BEYOND NEWTON: ROBUST METHODS FOR SOLVING LARGE NONLINEAR MODELS IN TROLL
Bibliographic record
Abstract
Newton's Method is an important algorithm for solving nonlinear systems of equations. For any solution algorithm, the principle concerns are robustness (finding a solution reliably) and efficiency (finding a solution quickly). Newton is simple in principle, but a useful implementation must deal with a variety of practical and theoretical obstacles.By using partial derivatives, Newton's Method can model the shape of the residual surface to provide quadratic convergence near the solution: the number of correct digits doubles each iteration. But far from the solution, the full Newton step may go too far, leading to divergence or oscillation. Or the full step may be illegal, leading to economic nonsense like negative prices and numerical problems like taking the log of a negative number. Automatic damping -- taking shorter steps along the Newton direction -- can improve global convergence in such cases.The most expensive part of Newton's Method is factoring the Jacobian matrix. The Jacobian can be very large in contemporary macroeconometric models, particularly forward-looking models that introduce simultaneity across time periods as well as between equations. Newton is impractical for such large models unless it can exploit the sparsity of the Jacobian. Sparse direct methods can be applied to the entire Jacobian; one advantage of direct methods is that after the first iteration, subsequent factorings can be extremely fast. An alternative is iterative Krylov subspace methods; their small memory requirements are an advantage, but they may fail to converge. In the case of forward-looking models, the stacked Jacobian has a block band-diagonal structure that can be exploited by direct methods or for preconditioning iterative methods.his paper describes enhancements to Newton's Method used in the TROLL modeling system and illustrates them with a variety of contemporary models.
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.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.005 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.003 |
| Scholarly communication | 0.002 | 0.003 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.003 | 0.004 |
| Insufficient payload (model declined to judge) | 0.011 | 0.004 |
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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".