Issue |
ITM Web Conf.
Volume 70, 2025
2024 2nd International Conference on Data Science, Advanced Algorithm and Intelligent Computing (DAI 2024)
|
|
---|---|---|
Article Number | 02006 | |
Number of page(s) | 8 | |
Section | Machine Learning in Healthcare and Finance | |
DOI | https://doi.org/10.1051/itmconf/20257002006 | |
Published online | 23 January 2025 |
Research Progress of Causal Inference in Bias Elimination
College of Computer and Big Data, Fuzhou University, Fuzhou, 350000, Fujian, China
Corresponding author: ronghong@ldy.edu.rs
The natural language processing (NLP) models have recently gained widespread attention and are increasingly being applied to real-world tasks. However, due to the presence of bias, the application of NLP models in specialized fields has led to various issues. This paper introduces various types of biases and. through a comparative analysis of multiple existing methods, explains why previous approaches cannot fundamentally solve the bias problem. Additionally, this paper proposes that using causal inference to eliminate bias is an advanced method, and research teams that reduce bias by building causal relationship models have already studied this approach. This paper conducts a detailed analysis of several cutting-edge studies, exploring the practical application of causal inference in bias elimination and the challenges involved. Experimental results indicate that although causal inference methods have eliminated bias to some extent, further research and optimization are still needed. Finally, this paper summarizes the previous sections and provides an outlook on future research directions.
© The Authors, published by EDP Sciences, 2025
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.