4 papers
On the Training Instability of Shuffling SGD with Batch Normalization
David X. Wu, Chulhee Yun, Suvrit Sra
We uncover how SGD interacts with batch normalization and can exhibit undesirable training dynamics such as divergence. More precisely, we study how Single Shuffle (SS) and Random…
QK Iteration: A Self-Supervised Representation Learning Algorithm for Image Similarity
David Wu, Yunnan Wu
Self-supervised representation learning is a fundamental problem in computer vision with many useful applications (e.g., image search, instance level recognition, copy detection).…
Nonuniform Distributions of Residues of Prime Sequences in Prime Moduli
David Wu
For positive integers , Dirichlet's theorem states that there are infinitely many primes in each reduced residue class modulo . A stronger form of the theorem states that the…
Distribution of Nonzero Digits in a Greedy Sequence of Powers of Two
David Wu
Understanding the distribution of digits in the expansions of perfect powers in different bases is difficult. Rather than consider the asymptotic digit distributions, we consider t…