New embedding models improve stance-aware argument retrieval

Researchers have developed new embedding models designed to improve stance-aware argument retrieval, a crucial step for downstream reasoning tasks. Existing models often prioritize topical relevance over correctly identifying whether an argument supports or attacks a claim. The study found that while contrastive training can help, it can also lead models to overemphasize polarity keywords, neglecting the semantic topic. To address this, a data-centric approach using a balanced argument curriculum and LLM-augmented, stance-inverted arguments was proposed, which helps models learn deeper directional logic and improves stance-aware argument retrieval. AI

IMPACT Enhances argument retrieval accuracy, potentially improving AI reasoning capabilities.

RANK_REASON The cluster contains a research paper detailing new methods for embedding models. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• alphaXiv
• arXiv
• CatalyzeX
• Connected Papers
• CORE Recommender
• DagsHub
• Embedding Models for Stance-Aware Argument Retrieval
• Gotit.pub
• Hugging Face
• Influence Flower
• Litmaps
• ScienceCast
• scite Smart Citations

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →