12 papers
On-the-Fly Adaptation to Quantization: Configuration-Aware LoRA for Efficient Fine-Tuning of Quantized LLMs
Rongguang Ye, Ming Tang, Edith C. H. Ngai
As increasingly large pre-trained models are released, deploying them on edge devices for privacy-preserving applications requires effective compression. Recent works combine quant…
OmniISR: A Unified Framework for Centralized and Federated Learning via Intermediate Supervision and Regularization
Wei-Bin Kou, Guangxu Zhu, Ming Tang +4
The global deployment of edge intelligence operates across heterogeneous legal frameworks. While some regions permit centralized learning (CL) via cloud data aggregation, others en…
Adverse Weather-Independent Framework Towards Autonomous Driving Perception through Temporal Correlation and Unfolded Regularization
Wei-Bin Kou, Guangxu Zhu, Rongguang Ye +5
Various adverse weather conditions such as fog and rain pose a significant challenge to autonomous driving (AD) perception tasks like semantic segmentation, object detection, etc.…
iMacHSR: Intermediate Multi-Access Heterogeneous Supervision and Regularization Scheme Toward Architecture-Agnostic Training
Wei-Bin Kou, Guangxu Zhu, Yichen Jin +4
While deep supervision is a powerful training strategy by supervising intermediate layers with auxiliary losses, it faces three underexplored problems: (I) Existing deep supervisio…
Enhancing Large Vision Model in Street Scene Semantic Understanding through Leveraging Posterior Optimization Trajectory
Wei-Bin Kou, Qingfeng Lin, Ming Tang +5
To improve the generalization of the autonomous driving (AD) perception model, vehicles need to update the model over time based on the continuously collected data. As time progres…
One-for-All Pruning: A Universal Model for Customized Compression of Large Language Models
Rongguang Ye, Ming Tang
Existing pruning methods for large language models (LLMs) focus on achieving high compression rates while maintaining model performance. Although these methods have demonstrated sa…