SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2501.18092 · NeurIPS · 2025

Learning Provably Improves the Convergence of Gradient Descent

Wei Lin, Qingyu Song, Hong Xu

arXiv · PDF · Open in the Atlas

Code that ran

1 of 1 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 1 failure of our instrument; 0 no recorded run), measured 2026-11-22; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2501.18092.json · badge: /api/ran/2501.18092.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
NetX-lab/MathL2OProof-Official canonical (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
FunctionStatusWhere it lives
unroll_BP Ran
ran; our contract draft was wrong, not the code
NetX-lab/MathL2OProof-Official/main_train_anlys.py
code served (permissive licence) · get_code("6eeb77dd307deac7")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
from collections import defaultdict

torch.manual_seed(0)

# Create a simple optimizee class
class SimpleOptimizee:
    def __init__(self):
        self.params = nn.Parameter(torch.randn(4))
        self.target = torch.zeros(4)
        
    def objective(self, compute_grad=False):
        if compute_grad:
            loss = ((self.params - self.target) ** 2).sum()
        else:
            with torch.no_grad():
                loss = ((self.params - self.target) ** 2).sum()
        return loss
    
    def detach_vars(self):
        self.params = self.params.detach()
        self.params.requires_grad = True

# Create a simple learned optimizer
class LearnedOptimizer(nn.Module):
    def __init__(self):
        super().__init__()
        self.step_size = nn.Parameter(torch.tensor(0.01))
        self.hidden_state = None
        
    def forward(self, optimizee, grad_method):
        # Compute gradient
        loss = optimizee.objective(compute_grad=True)
        if optimizee.params.grad is not None:
            optimizee.params.grad.zero_()
        loss.backward(retain_graph=True)
        
        # Update parameters
        with torch.no_grad():
            optimizee.params.data = optimizee.params.data - self.step_size * optimizee.params.grad
        
        # Clear gradients for next iteration
        if optimizee.params.grad is not None:
            optimizee.params.grad.zero_()
        
        return optimizee
    
    def detach_state(self):
        if self.hidden_state is not None:
            self.hidden_state = self.hidden_state.detach()

# Create opts object
class Opts:
    def __init__(self):
        self.loss_func = 'mean'  # or 'last', 'weighted_sum'
        self.grad_method = 'sgd'
        self.clip_grad = True

opts = Opts()

# Create optimizer and meta-optimizer
optimizer = LearnedOptimizer()
meta_optimizer = torch.optim.Adam(optimizer.parameters(), lr=0.001)

# Create optimizee
optimizees = SimpleOptimizee()

# Create training losses logs
training_losses_logs = defaultdict(list)

# Set parameters
unroll_length = 5
optimizer_training_steps = 5

ARGS = {
    'unroll_length': unroll_length,
    'optimizer_training_steps': optimizer_training_steps,
    'meta_optimizer': meta_optimizer,
    'optimizer': optimizer,
    'optimizees': optimizees,
    'opts': opts,
    'training_losses_logs': training_losses_logs
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block

Also present in other papers' code

Of 1 samples lifted from this paper's linked repositories, 1 carry a function identity and 0 do not. Of the 1 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 1 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

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

Abstract

Learn to Optimize (L2O) trains deep neural network-based solvers for optimization, achieving success in accelerating convex problems and improving non-convex solutions. However, L2O lacks rigorous theoretical backing for its own training convergence, as existing analyses often use unrealistic assumptions-a gap this work highlights empirically. We bridge this gap by proving the training convergence of L2O models that learn Gradient Descent (GD) hyperparameters for quadratic programming, leveraging the Neural Tangent Kernel (NTK) theory. We propose a deterministic initialization strategy to support our theoretical results and promote stable training over extended optimization horizons by mitigating gradient explosion. Our L2O framework demonstrates over 50% better optimality than GD and superior robustness over state-of-the-art L2O methods on synthetic datasets. The code of our method can be found from https://github.com/NetX-lab/MathL2OProof-Official.

For agents

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

get_harvested_code_for_paper("2501.18092")
get_code_for_paper("2501.18092")
have("2501.18092")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2501.18092.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2501.18092.svg)](https://syntology.ai/paper/2501.18092)

Connect an agent — have() is free.