Picture for Aleksandr Beznosikov

Aleksandr Beznosikov

Broadening Discovery through Structural Models: Multimodal Combination of Local and Structural Properties for Predicting Chemical Features

Add code
Feb 25, 2025
Viaarxiv icon

Variance Reduction Methods Do Not Need to Compute Full Gradients: Improved Efficiency through Shuffling

Add code
Feb 20, 2025
Viaarxiv icon

Accelerated Methods with Compressed Communications for Distributed Optimization Problems under Data Similarity

Add code
Dec 21, 2024
Viaarxiv icon

Label Privacy in Split Learning for Large Models with Parameter-Efficient Training

Add code
Dec 21, 2024
Viaarxiv icon

Just a Simple Transformation is Enough for Data Protection in Vertical Federated Learning

Add code
Dec 16, 2024
Viaarxiv icon

FRUGAL: Memory-Efficient Optimization by Reducing State Overhead for Scalable Training

Add code
Nov 12, 2024
Viaarxiv icon

Accelerated Stochastic ExtraGradient: Mixing Hessian and Gradient Similarity to Reduce Communication in Distributed and Federated Learning

Add code
Sep 22, 2024
Viaarxiv icon

Gradient Clipping Improves AdaGrad when the Noise Is Heavy-Tailed

Add code
Jun 06, 2024
Figure 1 for Gradient Clipping Improves AdaGrad when the Noise Is Heavy-Tailed
Figure 2 for Gradient Clipping Improves AdaGrad when the Noise Is Heavy-Tailed
Figure 3 for Gradient Clipping Improves AdaGrad when the Noise Is Heavy-Tailed
Figure 4 for Gradient Clipping Improves AdaGrad when the Noise Is Heavy-Tailed
Viaarxiv icon

Local Methods with Adaptivity via Scaling

Add code
Jun 02, 2024
Figure 1 for Local Methods with Adaptivity via Scaling
Figure 2 for Local Methods with Adaptivity via Scaling
Viaarxiv icon

Sparse Concept Bottleneck Models: Gumbel Tricks in Contrastive Learning

Add code
Apr 04, 2024
Figure 1 for Sparse Concept Bottleneck Models: Gumbel Tricks in Contrastive Learning
Figure 2 for Sparse Concept Bottleneck Models: Gumbel Tricks in Contrastive Learning
Figure 3 for Sparse Concept Bottleneck Models: Gumbel Tricks in Contrastive Learning
Figure 4 for Sparse Concept Bottleneck Models: Gumbel Tricks in Contrastive Learning
Viaarxiv icon