{"ID":2837353,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2511.18793","arxiv_id":"2511.18793","title":"NEZHA: A Zero-sacrifice and Hyperspeed Decoding Architecture for Generative Recommendations","abstract":"Generative Recommendation (GR), powered by Large Language Models (LLMs), represents a promising new paradigm for industrial recommender systems. However, their practical application is severely hindered by high inference latency, which makes them infeasible for high-throughput, real-time services and limits their overall business impact. While Speculative Decoding (SD) has been proposed to accelerate the autoregressive generation process, existing implementations introduce new bottlenecks: they typically require separate draft models and model-based verifiers, requiring additional training and increasing the latency overhead. In this paper, we address these challenges with NEZHA, a novel architecture that achieves hyperspeed decoding for GR systems without sacrificing recommendation quality. Specifically, NEZHA integrates a nimble autoregressive draft head directly into the primary model, enabling efficient self-drafting. This design, combined with a specialized input prompt structure, preserves the integrity of sequence-to-sequence generation. Furthermore, to tackle the critical problem of hallucination, a major source of performance degradation, we introduce an efficient, model-free verifier based on a hash set. We demonstrate the effectiveness of NEZHA through extensive experiments on public datasets and have successfully deployed the system on Taobao since October 2025, driving the billion-level advertising revenue and serving hundreds of millions of daily active users.","short_abstract":"Generative Recommendation (GR), powered by Large Language Models (LLMs), represents a promising new paradigm for industrial recommender systems. However, their practical application is severely hindered by high inference latency, which makes them infeasible for high-throughput, real-time services and limits their overa...","url_abs":"https://arxiv.org/abs/2511.18793","url_pdf":"https://arxiv.org/pdf/2511.18793v2","authors":"[\"Yejing Wang\",\"Shengyu Zhou\",\"Jinyu Lu\",\"Ziwei Liu\",\"Langming Liu\",\"Maolin Wang\",\"Wenlin Zhang\",\"Feng Li\",\"Wenbo Su\",\"Pengjie Wang\",\"Jian Xu\",\"Xiangyu Zhao\"]","published":"2025-11-24T05:53:46Z","proceeding":"cs.AI","tasks":"[\"cs.AI\",\"cs.LG\"]","methods":"[\"Large Language Model\",\"Language Model\"]","has_code":false}