On Parameter Selection for First-Order Methods: A Matrix Analysis Approach
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
First-order convex optimization algorithms are popular due to their computational attractiveness and applicability to a wide range of domains such as machine learning and control. Despite the substantial progress being made over the last few decades, some open questions related to their convergence remain unaddressed. In addition, majority of the first-order methods assume strong convexity to analyze both the stability of the method and derive an explicit convergence rate. In this manuscript, we relax the strong convexity condition, and then, lay out two main contributions. First, we provide a methodology where one can analyze the speed of convergence of the algorithm using the contractive theory and linear algebra. Second, we find explicit values of the tuning parameters of the Double Momentum Algorithm (which unifies many of the popular algorithms), ensuring stability for gradient L-Lipchitz functions. In this work, while an explicit convergence rate is not provided, the foundational results serve as a stepping stone in that direction, as we provide an explicit non-asymptotic rate. Furthermore, our numerical experiments demonstrate superior performance of the proposed method. Beyond optimization, we also apply our method to two-sided markets in non-cooperative game theory.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.006 | 0.011 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.014 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it