AI & ML
Plug-and-Play Multi-Licensee Robust Watermarking for Mixture-of-Experts LLMs
arXiv cs.CR0 views
arXiv:2611.23838v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) architectures enable high-capacity large language models (LLMs) with high sparse computation, making them attractive for deployment on private hardware. However, distributing model weights exposes valuable intellectual property (IP) to leakage and unauthorized redistribution, creating an urgent need to both verify model ownership and trace leaked copies to individual licensees. Existing model-watermarking techniques are ill-suited for meeting this need because they are fragile under MoE's dynamic routing or require costly retraining to embed a distinct watermark for each licensee, thus making scalable distribution difficult. To address these challenges, we present PPMoE, the first plug-and-play multi-licensee watermarking framework for MoE LLMs. Rather than training a separate model per user, PPMoE trains a compact set of mergeable incremental weight modules that steer the MoE router toward distinct, mutually compatible routing patterns. During model distribution, the model owner selectively merges unique combinations of these pre-trained, mutually compatible modules into the base model on the fly, embedding a licensee-specific watermark without additional training. PPMoE supports reliable multi-licensee attribution under both white- and black-box threat models. In the white-box setting, attribution is done by inspecting triggered routing decisions or the merged weight modules responsible for these decisions, whereas in the black-box setting, attribution relies on analyzing output token likelihood ratios. Extensive experiments across five state-of-the-art MoE models, seven benchmark datasets, and varying number of licensees show that PPMoE achieves very high attribution accuracy while preserving base model utility and maintaining strong robustness to various adaptive watermark-removal attacks.
Source:https://arxiv.org/abs/2611.23838PreviousSlice Isolation-as-a-Service: An Isolation Mechanism to Proactively Defend Cross-Slice Attacks in 5G-Advanced and BeyondNextWeighing the Envelope: Component Weight and Arbitration in Envelope-Layer Defenses for a Single LLM Agent
Related posts

Why Turn-Based Voice AI Feels Broken: Building a 52ms Barge-In Loop (HardCall)
0dev.to
Harnesses! The other kind
0dev.to
Apple may be developing its own AI agent that runs your Mac
0TechCrunch

Cosmos 3 Edge, Part 3: SO-101 Sim-to-Real and Robot Control on Jetson Thor
0dev.to
Seedance, Kling and Veo 3.1 API prices per second, compared across 6 sellers (November 2026)
0dev.to
Free AI Headshot Generator in 2026: What's Actually Free (and Why None Train on Your Face)
0dev.to Career