GLL: A Differentiable Graph Learning Layer for Neural Networks
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866908698743406592 |
|---|---|
| author | Brown, Jason Chen, Bohan Hardiman-Mostow, Harris Calder, Jeff Bertozzi, Andrea L. |
| author_facet | Brown, Jason Chen, Bohan Hardiman-Mostow, Harris Calder, Jeff Bertozzi, Andrea L. |
| contents | Standard deep learning architectures used for classification generate label predictions with a projection head and softmax activation function. Although successful, these methods fail to leverage the relational information between samples for generating label predictions. In recent works, graph-based learning techniques, namely Laplace learning, have been heuristically combined with neural networks for both supervised and semi-supervised learning (SSL) tasks. However, prior works approximate the gradient of the loss function with respect to the graph learning algorithm or decouple the processes; end-to-end integration with neural networks is not achieved. In this work, we derive backpropagation equations, via the adjoint method, for inclusion of a general family of graph learning layers into a neural network. The resulting method, distinct from graph neural networks, allows us to precisely integrate similarity graph construction and graph Laplacian-based label propagation into a neural network layer, replacing a projection head and softmax activation function for general classification task. Our experimental results demonstrate smooth label transitions across data, improved generalization and robustness to adversarial attacks, and improved training dynamics compared to a standard softmax-based approach. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2412_08016 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | GLL: A Differentiable Graph Learning Layer for Neural Networks Brown, Jason Chen, Bohan Hardiman-Mostow, Harris Calder, Jeff Bertozzi, Andrea L. Machine Learning 68T05, 68T07, 35R02 I.2.6; I.2.10; I.4.0 Standard deep learning architectures used for classification generate label predictions with a projection head and softmax activation function. Although successful, these methods fail to leverage the relational information between samples for generating label predictions. In recent works, graph-based learning techniques, namely Laplace learning, have been heuristically combined with neural networks for both supervised and semi-supervised learning (SSL) tasks. However, prior works approximate the gradient of the loss function with respect to the graph learning algorithm or decouple the processes; end-to-end integration with neural networks is not achieved. In this work, we derive backpropagation equations, via the adjoint method, for inclusion of a general family of graph learning layers into a neural network. The resulting method, distinct from graph neural networks, allows us to precisely integrate similarity graph construction and graph Laplacian-based label propagation into a neural network layer, replacing a projection head and softmax activation function for general classification task. Our experimental results demonstrate smooth label transitions across data, improved generalization and robustness to adversarial attacks, and improved training dynamics compared to a standard softmax-based approach. |
| title | GLL: A Differentiable Graph Learning Layer for Neural Networks |
| topic | Machine Learning 68T05, 68T07, 35R02 I.2.6; I.2.10; I.4.0 |
| url | https://arxiv.org/abs/2412.08016 |