Slow Convergence

Slow convergence in machine learning algorithms, hindering efficient model training, is a central research challenge. Current efforts focus on improving convergence speed in various contexts, including federated learning (across heterogeneous devices and data), reinforcement learning (especially with long-term dependencies), and distributed optimization (addressing asynchronous updates and communication bottlenecks). These investigations often involve adapting optimization algorithms (like SGD and Adam), employing novel architectures (such as Vision Transformers), or developing new techniques to manage data heterogeneity and communication delays, ultimately aiming to accelerate training and improve model performance in diverse applications.

Papers

October 14, 2024

FasterDiT: Towards Faster Diffusion Transformers Training without Architecture Modification
Jingfeng Yao, Wang Cheng, Wenyu Liu, Xinggang Wang
Supervised ImageNet Diffusion Transformer Signal to Noise Ratio Slow Convergence Iteration Head

October 11, 2024

Slow Convergence of Interacting Kalman Filters in Word-of-Mouth Social Learning
Vikram Krishnamurthy, Cristian Rojas
LeArning Abstract Kalman Filter Social Learning Slow Convergence

October 9, 2024

OledFL: Unleashing the Potential of Decentralized Federated Learning via Opposite Lookahead Enhancement
Qinglun Li, Miao Zhang, Mengzhu Wang, Quanjun Yin, Li Shen
Full Potential Decentralized Federated Learning Slow Convergence Centralized Federated Learning Lookahead Decoding

September 5, 2024

On the Convergence Rates of Federated Q-Learning across Heterogeneous Environments
Muxing Wang, Pengkun Yang, Lili Su
Reinforcement Learning Q Learning Convergence Rate Federated Prompt Cooperation Q Function Large Scale Multi Agent Heterogeneous Environment Slow Convergence

March 7, 2024

Mastering Memory Tasks with World Models
Mohammad Reza Samsami, Artem Zholus, Janarthanan Rajendran, Sarath Chandar
State Space Model World Model Long Term Memory Slow Convergence Symbolic Memory Current Reinforcement Learning

February 29, 2024

On the Convergence of Federated Learning Algorithms without Data Similarity
Ali Beikmohammadi, Sarit Khirirat, Sindri Magnússon
Federated Learning Early Stage Convergence Slow Convergence Federated Learning Algorithm Convergent Algorithm Data Similarity

February 6, 2024

The Challenges of the Nonlinear Regime for Physics-Informed Neural Networks
Andrea Bonfanti, Giuseppe Bruno, Cristina Cipriani
Technical Challenge Physic Informed Neural Network Neural Tangent Kernel Slow Convergence Optimal Regime

December 11, 2023

Asynchronous Distributed Optimization with Delay-free Parameters
Xuyang Wu, Changxin Liu, Sindri Magnusson, Mikael Johansson
Convergence Guarantee Distributed Optimization Consensus Algorithm Slow Convergence Asynchronous Algorithm

November 6, 2023

Signal Processing Meets SGD: From Momentum to Filter
Zhipeng Yao, Guiyuan Fu, Ying Li, Yu Zhang, Dazhou Li, Rui Yu
Stochastic Gradient Descent Learning Rate Signal Processing Residual Momentum Slow Convergence

October 13, 2023

Tackling Heterogeneity in Medical Federated learning via Vision Transformers
Erfan Darzi, Yiqing Shen, Yangming Ou, Nanna M. Sijtsema, P. M. A van Ooijen
Vision Transformer Federated Learning Regularization Model Slow Convergence

August 17, 2023

Enhancing Convergence Speed with Feature-Enforcing Physics-Informed Neural Networks: Utilizing Boundary Conditions as Prior Knowledge for Faster Convergence
Mahyar Jahaninasab, Mohamad Ali Bijarchi
Neural Network Deep Learning Model Loss Function Physic Informed Neural Network Prior Knowledge Faster Convergence Boundary Condition Slow Convergence

April 14, 2023

TimelyFL: Heterogeneity-aware Asynchronous Federated Learning with Adaptive Partial Training
Tuo Zhang, Lei Gao, Sunwoo Lee, Mi Zhang, Salman Avestimehr
Asynchronous Federated Learning Cross Device Federated Learning Slow Convergence Stale Gradient Partial Model Training

March 16, 2023

Efficient Diffusion Training via Min-SNR Weighting Strategy
Tiankai Hang, Shuyang Gu, Chen Li, Jianmin Bao, Dong Chen, Han Hu, Xin Geng, Baining Guo
Supervised ImageNet Denoising Diffusion Model Diffusion Training Slow Convergence Weighting Strategy

October 6, 2022

Federated Learning with Server Learning: Enhancing Performance for Non-IID Data
Van Sy Mai, Richard J. La, Tao Zhang
Training Data Raw Data System Performance Small Datasets Auxiliary Learning Slow Convergence

June 21, 2022

Federated Stochastic Approximation under Markov Noise and Heterogeneity: Applications in Reinforcement Learning
Sajad Khodadadian, Pranay Sharma, Gauri Joshi, Siva Theja Maguluri
Q Learning Federated Reinforcement Learning Linear Speedup Slow Convergence Markovian Sampling Federated Automatic Differentiation

May 10, 2022

A Communication-Efficient Distributed Gradient Clipping Algorithm for Training Deep Neural Networks
Mingrui Liu, Zhenxun Zhuang, Yunwei Lei, Chunyang Liao
Practical Algorithm Stochastic Gradient Descent Natural Gradient Communication Efficient Training Deep Communication Cost Slow Convergence Continuous Gradient

March 20, 2022

Robust Action Gap Increasing with Clipped Advantage Learning
Zhe Zhang, Yaozhong Gan, Xiaoyang Tan
Optimal Action Slow Convergence CLIP Training Advantage Learning Action Gap

February 17, 2022

Delay-adaptive step-sizes for asynchronous learning
Xuyang Wu, Sindri Magnusson, Hamid Reza Feyzmahdavian, Mikael Johansson
Machine Learning Step Size Slow Convergence Asynchronous Learning Asynchronous Algorithm Asynchronous Iteration

February 12, 2022

Maximizing Communication Efficiency for Large-scale Training via 0/1 Adam
Yucheng Lu, Conglong Li, Minjia Zhang, Christopher De Sa, Yuxiong He
Faster Training Gradient Compression Large Scale Training Slow Convergence Communication Optimization Convergence Criterion