File size: 886 Bytes
e0f62a9
 
 
 
 
a61880e
 
0a8284b
a61880e
 
0aec032
a61880e
0aec032
a61880e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
---
license: mit
tags:
  - neural-operator
  - attention
  - operator-learning
arxiv: 2605.31559
---
# FuncAttn: Functional Attention

This repository hosts the trained model weights for the PDE benchmarks reported in the paper [Functional Attention: From Pairwise Affinities to Functional Correspondences](https://huggingface.co/papers/2605.31559) (ICML 2026).

- **Project page:** https://xjffff.github.io/funcattn/
- **Code:** https://github.com/xjffff/funcattn

## Citation

```bibtex
@misc{xiao2026functionalattentionpairwiseaffinities,
      title={Functional Attention: From Pairwise Affinities to Functional Correspondences}, 
      author={Jiefang Xiao and Maolin Gao and Simon Weber and Guandao Yang and Daniel Cremers},
      year={2026},
      eprint={2605.31559},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2605.31559}, 
}
```