Learning Along the Arrow of Time: Hyperbolic Geometry for Backward-Compatible Representation Learning
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866913881331335168 |
|---|---|
| author | Bui, Ngoc Yang, Menglin Chen, Runjin Neves, Leonardo Ju, Mingxuan Ying, Rex Shah, Neil Zhao, Tong |
| author_facet | Bui, Ngoc Yang, Menglin Chen, Runjin Neves, Leonardo Ju, Mingxuan Ying, Rex Shah, Neil Zhao, Tong |
| contents | Backward compatible representation learning enables updated models to integrate seamlessly with existing ones, avoiding to reprocess stored data. Despite recent advances, existing compatibility approaches in Euclidean space neglect the uncertainty in the old embedding model and force the new model to reconstruct outdated representations regardless of their quality, thereby hindering the learning process of the new model. In this paper, we propose to switch perspectives to hyperbolic geometry, where we treat time as a natural axis for capturing a model's confidence and evolution. By lifting embeddings into hyperbolic space and constraining updated embeddings to lie within the entailment cone of the old ones, we maintain generational consistency across models while accounting for uncertainties in the representations. To further enhance compatibility, we introduce a robust contrastive alignment loss that dynamically adjusts alignment weights based on the uncertainty of the old embeddings. Experiments validate the superiority of the proposed method in achieving compatibility, paving the way for more resilient and adaptable machine learning systems. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_05826 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Learning Along the Arrow of Time: Hyperbolic Geometry for Backward-Compatible Representation Learning Bui, Ngoc Yang, Menglin Chen, Runjin Neves, Leonardo Ju, Mingxuan Ying, Rex Shah, Neil Zhao, Tong Machine Learning Backward compatible representation learning enables updated models to integrate seamlessly with existing ones, avoiding to reprocess stored data. Despite recent advances, existing compatibility approaches in Euclidean space neglect the uncertainty in the old embedding model and force the new model to reconstruct outdated representations regardless of their quality, thereby hindering the learning process of the new model. In this paper, we propose to switch perspectives to hyperbolic geometry, where we treat time as a natural axis for capturing a model's confidence and evolution. By lifting embeddings into hyperbolic space and constraining updated embeddings to lie within the entailment cone of the old ones, we maintain generational consistency across models while accounting for uncertainties in the representations. To further enhance compatibility, we introduce a robust contrastive alignment loss that dynamically adjusts alignment weights based on the uncertainty of the old embeddings. Experiments validate the superiority of the proposed method in achieving compatibility, paving the way for more resilient and adaptable machine learning systems. |
| title | Learning Along the Arrow of Time: Hyperbolic Geometry for Backward-Compatible Representation Learning |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2506.05826 |