5 papers
MT-DAO: Multi-Timescale Distributed Adaptive Optimizers with Local Updates
Alex Iacob, Andrej Jovanovic, Mher Safaryan +6
Training large models with distributed data parallelism (DDP) requires frequent communication of gradients across workers, which can saturate bandwidth. Infrequent communication st…
Loss-Transformation Invariance in the Damped Newton Method
Alexander Shestakov, Sushil Bohara, Samuel Horváth +2
The Newton method is a powerful optimization algorithm, valued for its rapid local convergence and elegant geometric properties. However, its theoretical guarantees are usually lim…
Differentially Private Clipped-SGD: High-Probability Convergence with Arbitrary Clipping Level
Saleh Vatan Khah, Savelii Chezhegov, Shahrokh Farahmand +2
Gradient clipping is a fundamental tool in Deep Learning, improving the high-probability convergence of stochastic first-order methods like SGD, AdaGrad, and Adam under heavy-taile…
DES-LOC: Desynced Low Communication Adaptive Optimizers for Training Foundation Models
Alex Iacob, Lorenzo Sani, Mher Safaryan +8
Scaling foundation model training with Distributed Data Parallel (DDP) methods is bandwidth-limited. Existing infrequent communication methods like Local SGD were designed to synch…
Convergence of Clipped-SGD for Convex -Smooth Optimization with Heavy-Tailed Noise
Savelii Chezhegov, Aleksandr Beznosikov, Samuel Horváth +1
Gradient clipping is a widely used technique in Machine Learning and Deep Learning (DL), known for its effectiveness in mitigating the impact of heavy-tailed noise, which frequentl…