Tag
This paper introduces Relational Over-Regularization (ROR) as a structural signal for detecting AI-generated text and proposes the Cross-Source Stylometric Fingerprint Graph (CSFG) framework to exploit this signal via graph-based methods, achieving high accuracy in detection.
MD-ProTector is a research paper proposing a prototype-based method for LLM-generated text detection, using multiple trainable reference vectors per class to capture within-class variation and improve robustness across domains, generators, languages, and adversarial edits.
Pangram raises $9 million to detect AI-generated content, launching new text and image detection models with over 99% accuracy for identifying AI-assisted writing.
PP-OCRv6 is a lightweight OCR model (34.5M parameters) that challenges large VLMs with its MetaFormer architecture, offering efficient text detection and recognition across multiple deployment scenarios.
PP-OCRv6 is the latest generation of PaddleOCR's universal OCR model family, offering three tiers from 1.5M to 34.5M parameters, supporting 50 languages, and achieving significant accuracy improvements over previous versions.
A research paper finds that base language models appear human to AI detectors, unlike instruction-tuned models. The authors propose a paraphrasing pipeline (HIP) that improves human-likeness while preserving semantics across model sizes.