用反思式进化搜索，让AI设计更可靠高效的神经网络结构。
RevoNAD: Reflective Evolutionary Exploration for Neural Architecture Design
- 多轮多专家共识提取设计线索，避免盲目生成。
- 根据反馈不确定性动态调整探索强度，提升稳定性。
- 兼顾精度、效率、多样性，适合实际部署的架构设计。
近年来，大语言模型（LLMs）推动了神经网络架构设计（NAD）系统的发展，使其能突破人工预定义搜索空间的限制。然而，基于LLM的生成仍面临挑战：令牌级设计循环是离散且不可微的，难以实现反馈对架构的平滑引导。此类方法常因缺乏有效推理而陷入冗余结构或偏离可行设计。本文提出RevoNAD，一种反思式进化协同器，有效连接基于LLM的推理与反馈对齐的架构搜索。首先，通过多轮多专家共识将零散的设计规则转化为有意义的架构线索；其次，自适应反思探索根据奖励方差调节探索程度，在反馈不确定时探索，在稳定时精炼；最后，帕累托引导的进化选择促进在准确率、效率、延迟、置信度和结构多样性上协同优化的架构。在CIFAR10、CIFAR100、ImageNet16-120、COCO-5K和Cityscape数据集上，RevoNAD均达到当前最优性能。消融实验与迁移研究进一步验证了其在实际可信赖、可部署架构设计中的有效性。
原文摘要 · Abstract (English)
Recent progress in leveraging large language models (LLMs) has enabled Neural Architecture Design (NAD) systems to generate new architecture not limited from manually predefined search space. Nevertheless, LLM-driven generation remains challenging: the token-level design loop is discrete and non-differentiable, preventing feedback from smoothly guiding architectural improvement. These methods, in turn, commonly suffer from mode collapse into redundant structures or drift toward infeasible designs when constructive reasoning is not well grounded. We introduce RevoNAD, a reflective evolutionary orchestrator that effectively bridges LLM-based reasoning with feedback-aligned architectural search. First, RevoNAD presents a Multi-round Multi-expert Consensus to transfer isolated design rules into meaningful architectural clues. Then, Adaptive Reflective Exploration adjusts the degree of exploration leveraging reward variance; it explores when feedback is uncertain and refines when stability is reached. Finally, Pareto-guided Evolutionary Selection effectively promotes architectures that jointly optimize accuracy, efficiency, latency, confidence, and structural diversity. Across CIFAR10, CIFAR100, ImageNet16-120, COCO-5K, and Cityscape, RevoNAD achieves state-of-the-art performance. Ablation and transfer studies further validate the effectiveness of RevoNAD in allowing practically reliable, and deployable neural architecture design.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。