Optimizing watermarks for large language models

2023/12/28 by Wouters, Bram · 2 citations
#Artificial Intelligence (cs.AI) #Computation and Language (cs.CL) #Cryptography and Security (cs.CR) #FOS: Computer and information sciences

paper · doi:10.48550/arxiv.2312.17295

Abstract

With the rise of large language models (LLMs) and concerns about potential misuse, watermarks for generative LLMs have recently attracted much attention. An important aspect of such watermarks is the trade-off between their identifiability and their impact on the quality of the generated text. This paper introduces a systematic approach to this trade-off in terms of a multi-objective optimization problem. For a large class of robust, efficient watermarks, the associated Pareto optimal solutions are identified and shown to outperform the currently default watermark.

Cited by

• SciPostGen: Bridging the Gap between Scientific Papers and Poster Layouts
• AdaDetectGPT: Adaptive Detection of LLM-Generated Text with Statistical Guarantees

Related

• Three Bricks to Consolidate Watermarks for Large Language Models
• Learning to Watermark: A Selective Watermarking Framework for Large Language Models via Multi-Objective Optimization
• Necessary and Sufficient Watermark for Large Language Models
• Towards Better Statistical Understanding of Watermarking LLMs
• Theoretically Grounded Framework for LLM Watermarking: A Distribution-Adaptive Approach