AI Autonomy: Self-initiated Open-world Continual Learning and Adaptation

被引:10
|
作者
Liu, Bing [1 ]
Mazumder, Sahisnu [2 ]
Robertson, Eric [3 ]
Grigsby, Scott [3 ]
机构
[1] Univ Illinois, Dept Comp Sci, Chicago, IL 60607 USA
[2] Intel Labs, Intelligent Syst Res, Santa Clara, CA USA
[3] PAR Govt Syst Corp, New York, NY USA
基金
美国国家科学基金会;
关键词
NETWORKS;
D O I
10.1002/aaai.12087
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
As more and more AI agents are used in practice, it is time to think about how to make these agents fully autonomous so that they can (1) learn by themselves continually in a self-motivated and self-initiated manner rather than being retrained offline periodically on the initiation of human engineers and (2) accommodate or adapt to unexpected or novel circumstances. As the real-world is an open environment that is full of unknowns or novelties, the capabilities of detecting novelties, characterizing them, accommodating/adapting to them, and gathering ground-truth training data and incrementally learning the unknowns/novelties become critical in making the AI agent more and more knowledgeable, powerful and self-sustainable over time. The key challenge here is how to automate the process so that it is carried out continually on the agent's own initiative and through its own interactions with humans, other agents and the environment just like human on-the-job learning. This paper proposes a framework (called SOLA) for this learning paradigm to promote the research of building autonomous and continual learning enabled AI agents. To show feasibility, an implemented agent is also described.
引用
收藏
页码:185 / 199
页数:15
相关论文
共 50 条
  • [41] Categorizing and Comparing Behaviors of Students Engaged in Self-initiated Learning Online
    Inventado, Paul Salvador
    Legaspi, Roberto
    Suarez, Merlin
    Numao, Masayuki
    THEORY AND PRACTICE OF COMPUTATION, 2012, 5 : 133 - 144
  • [42] Fluidity in the Networked Society - Self-initiated learning as a Digital Literacy Competence
    Levinsen, Karin Tweddell
    ELECTRONIC JOURNAL OF E-LEARNING, 2011, 9 (01): : 52 - 62
  • [43] Learning from self-initiated community forest management in Orissa, India
    Conroy, C
    Mishra, A
    Rai, A
    FOREST POLICY AND ECONOMICS, 2002, 4 (03) : 227 - 237
  • [44] Beyond the Known: Novel Class Discovery for Open-World Graph Learning
    Jin, Yucheng
    Xiong, Yun
    Fang, Juncheng
    Wu, Xixi
    He, Dongxiao
    Jia, Xing
    Zhao, Bingchen
    Yu, Philip S.
    DATABASE SYSTEMS FOR ADVANCED APPLICATIONS, PT VI, DASFAA 2024, 2024, 14855 : 117 - 133
  • [45] OpenWGL: open-world graph learning for unseen class node classification
    Man Wu
    Shirui Pan
    Xingquan Zhu
    Knowledge and Information Systems, 2021, 63 : 2405 - 2430
  • [46] A Simple Baseline for Open-World Tracking via Self-training
    Wang, Bingyang
    Li, Tanlin
    Wu, Jiannan
    Jiang, Yi
    Lu, Huchuan
    He, You
    PROCEEDINGS OF THE 31ST ACM INTERNATIONAL CONFERENCE ON MULTIMEDIA, MM 2023, 2023, : 2765 - 2774
  • [47] Dual Confidence Learning Network for Open-World Time Series Classification
    Lv, Junwei
    He, Ying
    Hu, Xuegang
    Cai, Desheng
    Chu, Yuqi
    Hu, Jun
    DATABASE SYSTEMS FOR ADVANCED APPLICATIONS, DASFAA 2022, PT II, 2022, : 574 - 589
  • [48] Confidence-Guided Open-World Semi-supervised Learning
    Li, Jibang
    Yang, Meng
    Feng, Mao
    PATTERN RECOGNITION AND COMPUTER VISION, PRCV 2023, PT IV, 2024, 14428 : 87 - 99
  • [49] Raman Spectroscopy in Open-World Learning Settings Using the Objectosphere Approach
    Balytskyi, Yaroslav
    Bendesky, Justin
    Paul, Tristan
    Hagen, Guy M.
    McNear, Kelly
    ANALYTICAL CHEMISTRY, 2022, 94 (44) : 15297 - 15306
  • [50] Open-world structured sequence learning via dense target encoding
    Zhang, Qin
    Liu, Ziqi
    Li, Qincai
    Xiang, Haolong
    Yu, Zhizhi
    Chen, Junyang
    Zhang, Peng
    Chen, Xiaojun
    INFORMATION SCIENCES, 2024, 680