9 papers
Not Only NTP: Extending Training Signal Coverage for Generative Recommendation
Changhao Li, Shuli Wang, Junwei Yin +6
Next-Token Prediction (NTP) carries two structural training signal limitations. First, NTP optimizes for single-step prediction only, placing no supervised pressure on learning lon…
Revisiting DAgger in the Era of LLM-Agents
Changhao Li, Rushi Qiang, Jiawei Huang +4
Long-horizon LM agents learn from multi-turn interaction, where a single early mistake can alter the subsequent state distribution and derail the whole trajectory. Existing recipes…
Exploration-Driven Optimization for Test-Time Large Language Model Reasoning
Changhao Li, Yuchen Zhuang, Chenxiao Gao +4
Post-training techniques combined with inference-time scaling significantly enhance the reasoning and alignment capabilities of large language models (LLMs). However, a fundamental…
Next-Scale Generative Reranking: A Tree-based Generative Rerank Method at Meituan
Shuli Wang, Changhao Li, Ke Fan +5
In modern multi-stage recommendation systems, reranking plays a critical role by modeling contextual information. Due to inherent challenges such as the combinatorial space complex…
MBGR: Multi-Business Prediction for Generative Recommendation at Meituan
Changhao Li, Junwei Yin, Zhilin Zeng +6
Generative recommendation (GR) has recently emerged as a promising paradigm for industrial recommendations. GR leverages Semantic IDs (SIDs) to reduce the encoding-decoding space a…
Sample Is Feature: Beyond Item-Level, Toward Sample-Level Tokens for Unified Large Recommender Models
Shuli Wang, Junwei Yin, Changhao Li +6
Scaling industrial recommender models has followed two parallel paradigms: \textbf{sample information scaling} -- enriching the information content of each training sample through…
DOS: Dual-Flow Orthogonal Semantic IDs for Recommendation in Meituan
Junwei Yin, Senjie Kou, Changhao Li +6
Semantic IDs serve as a key component in generative recommendation systems. They not only incorporate open-world knowledge from large language models (LLMs) but also compress the s…
You Only Evaluate Once: A Tree-based Rerank Method at Meituan
Shuli Wang, Yinqiu Huang, Changhao Li +6
Reranking plays a crucial role in modern recommender systems by capturing the mutual influences within the list. Due to the inherent challenges of combinatorial search spaces, most…
Entire Chain Uplift Modeling with Context-Enhanced Learning for Intelligent Marketing
Yinqiu Huang, Shuli Wang, Min Gao +6
Uplift modeling, vital in online marketing, seeks to accurately measure the impact of various strategies, such as coupons or discounts, on different users by predicting the Individ…