Power Equipment ChatGPT-type Model and Key Technologies

被引：0
|
作者：
Jiang X. [1 ]
Zang Y. [1 ]
Liu Y. [1 ]
Sheng G. [1 ]
Xu Y. [1 ]
Qian Q. [1 ]
机构：
[1] School of Electronic Information and Electrical Engineering, Shanghai Jiaotong University, Shanghai
来源：
Gaodianya Jishu/High Voltage Engineering | 2023年 / 49卷 / 10期
关键词：
digital twin; general artificial intelligence; power equipment ChatGPT; reinforcement learning from human feedback; transformer models;
D O I：
10.13336/j.1003-6520.hve.20231419
中图分类号：
学科分类号：
摘要：
ChatGPT (chat generative pre-trained transformer) is a new technology direction developed in the field of artificial intelligence in recent years, which covers digital functions such as device digital twin, device management, platform operation, etc., and is more characterized by versatility and generative human-machine dialogue. This paper firstly introduces the development status of ChatGPT, as well as the power equipment ChatGPT-type model and core technology architecture. It is illustrated that the large model has outstanding features such as excellent generalization ability, logical reasoning ability, and multimodal data analysis and generation ability. Then, the key technologies involved in ChatGPT-type large model for electric power equipment are analyzed from five aspects as follows: high arithmetic AI chip, corpus sample system construction, transformer-based generative pre-training model, multimodal algorithm embedded in big language model, and human feedback-based reinforcement learning technology. Finally, the feasibility and technical solutions for ChatGPT for power equipment to be carried out in the power industry are proposed, and the challenges and development directions of ChatGPT for power equipment in the future are summarized. © 2023 Science Press. All rights reserved.
引用
收藏
页码：4033 / 4045
页数：12
相关论文
共 38 条
  • [1] HAN X, ZHANG Z Y, DING N, Et al., Pre-trained models: past, present and future, AI Open, 2, pp. 225-250, (2021)
  • [2] PETERS M A, JACKSON L, PAPASTEPHANOU M, Et al., AI and the future of humanity: ChatGPT-4, philosophy and education-critical responses[J/OL], Educational Philosophy and Theory, pp. 1-35, (2023)
  • [3] ZHOU C, LI Q, LI C, Et al., A comprehensive survey on pretrained foundation models: a history from BERT to ChatGPT, (2023)
  • [4] LIU Y H, HAN T L, MA S Y, Et al., Summary of ChatGPT/GPT-4 research and perspective towards the future of large language models, (2023)
  • [5] GONG T, LYU C Q, ZHANG S L, Et al., MultiModal-GPT: a vision and language model for dialogue with humans, (2023)
  • [6] CAO Y H, LI S Y, LIU Y X, Et al., A comprehensive survey of AI-generated content (AIGC): a history of generative AI from GAN to ChatGPT, (2023)
  • [7] BUBECK S, CHANDRASEKARAN V, ELDAN R, Et al., Sparks of artificial general intelligence: early experiments with GPT-4, (2023)
  • [8] ELSE H., Abstracts written by ChatGPT fool scientists, Nature, 613, 7944, pp. 423-423, (2023)
  • [9] PLEVRIS V, PAPAZAFEIROPOULOS G, RIOS A J., Chatbots put to the test in math and logic problems: a preliminary comparison and assessment of ChatGPT-3.5, ChatGPT-4, and Google bard, (2023)
  • [10] GPT-4 technical report, (2023)
1234