SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2310.12487 · 2023

Improved Operator Learning by Orthogonal Attention

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 8 functions out of this paper's own repositories and ran 8 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
zhijie-group/orthogonal-neural-operator canonical 8 of 8
FunctionStatusWhere it lives
apply_2d_rotary_pos_emb Ran zhijie-group/orthogonal-neural-operator/ONOmodel2.py
code served (permissive licence) · get_code("6f4322bdb0cd6491")
apply_rotary_pos_emb Ran zhijie-group/orthogonal-neural-operator/ONOmodel2.py
code served (permissive licence) · get_code("b9007c9564c0612a")
central_diff Ran zhijie-group/orthogonal-neural-operator/Darcy_example.py
code served (permissive licence) · get_code("fbbefdd8c6f1550c")
central_diff Ran zhijie-group/orthogonal-neural-operator/time_gen.py
code served (permissive licence) · get_code("44cb0b9d7d877495")
count_parameters Ran zhijie-group/orthogonal-neural-operator/Darcy_example.py
code served (permissive licence) · get_code("54097d41b039f37b")
random_collate_fn Ran zhijie-group/orthogonal-neural-operator/NS_example2.py
code served (permissive licence) · get_code("7a755f3c02e8f45f")
random_collate_fn Ran zhijie-group/orthogonal-neural-operator/pla_example.py
code served (permissive licence) · get_code("532198a3d341e4b3")
rotate_half Ran zhijie-group/orthogonal-neural-operator/ONOmodel2.py
code served (permissive licence) · get_code("8549b7478752bf18")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Neural operators, as an efficient surrogate model for learning the solutions of PDEs, have received extensive attention in the field of scientific machine learning. Among them, attention-based neural operators have become one of the mainstreams in related research. However, existing approaches overfit the limited training data due to the considerable number of parameters in the attention mechanism. To address this, we develop an orthogonal attention based on the eigendecomposition of the kernel integral operator and the neural approximation of eigenfunctions. The orthogonalization naturally poses a proper regularization effect on the resulting neural operator, which aids in resisting overfitting and boosting generalization. Experiments on six standard neural operator benchmark datasets comprising both regular and irregular geometries show that our method can outperform competing baselines with decent margins.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2310.12487")
get_code_for_paper("2310.12487")
have("2310.12487")

Connect an agent — have() is free.