariance. We address this through theoretical analysis and a comprehensive empirical study focusing on point clouds. To enable controlled and fair comparisons, we introduce \texttt{Rapidash}, a unified grou…

Cited by 0SourcecodeScholar