PILOT:a pre-trained model-based continual learning toolbox  

作　　者：Hai-Long SUN Da-Wei ZHOU De-Chuan ZHAN Han-Jia YE 

机构地区：[1]School of Artificial Intelligence,Nanjing University,Nanjing 210023,China [2]National Key Laboratory for Novel Software Technology,Nanjing University,Nanjing 210023,China

出　　处：《Science China(Information Sciences)》2025年第4期379-380,共2页中国科学(信息科学)(英文版)

基　　金：partially supported by National Key R&D Program of China(Grant No.2022ZD0114805);National Natural Science Foundation of China(Grant Nos.62476123,62376118,62006112,62250069,61921006);Fundamental Research Funds for the Central Universities(Grant Nos.2024300373,14380021);Key Program of Jiangsu Science Foundation(Grant No.BK20243012);CCF-Tencent Rhino-Bird Open Research Fund(Grant No.RAGR20240101);AI&AI for Science Project of Nanjing University,Collaborative Innovation Center of Novel Software Technology and Industrialization。

摘　　要：The rapid advancements in deep learning have resulted in significant achievements across various fields.However,our ever-changing world often presents training data in a streaming format from an open environment.For example,while ChatGPT demonstrates exceptional inference capabilities,it struggles to provide users with the most up-to-date information.This challenge arises from the high costs associated with retraining a GPT model on new data daily.

关 键 词：deep learning PILOT continual learning training data pre trained model streaming format open environment 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]