发表机构
Institute of Fundamental Technological Research, Polish Academy of Sciences; Samsung AI Center（波兰科学院基础技术研究所; 三星人工智能中心）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
提出分层专家混合Transformer结合元学习框架，在权重空间实现隐式神经表示分类，在多分辨率基准达最优准确率，还揭示了INRs编码判别信息的结构特性。
AI 中文摘要
隐式神经表示（INRs）将信号编码为基于坐标的神经网络的权重，近期被提出作为下游学习的替代领域。尽管前景可观，但由于INRs参数的高维性和复杂结构，直接在权重空间进行分类仍具挑战性，且判别信息在INRs权重中的分布方式也尚未被充分理解。我们提出一种分层专家混合（HMoE）Transformer，其采用与底层隐式网络结构对齐的条件计算来处理INRs权重。结合用于塑造INRs参数以适配下游任务的元学习框架，我们的模型在标准基准测试中取得了最先进的准确率，覆盖从低分辨率数据集到高分辨率ImageNet-1K的范围。为深入理解INRs如何编码判别信息，我们开发了权重空间归因与剪枝方法，以识别与分类最相关的参数。这些分析揭示了INRs层内类别特定结构的形成方式，并证明了MoE架构适用于权重空间学习。我们的方法提升了权重空间分类器的性能与可解释性。
英文摘要
Implicit Neural Representations (INRs) encode signals as the weights of a coordinate-based neural network and have recently been proposed as an alternative domain for downstream learning. While promising, classification directly in weight space remains challenging due to the high dimensionality and complex structure of INR parameters. Furthermore, the way discriminative information is distributed across INR weights remains poorly understood. We propose a hierarchical Mixture-of-Experts (HMoE) Transformer that processes INR weights using conditional computation aligned with the structure of the underlying implicit network. Coupled with a meta-learning framework that shapes INR parameters for downstream tasks, our model achieves state-of-the-art accuracy across standard benchmarks, ranging from low-resolution datasets to high-resolution ImageNet-1K. To gain insight into how INRs encode discriminative information, we develop weight-space attribution and pruning methods that identify parameters most relevant for classification. These analyses reveal how class-specific structure emerges within INR layers and support the suitability of MoE architectures for weight-space learning. Our approach advances both the performance and interpretability of weight-space classifiers.
CommentsECCV 2026, 22 pages
