Scalable Meta-Learning via Mixed-Mode Differentiation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912421983027200 |
|---|---|
| author | Kemaev, Iurii Calian, Dan A Zintgraf, Luisa M Farquhar, Gregory van Hasselt, Hado |
| author_facet | Kemaev, Iurii Calian, Dan A Zintgraf, Luisa M Farquhar, Gregory van Hasselt, Hado |
| contents | Gradient-based bilevel optimisation is a powerful technique with applications in hyperparameter optimisation, task adaptation, algorithm discovery, meta-learning more broadly, and beyond. It often requires differentiating through the gradient-based optimisation itself, leading to "gradient-of-a-gradient" calculations with computationally expensive second-order and mixed derivatives. While modern automatic differentiation libraries provide a convenient way to write programs for calculating these derivatives, they oftentimes cannot fully exploit the specific structure of these problems out-of-the-box, leading to suboptimal performance. In this paper, we analyse such cases and propose Mixed-Flow Meta-Gradients, or MixFlow-MG -- a practical algorithm that uses mixed-mode differentiation to construct more efficient and scalable computational graphs yielding over 10x memory and up to 25% wall-clock time improvements over standard implementations in modern meta-learning setups. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_00793 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Scalable Meta-Learning via Mixed-Mode Differentiation Kemaev, Iurii Calian, Dan A Zintgraf, Luisa M Farquhar, Gregory van Hasselt, Hado Machine Learning Artificial Intelligence Gradient-based bilevel optimisation is a powerful technique with applications in hyperparameter optimisation, task adaptation, algorithm discovery, meta-learning more broadly, and beyond. It often requires differentiating through the gradient-based optimisation itself, leading to "gradient-of-a-gradient" calculations with computationally expensive second-order and mixed derivatives. While modern automatic differentiation libraries provide a convenient way to write programs for calculating these derivatives, they oftentimes cannot fully exploit the specific structure of these problems out-of-the-box, leading to suboptimal performance. In this paper, we analyse such cases and propose Mixed-Flow Meta-Gradients, or MixFlow-MG -- a practical algorithm that uses mixed-mode differentiation to construct more efficient and scalable computational graphs yielding over 10x memory and up to 25% wall-clock time improvements over standard implementations in modern meta-learning setups. |
| title | Scalable Meta-Learning via Mixed-Mode Differentiation |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2505.00793 |