The feature selection strategy of few-shot learning based on causal inference

Jining Zhang; Xiaodong Zhu; Yuanning Liu

doi:10.1117/12.2672724

28 February 2023 The feature selection strategy of few-shot learning based on causal inference

Jining Zhang, Xiaodong Zhu, Yuanning Liu

Proceedings Volume 12596, International Conference on Mechatronics Engineering and Artificial Intelligence (MEAI 2022); 125961Z (2023) https://doi.org/10.1117/12.2672724
Event: International Conference on Mechatronics Engineering and Artificial Intelligence (MEAI 2022), 2022, Changsha, China

Abstract

Statistical learning methods require large-scale data to make the significant generalized probability and observation error close to each other. Few-shot learning can alleviate this situation, but it cannot break through the limitations of statistical learning methods. The training model only depends on the correlation between data distributions. There may be potential risks in applying these models to decision-making in the natural environment. This paper studies feature selection in small sample regression analysis based on AutoMPG and MOP The performance of the two datasets on the regression task is first verified through three classical regression analysis models. Then, through the causal inference method, this paper analyzes the causal effect of the relationship between the features in the dataset and finds that two groups of features do not have a causal relationship. Finally, by setting up a simulation environment, this paper illustrates the potential risks of not considering the causal effect in feature selection.

Citation Download Citation

Jining Zhang, Xiaodong Zhu, and Yuanning Liu "The feature selection strategy of few-shot learning based on causal inference", Proc. SPIE 12596, International Conference on Mechatronics Engineering and Artificial Intelligence (MEAI 2022), 125961Z (28 February 2023); https://doi.org/10.1117/12.2672724

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available