TGR-EvoDAG is proposed, a multi-agent code evolution framework tailored for GR that enables the first autonomous evolution of a complete GR pipeline, and introduces stage-aware cost planning to optimize GPU-hour budgets, and active downstream co-adaptation to navigate temporary regressions and unlock cross-stage gains.
Abstract
Large Language Model (LLM) agents have enabled automated code evolution for algorithm discovery. However, current recursive self-improvement (RSI) research for sequential recommendation faces a dilemma: general-purpose frameworks lack domain-specific feedback, while recommender-specific frameworks rely on linear evolution. Crucially, naively integrating them fails when applied to Generative Recommendation (GR). Unlike end-to-end models, GR relies on a module-cascaded pipeline (tokenization, pre-training, and alignment). In this paradigm, existing evolution methods suffer from: i) cost blindness, squandering budgets by ignoring stage-asymmetric retraining overheads; and ii) short-sighted evaluation, discarding promising upstream innovations that temporarily regress metrics before downstream adaptation. To bridge these gaps, we propose TGR-EvoDAG, a multi-agent code evolution framework tailored for GR. TGR-EvoDAG models code evolution as a Directed Acyclic Graph (DAG) to maintain multiple evolution branches and merge complementary innovations, unifying advanced search with rich GR diagnostic feedback. We further introduce stage-aware cost planning to optimize GPU-hour budgets, and active downstream co-adaptation to navigate temporary regressions and unlock cross-stage gains. On public Amazon datasets, TGR-EvoDAG enables the first autonomous evolution of a complete GR pipeline, improving offline metrics by up to 37.4% and Agent4Rec simulated user satisfaction by 19.7% over the strongest baseline. Furthermore, in industrial deployment on Tencent's large-scale commercial media platform, the evolved GR achieves a 12.16% relative improvement in offline Hit@1 and a statistically significant 0.40% increase in online click-through rate (CTR).
Results characterize when runtime-controlled composition of coding-agent products improves execution accuracy, as well as when runtime-controlled composition of coding-agent products improves execution accuracy.
Zheng-Yu Chen, Lin-Feng Liu, Hong Li et al.· arXiv.org· 0 citations
This work introduces EvoSkillRec, a promotion-and-reuse framework for cumulative recommender architecture evolution that decomposes recommenders into atomic executable skills and represents architectures as typed skill genomes, with each skill equipped with input--output types, semantic annotations, and implementation...
Xiao-Peng Li, Kuo Cai, Bo Chen et al.· arXiv.org· 1 citation
This work proposes SLCA-GRPO, a framework incorporating Segment-Locked Credit Assignment (SLCA), a framework incorporating Segment-Locked Credit Assignment (SLCA) that decouples advantage estimation at the structural segment level within a single group of rollouts, without requiring additional rollouts from intermediat...
Ya-Ge Zhan, Shao-Bo Liu, Qiu-Nan Liu et al.· arXiv.org· 1 citation
Diffusion large language models (dLLMs) have advanced rapidly, yet strengthening their reasoning through reinforcement learning (RL) remains an open challenge. Any-order generation can suffer from entropy degradation, which narrows the reasoning paths available for exploration. Recent work mitigates this issue by apply...
Zhao-Yang Jia, Cheng-Wei Zhou, G. Datta et al.· 0 citations
Automated agentic workflow optimization relies on costly evaluations, making it essential to allocate a limited evaluation budget effectively. Multi-parent fusion can reuse designs from previously discovered workflows, but identifying promising parent combinations requires learning from limited fusion feedback. We intr...
Zhi-Wei Shang, Jia-Hang Sun, Ming-Rong Gong et al.· 0 citations
Training-free self-reflective protocols (generate, critique, revise) improve LLM reasoning but operate on a flat token stream: revisions cannot identify which prior step is unreliable, and stopping cannot exploit structural convergence. We introduce GraphEvoResearcher, a training-free protocol that lifts EvoResearcher'...
Su-Xing Liu, Byungwon Min, Zhi-Jian Zheng et al.· 0 citations