LATIDIA · Ciberseguridad

Muestreo especulativo de múltiples borradores con marca de agua a través de procesos de Poisson

arXiv:2609.21858v1 Tipo de anuncio: nuevo Resumen: Los modelos de lenguaje grandes (LLM) han logrado un rendimiento de vanguardia en una amplia gama de tareas, motivando dos aspectos importantes de la implementación: eficiencia de inferencia a

La noticia

arXiv:2609.21858v1 Announce Type: new Abstract: Large language models (LLMs) have achieved state-of-the-art performance across a wide range of tasks, motivating two important aspects of deployment: inference efficiency and output provenance, which can be tackled by speculative sampling and watermarking, respectively. However, recent works have shown that combining these two goals is highly nontrivial and can be potentially impossible. In this work, we develop a novel multi-draft speculative sampling algorithm based on Poisson processes that improves the frontier of this fundamental trade-off. The proposed algorithm has strong sampling efficiency on its own and, more interestingly, is naturally watermarkable: we can embed an unbiased watermark without degrading speculative acceptance. Moreover, our algorithm