通过数据集蒸馏发现星系特征
Discovering Galaxy Features via Dataset Distillation
AI总结：
本文提出利用数据集蒸馏技术从神经网络分类器中生成星系形态的合成原型图像，以逆向揭示神经网络用于星系分类的关键视觉特征，并引入自适应 Matching Trajectory 算法提升蒸馏性能。
摘要：
在许多应用中，神经网络（NNs）的分类性能已达到甚至超过人类水平。此外，神经网络很可能利用了与人类感知分类不同的底层特征。我们能否“逆向工程”相关特征以增强科学理解？在此，我们将这一思想应用于极具挑战性的星系分类任务：神经网络在该任务上已达到高性能，但神经网络在分类星系时“看到”了什么？是否存在人眼可能忽略、但有助于任务并提供新见解的形态特征？我们能否可视化早期演化的示踪物，或额外纳入的光谱数据？我们提出一种通过神经网络视角总结和可视化星系形态的新方法，利用数据集蒸馏（Dataset Distillation）——一种近期提出的深度学习方法，其主要目标是从大型数据集中蒸馏知识并将其浓缩为紧凑的合成数据集，使得在该合成数据集上训练的模型能够达到与在全量数据集上训练的模型相当的性能。我们构建了一个类别均衡、中等规模、高置信度的 Galaxy Zoo 2 数据集版本，并从我们精确的神经网络分类器出发进行数据集蒸馏，以生成星系形态特征的合成原型图像，证明了该方法的有效性。此外，作为独立贡献，我们引入了当前最先进的 Matching Trajectory 算法的自适应版本，以实现蒸馏过程的自动化，并在计算机视觉基准上展示了性能提升。
英文摘要：
In many applications, Neural Nets (NNs) have classification performance on par or even exceeding human capacity. Moreover, it is likely that NNs leverage underlying features that might differ from those humans perceive to classify. Can we "reverse-engineer" pertinent features to enhance our scientific understanding? Here, we apply this idea to the notoriously difficult task of galaxy classification: NNs have reached high performance for this task, but what does a neural net (NN) "see" when it classifies galaxies? Are there morphological features that the human eye might overlook that could help with the task and provide new insights? Can we visualize tracers of early evolution, or additionally incorporated spectral data? We present a novel way to summarize and visualize galaxy morphology through the lens of neural networks, leveraging Dataset Distillation, a recent deep-learning methodology with the primary objective to distill knowledge from a large dataset and condense it into a compact synthetic dataset, such that a model trained on this synthetic dataset achieves performance comparable to a model trained on the full dataset. We curate a class-balanced, medium-size high-confidence version of the Galaxy Zoo 2 dataset, and proceed with dataset distillation from our accurate NN-classifier to create synthesized prototypical images of galaxy morphological features, demonstrating its effectiveness. Of independent interest, we introduce a self-adaptive version of the state-of-the-art Matching Trajectory algorithm to automate the distillation process, and show enhanced performance on computer vision benchmarks.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
