Saved in:
Bibliographic Details
Main Authors: Sun, Yiqun, Huang, Qiang, Tung, Anthony K. H., Yu, Jun
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2505.24646
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908386174435328
author Sun, Yiqun
Huang, Qiang
Tung, Anthony K. H.
Yu, Jun
author_facet Sun, Yiqun
Huang, Qiang
Tung, Anthony K. H.
Yu, Jun
contents Semantic Text Embedding is a fundamental NLP task that encodes textual content into vector representations, where proximity in the embedding space reflects semantic similarity. While existing embedding models excel at capturing general meaning, they often overlook ideological nuances, limiting their effectiveness in tasks that require an understanding of political bias. To address this gap, we introduce PRISM, the first framework designed to Produce inteRpretable polItical biaS eMbeddings. PRISM operates in two key stages: (1) Controversial Topic Bias Indicator Mining, which systematically extracts fine-grained political topics and their corresponding bias indicators from weakly labeled news data, and (2) Cross-Encoder Political Bias Embedding, which assigns structured bias scores to news articles based on their alignment with these indicators. This approach ensures that embeddings are explicitly tied to bias-revealing dimensions, enhancing both interpretability and predictive power. Through extensive experiments on two large-scale datasets, we demonstrate that PRISM outperforms state-of-the-art text embedding models in political bias classification while offering highly interpretable representations that facilitate diversified retrieval and ideological analysis. The source code is available at https://github.com/dukesun99/ACL-PRISM.
format Preprint
id arxiv_https___arxiv_org_abs_2505_24646
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle PRISM: A Framework for Producing Interpretable Political Bias Embeddings with Political-Aware Cross-Encoder
Sun, Yiqun
Huang, Qiang
Tung, Anthony K. H.
Yu, Jun
Computation and Language
Semantic Text Embedding is a fundamental NLP task that encodes textual content into vector representations, where proximity in the embedding space reflects semantic similarity. While existing embedding models excel at capturing general meaning, they often overlook ideological nuances, limiting their effectiveness in tasks that require an understanding of political bias. To address this gap, we introduce PRISM, the first framework designed to Produce inteRpretable polItical biaS eMbeddings. PRISM operates in two key stages: (1) Controversial Topic Bias Indicator Mining, which systematically extracts fine-grained political topics and their corresponding bias indicators from weakly labeled news data, and (2) Cross-Encoder Political Bias Embedding, which assigns structured bias scores to news articles based on their alignment with these indicators. This approach ensures that embeddings are explicitly tied to bias-revealing dimensions, enhancing both interpretability and predictive power. Through extensive experiments on two large-scale datasets, we demonstrate that PRISM outperforms state-of-the-art text embedding models in political bias classification while offering highly interpretable representations that facilitate diversified retrieval and ideological analysis. The source code is available at https://github.com/dukesun99/ACL-PRISM.
title PRISM: A Framework for Producing Interpretable Political Bias Embeddings with Political-Aware Cross-Encoder
topic Computation and Language
url https://arxiv.org/abs/2505.24646