SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2402.03655 · ICML · 2024

Operator SVD with Neural Networks via Nested Low-Rank Approximation

Yuheng Bu, Gregory Wornell, J Ryu, Xiangxiang Xu, H Melihcan Erol, Lizhong Zheng

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 7 functions out of this paper's own repositories and ran 6 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
jongharyu/neural-svd canonical 6 of 7
FunctionStatusWhere it lives
NestedLoRA Ran jongharyu/neural-svd/methods/nestedlora.py
code served (permissive licence) · get_code("595283443161fcb7")
NestedLoRALossFunctionEVD Ran jongharyu/neural-svd/methods/nestedlora.py
code served (permissive licence) · get_code("1d596dbaf553ca99")
NestedLoRALossFunctionSVD Ran jongharyu/neural-svd/methods/nestedlora.py
code served (permissive licence) · get_code("d527b1dbdd9f9b76")
compute_lambda Ran jongharyu/neural-svd/methods/nestedlora.py
code served (permissive licence) · get_code("f0d2fe3cdb042e8b")
get_joint_nesting_masks Ran jongharyu/neural-svd/methods/nestedlora.py
code served (permissive licence) · get_code("0003166a87311c4a")
get_sequential_nesting_masks Ran jongharyu/neural-svd/methods/nestedlora.py
code served (permissive licence) · get_code("34382e86673d0ad9")
compute_loss_metric Not yet run jongharyu/neural-svd/methods/nestedlora.py
code served (permissive licence) · get_code("5497e0c532fa215a")

Repositories linked to this paper

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

Abstract

Computing eigenvalue decomposition (EVD) of a given linear operator, or finding its leading eigenvalues and eigenfunctions, is a fundamental task in many machine learning and scientific computing problems. For high-dimensional eigenvalue problems, training neural networks to parameterize the eigenfunctions is considered as a promising alternative to the classical numerical linear algebra techniques. This paper proposes a new optimization framework based on the low-rank approximation characterization of a truncated singular value decomposition, accompanied by new techniques called nesting for learning the top-L singular values and singular functions in the correct order. The proposed method promotes the desired orthogonality in the learned functions implicitly and efficiently via an unconstrained optimization formulation, which is easy to solve with off-the-shelf gradient-based optimization algorithms. We demonstrate the effectiveness of the proposed optimization framework for use cases in computational physics and machine learning.

For agents

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

get_harvested_code_for_paper("2402.03655")
get_code_for_paper("2402.03655")
have("2402.03655")

Connect an agent — have() is free.