Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2023
|
| Schlagworte: | |
| Online-Zugang: | https://arxiv.org/abs/2310.19852 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866917976377131008 |
|---|---|
| author | Ji, Jiaming Qiu, Tianyi Chen, Boyuan Zhang, Borong Lou, Hantao Wang, Kaile Duan, Yawen He, Zhonghao Vierling, Lukas Hong, Donghai Zhou, Jiayi Zhang, Zhaowei Zeng, Fanzhi Dai, Juntao Pan, Xuehai Ng, Kwan Yee O'Gara, Aidan Xu, Hua Tse, Brian Fu, Jie McAleer, Stephen Yang, Yaodong Wang, Yizhou Zhu, Song-Chun Guo, Yike Gao, Wen |
| author_facet | Ji, Jiaming Qiu, Tianyi Chen, Boyuan Zhang, Borong Lou, Hantao Wang, Kaile Duan, Yawen He, Zhonghao Vierling, Lukas Hong, Donghai Zhou, Jiayi Zhang, Zhaowei Zeng, Fanzhi Dai, Juntao Pan, Xuehai Ng, Kwan Yee O'Gara, Aidan Xu, Hua Tse, Brian Fu, Jie McAleer, Stephen Yang, Yaodong Wang, Yizhou Zhu, Song-Chun Guo, Yike Gao, Wen |
| contents | AI alignment aims to make AI systems behave in line with human intentions and values. As AI systems grow more capable, so do risks from misalignment. To provide a comprehensive and up-to-date overview of the alignment field, in this survey, we delve into the core concepts, methodology, and practice of alignment. First, we identify four principles as the key objectives of AI alignment: Robustness, Interpretability, Controllability, and Ethicality (RICE). Guided by these four principles, we outline the landscape of current alignment research and decompose them into two key components: forward alignment and backward alignment. The former aims to make AI systems aligned via alignment training, while the latter aims to gain evidence about the systems' alignment and govern them appropriately to avoid exacerbating misalignment risks. On forward alignment, we discuss techniques for learning from feedback and learning under distribution shift. On backward alignment, we discuss assurance techniques and governance practices.
We also release and continually update the website (www.alignmentsurvey.com) which features tutorials, collections of papers, blog posts, and other resources. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2310_19852 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | AI Alignment: A Comprehensive Survey Ji, Jiaming Qiu, Tianyi Chen, Boyuan Zhang, Borong Lou, Hantao Wang, Kaile Duan, Yawen He, Zhonghao Vierling, Lukas Hong, Donghai Zhou, Jiayi Zhang, Zhaowei Zeng, Fanzhi Dai, Juntao Pan, Xuehai Ng, Kwan Yee O'Gara, Aidan Xu, Hua Tse, Brian Fu, Jie McAleer, Stephen Yang, Yaodong Wang, Yizhou Zhu, Song-Chun Guo, Yike Gao, Wen Artificial Intelligence AI alignment aims to make AI systems behave in line with human intentions and values. As AI systems grow more capable, so do risks from misalignment. To provide a comprehensive and up-to-date overview of the alignment field, in this survey, we delve into the core concepts, methodology, and practice of alignment. First, we identify four principles as the key objectives of AI alignment: Robustness, Interpretability, Controllability, and Ethicality (RICE). Guided by these four principles, we outline the landscape of current alignment research and decompose them into two key components: forward alignment and backward alignment. The former aims to make AI systems aligned via alignment training, while the latter aims to gain evidence about the systems' alignment and govern them appropriately to avoid exacerbating misalignment risks. On forward alignment, we discuss techniques for learning from feedback and learning under distribution shift. On backward alignment, we discuss assurance techniques and governance practices. We also release and continually update the website (www.alignmentsurvey.com) which features tutorials, collections of papers, blog posts, and other resources. |
| title | AI Alignment: A Comprehensive Survey |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2310.19852 |