用大模型自动配置神经网络通道，发现更高效且有规律的结构。
Scaling Closed-Loop Feature Channel Configuration with LLMs
- 通过代码生成与精度反馈闭环优化通道配置
- 250个候选模型中最佳准确率达36.76%，参数量降至1180万
- 揭示非2的幂通道宽度与中后段扩张的结构规律
基于大语言模型的闭环通道配置搜索初现成效，表明可通过可执行代码生成与精度反馈直接优化神经网络宽度。然而此前结果来自较稀疏的有效评估，尚不确定该优化行为能否推广至更密集采样，并且在更多生成网络下是否出现新的架构规律。为此，将同一搜索设置扩展至每轮微调周期生成250个候选网络，共完成8轮完整循环，分析2000个生成候选，经任务与元数据过滤后获得462次验证的CIFAR-100评估。每轮平均准确率呈现正线性趋势，斜率为9.87e-4（p=0.043）；表现最优的前沿提升更为显著：最高准确率由0.3144升至0.3676，且前五和前十轮平均值均呈上升趋势。大规模实验还揭示了更高的参数效率：最优模型达到0.3676准确率仅需1180万参数，远低于早期高性能模型（0.3144准确率，1665万参数）。此外，更大样本暴露了此前稀疏观测难以捕捉的架构规律：41.8%的验证候选采用非2的幂通道宽度，且性能最强的模型普遍呈现中早期通道适中、中后段扩展的结构性分配模式。这些发现表明，初始研究中的通道搜索信号在大规模下依然有效。
原文摘要 · Abstract (English)
Promising initial results in closed-loop large-language-model-based channel-configuration search demonstrated that neural-network widths can be optimized directly through executable code generation and accuracy feedback. However, those results were obtained from a relatively sparse set of valid evaluations, leaving open whether the observed optimization behavior transfers to a denser sampling regime and whether additional architectural regularities emerge when more generated networks are evaluated. To test this, the same search setting is scaled to 250 candidate networks per fine-tuning cycle. The analysis covers 2000 generated candidates from 8 complete cycles, yielding 462 verified CIFAR-100 evaluations after task and metadata filtering. Per-cycle mean accuracy exhibits a positive linear trend with slope 9.87e-4 (p=0.043), while the high-performing frontier improves more strongly: the best observed accuracy increases from 0.3144 to 0.3676, and both the top-5 and top-10 cycle-level means exhibit positive trends. The scaled run also reveals improved parameter efficiency. The best model reaches 0.3676 with 11.8M parameters, compared with an early high-performing model at 0.3144 with 166.5M parameters. Beyond accuracy, the larger sample exposes architectural regularities that were difficult to assess from sparse observations. Non-power-of-two channel widths occur in 41.8% of verified candidates, and the strongest models share structured channel-allocation patterns characterized by moderate early widths and expanded middle or later blocks. These findings indicate that the channel-search signal observed in the initial study transfers
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。