← Back to all development cases
This case was judged by an anonymized AI reviewer and has not been independently reviewed by people; inspect the draft, cards, and both outputs to assess it.
AutocompleteWith cards better249 / 280 · beu-b57fcfa747777665

Rank-one Riemannian Subspace Descent for Nonlinear Matrix Equations

Mathematics · 2601.14933v1

FLOWING EVIDENCE BENCHMARK

How do we compare writing with and without cards?

At the same writing position, the same model produces one first result with cards and one without. All 280 pairs enter the overall anonymized review, which assesses whether a decision not to continue fits the draft.

Overall: 138 with-card wins, 26 ties, 107 no-card wins, 5 pairs where neither was usable, and 4 inconclusive judgments. Both sides wrote text in 218 pairs; 48 had a deliberate stop and 14 included an invalid output.

SAME DRAFT · WITH OR WITHOUT CARDS

01 · FIXED WRITING POSITION

MANUSCRIPT

Same writing position ▌

Matched autocomplete at the same draft position

02 · TWO MATCHED VERSIONS

SHARED BY BOTH

Draft, writing task, and generation model

A · With cards

Relevant literature cards supplied

B · Without cards

No literature cards supplied

Generation model

Same model and version

A → first output

B → first output

Anonymized AI judge

First outputs are anonymized as X and Y

Continuation review: accuracy, fit to the writing task and usability
Returns: X preferred / tie / Y preferred / both unusable

Review again with the answer order swapped

HOW DOES BLIND JUDGING WORK?

① Anonymize both outputs
The AI judge sees the same draft and both first outputs without knowing which received cards.

② Compare and swap order
The judge applies task-specific criteria in X/Y and then Y/X order.

③ Review disagreements
A third pass resolves disagreements. Incomplete reviews remain in the denominator.

How are positions and cards selected?

We selected 40 writing positions in each of seven fields, for 280 comparisons. The position and relevant literature cards were set before each comparison, so both versions began from the same draft.

The main result retains every frozen position, including places where cards might not add a useful next move. This avoids selecting cases based on their outputs, but 49% describes the whole cohort, not a win rate where evidence clearly had an opportunity to help. This cohort did not record that opportunity label for each position before generation, so those two groups cannot yet be separated reliably.

CARD USE · 140/280

Reported card use in continuations with text

The model reports whether a card affected its continuation. In 140 of 280 tests it produced continuation text and reported card use. The previous count of 150 included 10 cases without continuation text and has been corrected.

COMPARISON · 70 / 7 / 30

Which written version was better?

Of the 107 reported-use cases where both versions produced text, the blinded AI judge preferred the with-card version 70 times, found 7 about the same, and preferred the without-card version 30 times.

How are the two table percentages calculated?

ADOPTION COVERAGE

140reported use with text
÷
280all positions
=
50%

WITH-CARD VERSION PREFERRED

70with cards better
÷
107comparable tests
=
65%

The percentages use different groups: 50% counts continuations with text reporting card use among all 280 tests; 65% covers the 107 such cases with text on both sides. The generating model reports card use, while an anonymized AI judge compares writing quality.

Card use is self-reported by the generator. The anonymized AI judge sees the draft and source snippets, so source-aware preference remains possible; there was no independent human review. The overall result describes these 280 frozen positions, not a win probability for any writing task.

Manuscript writing position

Text excerpt · not a PDF page

Research manuscript · excerpt

Rank-one Riemannian Subspace Descent for Nonlinear Matrix Equations

Introduction / related work

22], [28]–[30], quasi-gradie nt- based inversion-free methods [22], [31], and accelerated a l- gorithms [32], [33]. Despite this breadth, the per-iterati on complexity of all these algorithms remains O(n3) or higher for the general case, making them unsuitable for large-scale pr ob- lems. Algorithms achieving lower per-iteration complexit ies typically rely on additional assumptions such as X being low rank and coefficient matrices being sparse [3], [11], [34]–[ 37]. However, without such low-rank and sparsity assumptions, t he per-iteration cost of these algorithms remains O(n3), limiting their applicability to high-dimensional problems [38]. Riemannian geometric methods have gained significant at- tention [39]–[43] for their ability to leverage the intrins ic geometry of SPD constraint. To reduce per-iteration costs, subspace/coordinate descent algorithms that update only a subset of directions [44]–[46] have recently been extended to manifold settings [47]–[50]. For the SPD manifold in particular, subspace descent algorithms were proposed in [ 43], [51]. Of these, [43] is the only method that reduces the dominant per-iteration cost on the SPD manifold to O(n2) flops for the general

↓ Next paragraph was generated here

The highlighted text is near the writing position. This is a layout of the evaluation input, not a PDF screenshot.

A matched comparison at the same writing positionOverall AI review: With cards better

Both arms use the same model, task and manuscript context. Only the retrieved passages differ. The first recorded output from each arm is shown below.

Read overall AI review rationale

A = without cards · B = with cards

Both outputs correctly complete the sentence comparing [43] and [51] on per-iteration cost. Output A states [51] retains O(n3) cost, which is plausible but creates a sharp contrast that seems slightly overstated—if [51] were truly O(n3), it would be unremarkable and not worth mentioning alongside [43]'s O(n2) breakthrough. Output B is more nuanced and better supported by the sources: source [2] explicitly describes 'uni-directional and multi-directional Riemannian subspace descent variants' with 'per-iteration complexities of O(n) and O(n2) respectively,' and notes this applies 'over a broad class of functions' where 'Riemannian gradients can be calculated efficiently.' Output B's qualification about 'functions whose Riemannian gradients admit efficient computation' accurately reflects this source-supported distinction, making it more precise and less likely to overstate [51]'s limitations. Output B also better preserves the paper's argumentative trajectory: the draft emphasizes that algorithms achieving lower complexity 'typically rely on additional assumptions,' and Output B's qualified claim about [51] maintains this thematic consistency about conditional improvements rather than reverting to a blanket O(n3) claim.

Evidence

Same model and position · retrieved passages supplied · first output

Preferred by blind review

case, whereas [51] attains O(n) per-iteration complexity only for functions whose Riemannian gradients admit efficient computation.

Passages supplied to the Evidence version

Card 1 · Riemannian coordinate descent algorithms on matrix manifolds

Many machine learning applications are naturally formulated as optimization problems on Riemannian manifolds. The main idea behind Riemannian optimization is to maintain the feasibility of the variables while moving along a descent direction on the manifold. This results in updating all the variable…

Read full passage excerpt

Many machine learning applications are naturally formulated as optimization problems on Riemannian manifolds. The main idea behind Riemannian optimization is to maintain the feasibility of the variables while moving along a descent direction on the manifold. This results in updating all the variables at every iteration. In this work, we provide a general framework for developing computationally efficient coordinate descent (CD) algorithms on matrix manifolds that allows updating only a few variables at every iteration while adhering to the manifold constraint. In particular, we propose CD algorithms for various manifolds such as Stiefel, Grassmann, (generalized) hyperbolic, symplectic, and symmetric positive (semi)definite. While the cost per iteration of the proposed CD algorithms is low, we further develop a more efficient variant via a first-order approximation of the objective function. We analyze their convergence and complexity, and empirically illustrate their efficacy in severa …

Card 2 · Low-complexity subspace-descent over symmetric positive definite manifold

This work puts forth low-complexity Riemannian subspace descent algorithms for the minimization of functions over the symmetric positive definite (SPD) manifold. Different from the existing Riemannian gradient descent variants, the proposed approach utilizes carefully chosen subspaces that allow the…

Read full passage excerpt

This work puts forth low-complexity Riemannian subspace descent algorithms for the minimization of functions over the symmetric positive definite (SPD) manifold. Different from the existing Riemannian gradient descent variants, the proposed approach utilizes carefully chosen subspaces that allow the update to be written as a product of the Cholesky factor of the iterate and a sparse matrix. The resulting updates avoid the costly matrix operations like matrix exponentiation and dense matrix multiplication, which are generally required in almost all other Riemannian optimization algorithms on SPD manifold. We further identify a broad class of functions, arising in diverse applications, such as kernel matrix learning, covariance estimation of Gaussian distributions, maximum likelihood parameter estimation of elliptically contoured distributions, and parameter estimation in Gaussian mixture model problems, over which the Riemannian gradients can be calculated efficiently. The proposed uni- …

Card 3 · Riemannian Tensor Completion with Side Information

By restricting the iterate on a nonlinear manifold, the recently proposed Riemannian optimization methods prove to be both efficient and effective in low rank tensor completion problems. However, existing methods fail to exploit the easily accessible side information, due to their format mismatch. C…

Read full passage excerpt

By restricting the iterate on a nonlinear manifold, the recently proposed Riemannian optimization methods prove to be both efficient and effective in low rank tensor completion problems. However, existing methods fail to exploit the easily accessible side information, due to their format mismatch. Consequently, there is still room for improvement in such methods. To fill the gap, in this paper, a novel Riemannian model is proposed to organically integrate the original model and the side information by overcoming their inconsistency. For this particular model, an efficient Riemannian conjugate gradient descent solver is devised based on a new metric that captures the curvature of the objective.Numerical experiments suggest that our solver is more accurate than the state-of-the-art without compromising the efficiency.

Card 4 · Variable projection framework for the reduced-rank matrix approximation problem by weighted least-squares

In this monograph, we review and develop variable projection Gauss-Newton, Levenberg-Marquardt and Newton methods for the Weighted Low-Rank Approximation (WLRA) problem, which has now an increasing number of applications in many scientific fields. Particular attention is drawn at the robustness, eff…

Read full passage excerpt

In this monograph, we review and develop variable projection Gauss-Newton, Levenberg-Marquardt and Newton methods for the Weighted Low-Rank Approximation (WLRA) problem, which has now an increasing number of applications in many scientific fields. Particular attention is drawn at the robustness, efficiency and scalability of these variable projection second-order algorithms such that they can be used also on larger datasets now commonly found in many practical problems for which only first-order algorithms based on sequential repetitions of local optimization (e.g., majorization, Expectation-Maximization or alternating least-squares methods) or variations of gradient descent (e.g., conjugate, proximal or stochastic gradient descent methods), or hybrid algorithms from these two classes of methods, were only feasible due to their lower cost and memory requirement per iteration. In parallel with this review of variable projection algorithms, we develop new formulae for the Jacobian and He …

Card 5 · Low-rank tensor completion: a Riemannian manifold preconditioning approach

We propose a novel Riemannian manifold preconditioning approach for the tensor completion problem with rank constraint. A novel Riemannian metric or inner product is proposed that exploits the least-squares structure of the cost function and takes into account the structured symmetry that exists in…

Read full passage excerpt

We propose a novel Riemannian manifold preconditioning approach for the tensor completion problem with rank constraint. A novel Riemannian metric or inner product is proposed that exploits the least-squares structure of the cost function and takes into account the structured symmetry that exists in Tucker decomposition. The specific metric allows to use the versatile framework of Riemannian optimization on quotient manifolds to develop preconditioned nonlinear conjugate gradient and stochastic gradient descent algorithms for batch and online setups, respectively. Concrete matrix representations of various optimization-related ingredients are listed. Numerical comparisons suggest that our proposed algorithms robustly outperform state-of-the-art algorithms across different synthetic and real-world datasets.

Model-reported card use

Card 2 · fact

Affected output: [51] attains O(n) per-iteration complexity

Model explanation: Supplies the concrete O(n) per-iteration complexity value for the subspace descent method, which the draft had left as an unfinished clause.

Card 2 · scope

Affected output: only for functions whose Riemannian gradients admit efficient computation

Model explanation: Narrows the O(n) complexity claim to the restricted class of functions with efficiently computable Riemannian gradients, preventing an overgeneralized field-wide claim.

No Evidence

Same model and position · no retrieved passages · first output

Not preferred in this comparison

case, whereas [51] retains an O(n3) per-iteration cost.

Earlier card-use comparison

The earlier diagnostic compared only cases with reported card use and text on both sides; its scope differs from the overall AI review above.

The blinded model review preferred the Evidence version. Its original rationale appears below.

Read the original model judgment

Output A is grounded in Source [51] ("Low-complexity subspace-descent over symmetric positive definite manifold"), which explicitly states that [43] achieves O(n²) per-iteration complexity for the general case, while the proposed algorithms (which include [51]'s method) achieve O(n) and O(n²) complexities respectively, with the O(n) variant specifically for cases where "Riemannian gradients can be calculated efficiently." The draft already establishes [43] as O(n²) for the general case, and Source [51] confirms its own method has lower complexity with the efficient gradient caveat. Output B's claim that [51] 'retains an O(n³) per-iteration cost' directly contradicts Source [51], which advertises O(n) and O(n²) complexities versus existing methods' O(n³). This is a material factual error making B unsupported and unusable.

The passages are input material; their presence does not mean an output used them or that they support every claim. Draft and source passages are excerpted; no simulated scores or PDF appear here.