New HiTMS framework boosts LLM linguistic steganography throughput

Researchers have developed HiTMS, a novel framework designed to improve the efficiency and security of linguistic steganography using large language models. This new method allows for the distribution of secret data across multiple responses generated in successive interaction rounds, significantly increasing throughput compared to single-stream approaches. HiTMS also incorporates a self-describing frame and a key-derived schedule to conceal the number of active streams and ensure exact recovery of information, reducing the detectability by steganalyzers. AI

IMPACT Enhances data security and privacy by improving the efficiency of concealing information within LLM outputs.

RANK_REASON The item describes a new framework for linguistic steganography presented in an arXiv paper. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• alphaXiv
• arXiv
• CatalyzeX
• DagsHub
• GitHub
• Gotit.pub
• HiTMS
• Hugging Face
• ScienceCast

• paper
• other

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