One paper on theoretical framework for watermarking Large Language Models (LLMs) that jointly optimizes both the watermarking scheme and the detection process is accepted by NeurIPS 2025! Congrats to my collaborators!