HyperSparse Neural Networks: Shifting Exploration to Exploitation through Adaptive Regularization
Sparse neural networks are a key factor in developing resource-efficient machine learning applications. We propose the novel and powerful sparse learning method Adaptive Regularized Training (ART) to compress dense into sparse networks. Instead of the commonly used binary mask during training to reduce the number of model weights, we inherently shrink weights close to zero in an iterative manner with increasing weight regularization. Our method compresses the pre-trained model knowledge into the weights of highest magnitude. Therefore, we introduce a novel regularization loss named HyperSparse that exploits the highest weights while conserving the ability of weight exploration. Extensive experiments on CIFAR and TinyImageNet show that our method leads to notable performance gains compared to other sparsification methods, especially in extremely high sparsity regimes up to 99.8 percent model sparsity. Additional investigations provide new insights into the patterns that are encoded in weights with high magnitudes.
Code (1)
Tasks
Sparse LearningSimilar Papers 제목 키워드 기반
Reward Shifting for Optimistic Exploration and Conservative Exploitation
In this work, we study the simple yet universally applicable case of reward shaping, the linear transformation, in value-based Deep Reinforcement Learning. We show that reward shifting, as the simplest linear reward tran…
continuous-controlContinuous ControlDeep Reinforcement LearningOffline RLOptimistic Curiosity Exploration and Conservative Exploitation with Linear Reward Shaping
In this work, we study the simple yet universally applicable case of reward shaping in value-based Deep Reinforcement Learning (DRL). We show that reward shifting in the form of the linear transformation is equivalent to…
continuous-controlContinuous ControlDeep Reinforcement LearningOffline RLThe Exploration-Exploitation Dilemma Revisited: An Entropy Perspective
The imbalance of exploration and exploitation has long been a significant challenge in reinforcement learning. In policy optimization, excessive reliance on exploration reduces learning efficiency, while over-dependence …
MuJoCoOpt3DGS: Optimizing 3D Gaussian Splatting with Adaptive Exploration and Curvature-Aware Exploitation
3D Gaussian Splatting (3DGS) has emerged as a leading framework for novel view synthesis, yet its core optimization challenges remain underexplored. We identify two key issues in 3DGS optimization: entrapment in suboptim…
Novel View SynthesisLearn the Ropes, Then Trust the Wins: Self-imitation with Progressive Exploration for Agentic Reinforcement Learning
Reinforcement learning (RL) is the dominant paradigm for sharpening strategic tool use capabilities of LLMs on long-horizon, sparsely-rewarded agent tasks, yet it faces a fundamental challenge of exploration-exploitation…
Reinforcement Learning