Typed states for the displayed outbound observations.
Source: paper_references, paper_reference_links, observed 2026-08-08T11:46:47.840408Z
Paper Citation Record · LEDGER
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension
As of 16 August 2026, this Paper Citation Record lists 36 of 36 outbound references and 3 inbound Pith citation observations for arXiv:2502.07752.
A citation records a reference. It does not transfer a finding from one paper to another.
Coverage vector
One-hop event checks from named stored sources.
Source: scholarly_work_events, retraction_status_cache, observed 2026-08-16T06:30:59.297886+00:00
Pith citing papers itemized under the disclosed page cap.
Source: paper_references, paper_reference_links, observed 2026-08-02T08:27:40.373666Z
A source-named dated measurement, never combined with another source.
Source: pith, observed 2026-07-08T22:25:39.545387Z
Reference resolution
36 of 36 outbound references displayed
- verified exact4
- verified fuzzy10
- unresolved22
- parse uncertain0
- malformed identifier0
- metadata mismatch0
External citation measurements
No source-named external measurement is stored.
Outbound references
Observation 5bfff0f7-79e4-4529-936b-889f94da49b4 · outbound
This paper cites Scalable Second Order Optimization for Deep Learning.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension Scalable Second Order Optimization for Deep Learning
Reference 1
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.441001Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 7c25bd3b-2481-4d21-8029-2ef4fa51b299 · outbound
This paper cites M1 0 0 0 M2 0 0 0 M3 # Example of Diagv(·) This will stack the vector element into a pure diagonal matrix: Diagv([a11, a22, a33]T ) =.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension M1 0 0 0 M2 0 0 0 M3 # Example of Diagv(·) This will stack the vector element into a pure diagonal matrix: Diagv([a11, a22, a33]T ) =
Reference 2
- Resolution
- verified fuzzy
- raw_fallback, observed 2026-08-08T11:46:48.475556Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-16T06:30:59.297886+00:00.
Observation d0ed36c7-5554-4911-9910-5b40f9a75dab · outbound
This paper cites From the Theorem D.1, the iterative procedure for Q can be simply obtained by taking the diagonals of M: Q = Diag E GSGT ∥S∥2 F.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension From the Theorem D.1, the iterative procedure for Q can be simply obtained by taking the diagonals of M: Q = Diag E GSGT ∥S∥2 F
Reference 3
- Resolution
- verified fuzzy
- raw_fallback, observed 2026-08-08T11:46:48.401760Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-16T06:30:59.297886+00:00.
Observation 41d59989-3507-4b07-a1f4-167591649010 · outbound
This paper cites Effects of last layer One crucial setup difference during evaluation for low-rank methods is whether the last layer is trained by full-rank Adam or not.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension Effects of last layer One crucial setup difference during evaluation for low-rank methods is whether the last layer is trained by full-rank Adam or not
Reference 4
- Resolution
- verified fuzzy
- raw_fallback, observed 2026-08-08T11:46:48.351953Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-16T06:30:59.297886+00:00.
Observation 6d6d2a75-85dc-41f1-b244-cfc84dcdd9f9 · outbound
This paper cites LoRA: Low-Rank Adaptation of Large Language Models.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension LoRA: Low-Rank Adaptation of Large Language Models
Reference 5
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.457800Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 3f0ec22f-5055-46d2-a09d-78f7a9910efd · outbound
This paper cites FAdam: Adam is a natural gradient optimizer using diagonal empirical Fisher information.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension FAdam: Adam is a natural gradient optimizer using diagonal empirical Fisher information
Reference 6
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.479823Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 27822886-f0b7-4bb6-b64a-620ff692c6cc · outbound
This paper cites For 8-bit optimizers, we assume weights are stored in BF16, but optimizer states use FP8.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension For 8-bit optimizers, we assume weights are stored in BF16, but optimizer states use FP8
Reference 9
- Resolution
- verified exact
- raw_fallback, observed 2026-08-08T11:46:47.934780Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-16T06:30:59.297886+00:00.
Observation 1847dc01-59e5-4b35-9f69-8e0a19be6459 · outbound
This paper cites an unresolved cited work.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension Unresolved cited work
Reference 10
- Resolution
- unresolved
- raw_fallback, observed 2026-08-08T11:46:48.364266Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-16T06:30:59.297886+00:00.
Observation af1dfbbe-8afb-4c20-93c2-cca4ee01e73c · outbound
This paper cites SGDR: Stochastic Gradient Descent with Warm Restarts.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension SGDR: Stochastic Gradient Descent with Warm Restarts
Reference 11
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.747483Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 34d89f11-b35a-49a3-a208-31cb801b1f64 · outbound
This paper cites A New Perspective on Shampoo's Preconditioner.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension A New Perspective on Shampoo's Preconditioner
Reference 13
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.755604Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 07dcae0d-99ef-421d-9a27-abaf2e1bb770 · outbound
This paper cites Curvature-Informed SGD via General Purpose Lie-Group Preconditioners.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension Curvature-Informed SGD via General Purpose Lie-Group Preconditioners
Reference 14
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.759576Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 1ee16ce6-d40b-4be5-b189-fa3c6654258f · outbound
This paper cites Maintaining Structural Integrity in Parameter Spaces for Parameter Efficient Fine-tuning.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension Maintaining Structural Integrity in Parameter Spaces for Parameter Efficient Fine-tuning
Reference 15
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.763401Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 87834fae-442a-46ee-884f-0294812c3003 · outbound
This paper cites Connection of diagonal hessian estimates to natural gradients in stochastic optimization.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension Connection of diagonal hessian estimates to natural gradients in stochastic optimization
Reference 16
- Resolution
- verified fuzzy
- raw_fallback, observed 2026-08-08T11:46:48.486121Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-16T06:30:59.297886+00:00.
Observation b3c55a12-7f0f-436f-bc2d-a67a591c09dd · outbound
This paper cites Understanding Self-supervised Learning with Dual Deep Networks.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension Understanding Self-supervised Learning with Dual Deep Networks
Reference 17
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.771026Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 6bd0fe42-d56c-4794-83c5-42f239b0fc19 · outbound
This paper cites SOAP: Improving and Stabilizing Shampoo using Adam.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension SOAP: Improving and Stabilizing Shampoo using Adam
Reference 19
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.778814Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 938f875f-cad8-42ce-916b-ecc6ec5e8f30 · outbound
This paper cites No More Adam: Learning Rate Scaling at Initialization is All You Need.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension No More Adam: Learning Rate Scaling at Initialization is All You Need
Reference 20
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.782800Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 3b7c6b6a-edde-44c1-ba10-87bc84748d90 · outbound
This paper cites Large Batch Optimization for Deep Learning: Training BERT in 76 minutes.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension Large Batch Optimization for Deep Learning: Training BERT in 76 minutes
Reference 22
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.790437Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 87ec5990-80e3-4c54-abd5-0c85ea4bf764 · outbound
This paper cites Adam-mini: Use Fewer Learning Rates To Gain More.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension Adam-mini: Use Fewer Learning Rates To Gain More
Reference 23
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.795211Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 41446892-5841-44ff-806f-c9e190c9a2f9 · outbound
This paper cites GaLore: Memory-Efficient LLM Training by Gradient Low-Rank Projection.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension GaLore: Memory-Efficient LLM Training by Gradient Low-Rank Projection
Reference 24
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.799002Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 681d9b23-720e-479d-9aa9-2a74eff61afc · outbound
This paper cites for t = 1,.
Reference 26
- Resolution
- verified fuzzy
- raw_fallback, observed 2026-08-08T11:46:48.464545Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-16T06:30:59.297886+00:00.
Observation 9b1a17ee-3f19-456a-a908-df93999b7bc1 · outbound
This paper cites Note that our paper assumes Vec(·) is stacking columns of matrix whereas Gupta et al.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension Note that our paper assumes Vec(·) is stacking columns of matrix whereas Gupta et al
Reference 27
- Resolution
- verified fuzzy
- raw_fallback, observed 2026-08-08T11:46:48.452711Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-16T06:30:59.297886+00:00.
Observation 2da2c5cc-9156-4851-b16c-d659a268811b · outbound
This paper cites an unresolved cited work.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension Unresolved cited work
Reference 28
- Resolution
- unresolved
- raw_fallback, observed 2026-08-08T11:46:48.439953Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-16T06:30:59.297886+00:00.
Observation 49c3e7fc-8115-4ce3-b3b4-7f35fb270d6c · outbound
This paper cites In practice, SW AN proposes to compute the(GGT )− 1 2 using Newton-Schulz iterations.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension In practice, SW AN proposes to compute the(GGT )− 1 2 using Newton-Schulz iterations
Reference 29
- Resolution
- verified fuzzy
- raw_fallback, observed 2026-08-08T11:46:48.427956Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-16T06:30:59.297886+00:00.
Observation 2e1d9700-dd5f-4773-8de0-7500648dca74 · outbound
This paper cites At the same time, GaLore [Zhao et al., 2024a] popularizes the use of low-rank optimizers, which demonstrates on-par performance compared to full-rank Adam training.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension At the same time, GaLore [Zhao et al., 2024a] popularizes the use of low-rank optimizers, which demonstrates on-par performance compared to full-rank Adam training
Reference 30
- Resolution
- verified fuzzy
- raw_fallback, observed 2026-08-08T11:46:48.415008Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-16T06:30:59.297886+00:00.
Observation d3e11205-1153-40ae-b572-72d1e534efd7 · outbound
This paper cites Then, the normalization step of Lamb and Lars can be viewed as a 1-sample approximation to FIM* under the structure considered in Sec.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension Then, the normalization step of Lamb and Lars can be viewed as a 1-sample approximation to FIM* under the structure considered in Sec
Reference 32
- Resolution
- verified fuzzy
- raw_fallback, observed 2026-08-08T11:46:48.389263Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-16T06:30:59.297886+00:00.
Observation 262c1a26-ccc7-45a0-80ed-caa8d5b905e6 · outbound
This paper cites U T U T c G ⊙2# =[U , Uc] U T G U T c G vuutE.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension U T U T c G ⊙2# =[U , Uc] U T G U T c G vuutE
Reference 33
- Resolution
- verified fuzzy
- raw_fallback, observed 2026-08-08T11:46:48.376523Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-16T06:30:59.297886+00:00.
Observation 26634398-10a7-4e11-8a13-13b53a4aaa42 · outbound
This paper cites Large Batch Training of Convolutional Networks.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension Large Batch Training of Convolutional Networks
Reference 2008
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.786590Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 8d0ec684-065f-49f1-905a-3d8e8d99d594 · outbound
This paper cites Efficient Approximations of the Fisher Matrix in Neural Networks using Kronecker Product Singular Value Decomposition.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension Efficient Approximations of the Fisher Matrix in Neural Networks using Kronecker Product Singular Value Decomposition
Reference 2014
- Resolution
- verified exact
- local_arxiv, observed 2026-08-08T11:46:48.234529Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-16T06:30:59.297886+00:00.
Observation 860c6b52-a841-4eaf-9efe-f4947bd814be · outbound
This paper cites SWAN: SGD with Normalization and Whitening Enables Stateless LLM Training.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension SWAN: SGD with Normalization and Whitening Enables Stateless LLM Training
Reference 2016
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.751693Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 91c9ad27-332d-4c82-812c-716d28f52cc3 · outbound
This paper cites Preconditioner on Matrix Lie Group for SGD.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension Preconditioner on Matrix Lie Group for SGD
Reference 2017
- Resolution
- verified exact
- local_arxiv, observed 2026-08-08T11:46:48.218281Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-16T06:30:59.297886+00:00.
Observation df762821-163a-4888-8ce3-1a2c866041c0 · outbound
This paper cites Fira: Can we achieve full-rank training of llms under low-rank constraint? arXiv preprint arXiv:2410.01623, 2024a.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension Fira: Can we achieve full-rank training of llms under low-rank constraint? arXiv preprint arXiv:2410.01623, 2024a
Reference 2018
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.446203Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation f55c5c32-f0e7-4f6e-9262-f79e92191c9b · outbound
This paper cites On Empirical Comparisons of Optimizers for Deep Learning.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension On Empirical Comparisons of Optimizers for Deep Learning
Reference 2019
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.449567Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 1cb85550-a248-4f7a-b11c-f9226b101e92 · outbound
This paper cites LLaMA: Open and Efficient Foundation Language Models.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension LLaMA: Open and Efficient Foundation Language Models
Reference 2020
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.775109Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation d031e4fd-ea87-48a9-a5c7-62fcc633626b · outbound
This paper cites The Llama 3 Herd of Models.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension The Llama 3 Herd of Models
Reference 2021
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.453851Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 98d71768-a389-41a1-9b93-9794e8b2545b · outbound
This paper cites Can We Remove the Square-Root in Adaptive Gradient Methods? A Second-Order Perspective.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension Can We Remove the Square-Root in Adaptive Gradient Methods? A Second-Order Perspective
Reference 2023
- Resolution
- verified exact
- local_arxiv, observed 2026-08-08T11:46:48.200943Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-16T06:30:59.297886+00:00.
Observation dc51d578-f541-4b7b-8309-3d1ace46c9da · outbound
This paper cites Adam: A Method for Stochastic Optimization.
Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension Adam: A Method for Stochastic Optimization
Reference 2024
- Resolution
- unresolved
- no resolver link, observed 2026-08-08T11:46:47.554330Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Pith citing papers
Observation 91f4ecc9-1ff4-4058-9bf0-0ff00be0822d · inbound
OmniOpt: Taxonomy, Geometry, and Benchmarking of Modern Optimizers cites this paper.
OmniOpt: Taxonomy, Geometry, and Benchmarking of Modern Optimizers Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension
Reference 37
- Resolution
- unresolved
- no resolver link, observed 2026-07-11T22:10:49.683444Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 88a4455d-ebb9-4db0-a1a4-4857364a2afe · inbound
No Subspace to Track: Non-Identifiability and Optimizer State in Low-Rank Training cites this paper.
No Subspace to Track: Non-Identifiability and Optimizer State in Low-Rank Training Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension
Reference 5
- Resolution
- metadata mismatch
- local_arxiv, observed 2026-07-08T22:25:39.546940Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-16T06:30:59.297886+00:00.
Observation cf227b65-8840-4ae3-a742-e515ac419a16 · inbound
No Subspace to Track: Non-Identifiability and Optimizer State in Low-Rank Training cites this paper.
No Subspace to Track: Non-Identifiability and Optimizer State in Low-Rank Training Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension
Reference 2025
- Resolution
- unresolved
- no resolver link, observed 2026-08-02T08:27:40.373666Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.