paper-with-me

홈 › Papers

Compacting, Picking and Growing for Unforgetting Continual Learning

2019-10-15 · NeurIPS 2019 12 · Steven C. Y. Hung, Cheng-Hao Tu, Cheng-En Wu, Chien-Hung Chen, Yi-Ming Chan, Chu-Song Chen

Continual lifelong learning is essential to many applications. In this paper, we propose a simple but effective approach to continual deep learning. Our approach leverages the principles of deep model compression, critical weights selection, and progressive networks expansion. By enforcing their integration in an iterative manner, we introduce an incremental learning method that is scalable to the number of sequential tasks in a continual learning process. Our approach is easy to implement and owns several favorable characteristics. First, it can avoid forgetting (i.e., learn new tasks while remembering all previous tasks). Second, it allows model expansion but can maintain the model compactness when handling sequential tasks. Besides, through our compaction and selection/expansion mechanism, we show that the knowledge accumulated through learning previous tasks is helpful to build a better model for the new tasks compared to training the models independently with tasks. Experimental results show that our approach can incrementally learn a deep model tackling multiple tasks without forgetting, while the model compactness is maintained with the performance more satisfiable than individual task training.

📄 PDF Abstract BibTeX arXiv:1910.06562

Code (2)

ivclab/CPG 공식 구현 pytorch
p0werweirdo/tagfcl pytorch

Tasks

Age And Gender ClassificationContinual LearningFace VerificationFacial Expression Recognition (FER)Incremental LearningLifelong learningModel Compression

Similar Papers 제목 키워드 기반

PRANC: Pseudo RAndom Networks for Compacting deep models

2022-06-16 · ICCV 2023 1 · Parsa Nooralinejad, Ali Abbasi, Soroush Abbasi Koohpayegani, Kossar Pourahmadi Meibodi 외

We demonstrate that a deep model can be reparametrized as a linear combination of several randomly initialized and frozen deep models in the weight space. During training, we seek local minima that reside within the subs…

image-classificationImage Classification

EXTENDING CONDITIONAL CONVOLUTION STRUCTURES FOR ENHANCING MULTITASKING CONTINUAL LEARNING

2020-12-07 · Cheng-Hao Tu, Cheng-En Wu, Chu-Song Chen

Conditional operations have received much attention in recent deep learning studies to facilitate the prediction accuracy of a model. A recent advance toward this direction is the conditional parametric convolutions (Con…

Computational EfficiencyContinual LearningMulti-Task Learning

Compacting Deep Neural Networks for Internet of Things: Methods and Applications

2021-03-20 · Ke Zhang, Hanbo Ying, Hong-Ning Dai, Lin Li 외

Deep Neural Networks (DNNs) have shown great success in completing complex tasks. However, DNNs inevitably bring high computational cost and storage consumption due to the complexity of hierarchical structures, thereby h…

DiversityKnowledge DistillationModel Compression

Gradually Compacting Large Language Models for Reasoning Like a Boiling Frog

2026-02-04 · Yiran Zhao, Shengyang Zhou, Zijian Wu, Tongyan Hu 외 arxiv

Large Language Models (LLMs) have demonstrated impressive reasoning capabilities, but their substantial size often demands significant computational resources. To reduce resource consumption and accelerate inference, it …

Reinforcement LearningMathematical ReasoningCode Generation

Autonomous Strawberry Picking Robotic System (Robofruit)

2023-01-10 · Soran Parsa, Bappaditya Debnath, Muhammad Arshad Khan, Amir Ghalamzan E.

Challenges in strawberry picking made selective harvesting robotic technology demanding. However, selective harvesting of strawberries is complicated forming a few scientific research questions. Most available solutions …