面向视觉的预算自适应评估轻量神经架构搜索
Vision-Oriented Lightweight Neural Architecture Search with Budget-Adaptive Evaluation
用六个训练开销极小的微型视觉任务作为代理信号来给候选架构排序，可在接近免训练方法的成本下获得基于训练的可靠性，适合直接替换现有 NAS 流程中的评估器。
原文依据 we propose a vision-oriented lightweight training-based NAS framework
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论在基于深度学习的计算机视觉社区中，神经架构搜索（NAS）已成为获取任务最优网络结构的事实标准工具。然而，NAS 方法陷入了精度与效率的根本性两难：基于训练的方法能给出可靠的性能，但搜索成本高得难以承受；免训练策略速度极快，却往往产生相对不可靠的排序。为调和这一冲突，我们提出一个面向视觉的轻量级基于训练的 NAS 框架。我们首先设计了六个微型视觉任务，其训练时间可忽略不计，但它们共同探测了广泛的表征能力谱系。在这些任务的基础上，我们引入一个预算自适应的性能评估器，以在给定预算限制内产生所能达到的最准确排序。在主流 NAS 基准上的实验表明，我们的方法取得了高于现有方法的排序相关性。此外，我们由流行的神经模块构建了一个搜索空间，并以接近免训练方法的成本运行我们的方法；在相同训练配方下，所发现的架构超越了当前最优。我们的代码可在 https://github.com/fanyi-plus/tf-nas 获取。
英文原文摘要 1239 字符
In the deep-learning-based computer vision community, Neural Architecture Search (NAS) has become the de-facto tool for acquiring task-optimal network structures. Nevertheless, NAS methods are trapped in a fundamental accuracy-efficiency dilemma: training-based approaches deliver reliable performance but incur prohibitive search costs, whereas training-free strategies are ultra-fast but often yield relatively unreliable rankings. To reconcile this conflict, we propose a vision-oriented lightweight training-based NAS framework. We first design six micro vision tasks whose training time is negligible, yet together they probe a broad spectrum of representational capacities. Built upon these tasks, we introduce a budget-adaptive performance evaluator to produce the most accurate ranking attainable within the limit. Experiments on popular NAS benchmarks show that our method achieves a ranking correlation higher than existing methods. Furthermore, we construct a search space from prevalent neural blocks and run our method at a cost close to training-free methods; the discovered architecture surpasses the current state-of-the-art under identical training recipes. Our codes are available at https://github.com/fanyi-plus/tf-nas.