Improved Generalization of Weight Space Networks via Augmentations
This work addresses overfitting issues in weight space models, which is an incremental improvement for applications like neural fields and neural network inference.
The paper tackled the problem of overfitting in deep weight space networks by addressing the lack of diversity in datasets, and the result was that their proposed MixUp method improved classification performance similarly to having up to 10 times more data and yielded 5-10% gains in downstream classification.
Learning in deep weight spaces (DWS), where neural networks process the weights of other neural networks, is an emerging research direction, with applications to 2D and 3D neural fields (INRs, NeRFs), as well as making inferences about other types of neural networks. Unfortunately, weight space models tend to suffer from substantial overfitting. We empirically analyze the reasons for this overfitting and find that a key reason is the lack of diversity in DWS datasets. While a given object can be represented by many different weight configurations, typical INR training sets fail to capture variability across INRs that represent the same object. To address this, we explore strategies for data augmentation in weight spaces and propose a MixUp method adapted for weight spaces. We demonstrate the effectiveness of these methods in two setups. In classification, they improve performance similarly to having up to 10 times more data. In self-supervised contrastive learning, they yield substantial 5-10% gains in downstream classification.