Survey maps fake review detection from PLMs to LLMs

A new survey paper examines the evolving landscape of fake review detection, tracing methods from traditional machine learning to advanced techniques utilizing pre-trained language models (PLMs) and large language models (LLMs). The paper organizes 211 studies from 2018 to early 2026 based on evidence sources and fusion levels, including text, sentiment, user behavior, temporal data, and multimodal content. It analyzes performance trends on benchmark datasets like Amazon and Yelp, while also highlighting limitations in current evaluation procedures and identifying open challenges such as adversarial generation, cross-domain transfer, and interpretability for AI-generated deceptive content. AI

IMPACT Provides a comprehensive overview of AI techniques for combating deceptive online content, aiding researchers and platforms in developing more robust detection systems.

RANK_REASON The item is a survey paper published on arXiv detailing research on fake review detection methods. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• arXiv
• Fake Review Detection
• Large Language Models
• Pre-trained Language Models
• Yelp

• paper
• other

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