Landscaper: Understanding Loss Landscapes Through Multi-Dimensional Topological Analysis
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866912916931870720 |
|---|---|
| author | Chen, Jiaqing Hadler, Nicholas Xie, Tiankai Hnatyshyn, Rostyslav Geniesse, Caleb Yang, Yaoqing Mahoney, Michael W. Perciano, Talita Hartwig, John F. Maciejewski, Ross Weber, Gunther H. |
| author_facet | Chen, Jiaqing Hadler, Nicholas Xie, Tiankai Hnatyshyn, Rostyslav Geniesse, Caleb Yang, Yaoqing Mahoney, Michael W. Perciano, Talita Hartwig, John F. Maciejewski, Ross Weber, Gunther H. |
| contents | Loss landscapes are a powerful tool for understanding neural network optimization and generalization, yet traditional low-dimensional analyses often miss complex topological features. We present Landscaper, an open-source Python package for arbitrary-dimensional loss landscape analysis. Landscaper combines Hessian-based subspace construction with topological data analysis to reveal geometric structures such as basin hierarchy and connectivity. A key component is the Saddle-Minimum Average Distance (SMAD) for quantifying landscape smoothness. We demonstrate Landscaper's effectiveness across various architectures and tasks, including those involving pre-trained language models, showing that SMAD captures training transitions, such as landscape simplification, that conventional metrics miss. We also illustrate Landscaper's performance in challenging chemical property prediction tasks, where SMAD can serve as a metric for out-of-distribution generalization, offering valuable insights for model diagnostics and architecture design in data-scarce scientific machine learning scenarios. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_07135 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Landscaper: Understanding Loss Landscapes Through Multi-Dimensional Topological Analysis Chen, Jiaqing Hadler, Nicholas Xie, Tiankai Hnatyshyn, Rostyslav Geniesse, Caleb Yang, Yaoqing Mahoney, Michael W. Perciano, Talita Hartwig, John F. Maciejewski, Ross Weber, Gunther H. Machine Learning Artificial Intelligence Loss landscapes are a powerful tool for understanding neural network optimization and generalization, yet traditional low-dimensional analyses often miss complex topological features. We present Landscaper, an open-source Python package for arbitrary-dimensional loss landscape analysis. Landscaper combines Hessian-based subspace construction with topological data analysis to reveal geometric structures such as basin hierarchy and connectivity. A key component is the Saddle-Minimum Average Distance (SMAD) for quantifying landscape smoothness. We demonstrate Landscaper's effectiveness across various architectures and tasks, including those involving pre-trained language models, showing that SMAD captures training transitions, such as landscape simplification, that conventional metrics miss. We also illustrate Landscaper's performance in challenging chemical property prediction tasks, where SMAD can serve as a metric for out-of-distribution generalization, offering valuable insights for model diagnostics and architecture design in data-scarce scientific machine learning scenarios. |
| title | Landscaper: Understanding Loss Landscapes Through Multi-Dimensional Topological Analysis |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2602.07135 |