SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2302.08224 · NeurIPS · 2023

DIFUSCO: Graph-based Diffusion Solvers for Combinatorial Optimization

Yiming Yang, Zhiqing Sun

arXiv · PDF · Open in the Atlas

Code that ran

7 of 9 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 6 have no contract verdict; 1 failure of our instrument; 2 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; 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/2302.08224.json · badge: /api/ran/2302.08224.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
edward-sun/difusco canonical (classifier label)
classifier basis not recorded on the link
7 ran · 2 no recorded run
FunctionStatusWhere it lives
batched_two_opt_torch Ran
ran on a synthesized input
edward-sun/difusco/difusco/utils/tsp_utils.py
code served (permissive licence) · get_code("9587e3297756f4d3")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) 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 numpy as np

torch.manual_seed(0)

# Create small point cloud: batch_size=2, num_points=4, coordinates=2D
num_points = 4
batch_size = 2
points = np.random.rand(num_points, 2).astype(np.float32)

# Create initial tours: batch_size=2, tour_length=num_points+1 (closed tour)
# Each tour is a sequence of point indices, with first point repeated at end
tour = np.array([
    [0, 1, 2, 3, 0],
    [0, 2, 3, 1, 0]
], dtype=np.int64)

ARGS = {
    'points': points,
    'tour': tour,
    'max_iterations': 10,
    'device': 'cpu'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, 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
get_one_cycle Ran
ran on a synthesized input
edward-sun/difusco/difusco/utils/lr_schedulers.py
code served (permissive licence) · get_code("0c44d4af4c356ee5")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned LambdaLR. 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 torch.optim import SGD

torch.manual_seed(0)

# Build a minimal model
model = nn.Linear(4, 3)

# Create an optimizer
optimizer = SGD(model.parameters(), lr=0.1)

# Define the number of training steps
num_training_steps = 100

ARGS = {
    "optimizer": optimizer,
    "num_training_steps": num_training_steps,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type LambdaLR — a LambdaLR output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a LambdaLR output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
get_schedule_fn Ran
ran on a synthesized input
edward-sun/difusco/difusco/utils/lr_schedulers.py
code served (permissive licence) · get_code("b04bf96acaa8e607")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned partial. 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 torch.optim import SGD
from functools import partial
from torch.optim.lr_scheduler import LambdaLR

torch.manual_seed(0)

# Create a minimal model
class TinyModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def forward(self, x):
        return self.linear(x)

# Instantiate model and optimizer
model = TinyModel()
optimizer = SGD(model.parameters(), lr=0.1)

# Define arguments
ARGS = {
    'scheduler': 'one-cycle',
    'num_training_steps': 100
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type partial — a partial output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a partial output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
mis_decode_np Ran
ran on a synthesized input
edward-sun/difusco/difusco/utils/mis_utils.py
code served (permissive licence) · get_code("0b45baf887c9bce6")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 1, dtype family int, shape (5,) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape 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 numpy as np
from scipy import sparse

# Create predictions: a 1D array of probabilities
predictions = np.array([0.9, 0.3, 0.7, 0.2, 0.5])

# Create a sparse adjacency matrix (5x5 for a 5-node graph)
# Using COO format which has .tocsr() method
row = np.array([0, 0, 1, 2, 2, 3, 4])
col = np.array([1, 2, 3, 0, 4, 1, 2])
data = np.array([1, 1, 1, 1, 1, 1, 1])
adj_matrix = sparse.coo_matrix((data, (row, col)), shape=(5, 5))

ARGS = {
    "predictions": predictions,
    "adj_matrix": adj_matrix
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family int, shape (5,) — shape (5,) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
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
numpy_merge Ran
ran on a synthesized input
edward-sun/difusco/difusco/utils/tsp_utils.py
code served (permissive licence) · get_code("3f4c5b6a015b2fa5")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) 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 numpy as np

# Construct valid arguments for numpy_merge function
# The function expects:
# - points: array of shape (n_points, n_dims) - used to compute pairwise distances
# - adj_mat: square matrix of shape (n_points, n_points) - adjacency/weight matrix

# Create a small set of points in 2D space
points = np.array([
    [0.0, 0.0],
    [1.0, 0.0],
    [2.0, 0.0],
    [0.0, 1.0]
], dtype=np.float32)

# Create a small adjacency matrix with 4 nodes
# The function uses this to compute weights and sort edges
adj_mat = np.array([
    [0.0, 0.5, 1.0, 2.0],
    [0.5, 0.0, 0.3, 1.5],
    [1.0, 0.3, 0.0, 1.2],
    [2.0, 1.5, 1.2, 0.0]
], dtype=np.float32)

ARGS = {
    "points": points,
    "adj_mat": adj_mat
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — 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
run_sparse_layer Ran
ran on a synthesized input
edward-sun/difusco/difusco/models/gnn_encoder.py
code served (permissive licence) · get_code("bee320771d4e98eb")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned function. 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

torch.manual_seed(0)

# Define minimal mock modules that satisfy the function's usage

class MockSparseLayer(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear_x = nn.Linear(4, 4)
        self.linear_e = nn.Linear(3, 3)
    
    def forward(self, x_in, e_in, adj_matrix, mode='direct', edge_index=None, sparse=True):
        # Returns transformed x and e
        x = self.linear_x(x_in)
        e = self.linear_e(e_in)
        return x, e

class MockTimeLayer(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(5, 4)
    
    def forward(self, time_emb):
        # Takes time embedding and returns same shape as x or e depending on context
        return self.linear(time_emb)

class MockOutLayer(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(3, 3)
    
    def forward(self, e):
        # Takes edge features and returns same shape
        return self.linear(e)

# Create instances
layer = MockSparseLayer()
time_layer = MockTimeLayer()
out_layer = MockOutLayer()

# Create sparse adjacency matrix (2x2 for small graph)
adj_matrix = torch.sparse_coo_tensor(
    indices=torch.tensor([[0, 1], [1, 0]]),
    values=torch.tensor([1.0, 1.0]),
    size=(2, 2)
)

# Create edge index for sparse representation
edge_index = torch.tensor([[0, 1], [1, 0]])

ARGS = {
    'layer': layer,
    'time_layer': time_layer,
    'out_layer': out_layer,
    'adj_matrix': adj_matrix,
    'edge_index': edge_index,
    'add_time_on_edge': True
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
zero_module Ran
ran; our contract draft was wrong, not the code
edward-sun/difusco/difusco/models/nn.py
code served (permissive licence) · get_code("129b804760b3115f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned SimpleModule. 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

torch.manual_seed(0)

# Create a simple module with some parameters
class SimpleModule(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
        self.conv = nn.Conv1d(2, 2, 3)
    
    def forward(self, x):
        return x

module = SimpleModule()

# Verify the module has non-zero parameters before zeroing
for p in module.parameters():
    assert p.abs().sum() > 0, "Parameters should be non-zero initially"

ARGS = {
    "module": module
}
Setup block drafted by
a language model, Claude Haiku 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 SimpleModule — a SimpleModule output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a SimpleModule output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
avg_pool_nd Not yet run
unverified (harvested, no recorded run)
edward-sun/difusco/difusco/models/nn.py
code served (permissive licence) · get_code("ecd0fc28815b65ae")
How it ran

no invocation recorded.

conv_nd Not yet run
unverified (harvested, no recorded run)
edward-sun/difusco/difusco/models/nn.py
code served (permissive licence) · get_code("fe4eb545bbb728e0")
How it ran

no invocation recorded.

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 distinct functions with an identity, 8 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 3 the same month, 8 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.

FunctionAlso present inEarliest-datedOthers named
zero_module 196 other papers
31 dated earlier, 2 the same month, 162 later, 1 same year, order not determinable · 49 only as an assembly brick
2006.11239 (2020-06)
Denoising Diffusion Probabilistic Models
2010.02502 (2020-10)
Denoising Diffusion Implicit Models
2105.05233 (2021-05)
Diffusion Models Beat GANs on Image Synthesis
and 193 more not named here
conv_nd 121 other papers
29 dated earlier, 2 the same month, 89 later, 1 same year, order not determinable · 32 only as an assembly brick
2006.11239 (2020-06)
Denoising Diffusion Probabilistic Models
2010.02502 (2020-10)
Denoising Diffusion Implicit Models
2111.11672 (2021-11)
Few-shot Image Generation with Mixup-based Distance Learning
and 118 more not named here
avg_pool_nd 113 other papers
30 dated earlier, 2 the same month, 80 later, 1 same year, order not determinable · 31 only as an assembly brick
2006.11239 (2020-06)
Denoising Diffusion Probabilistic Models
2010.02502 (2020-10)
Denoising Diffusion Implicit Models
2110.11291 (2021-10)
Likelihood Training of Schrödinger Bridge using Forward-Backward SDEs Theory
and 110 more not named here
batched_two_opt_torch 1 other paper
1 later
2501.13767 (2025-01)
An Efficient Diffusion-based Non-Autoregressive Solver for Traveling Salesman Problem
get_one_cycle 1 other paper
1 later
2501.13767 (2025-01)
An Efficient Diffusion-based Non-Autoregressive Solver for Traveling Salesman Problem
get_schedule_fn 1 other paper
1 later
2501.13767 (2025-01)
An Efficient Diffusion-based Non-Autoregressive Solver for Traveling Salesman Problem
numpy_merge 1 other paper
1 later
2501.13767 (2025-01)
An Efficient Diffusion-based Non-Autoregressive Solver for Traveling Salesman Problem
run_sparse_layer 1 other paper
1 later
2502.02941 (2025-02)
Fast T2T: Optimization Consistency Speeds Up Diffusion-Based Training-to-Testing Solving for Combinatorial Optimization

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

Neural network-based Combinatorial Optimization (CO) methods have shown promising results in solving various NP-complete (NPC) problems without relying on hand-crafted domain knowledge. This paper broadens the current scope of neural solvers for NPC problems by introducing a new graph-based diffusion framework, namely DIFUSCO. Our framework casts NPC problems as discrete {0, 1}-vector optimization problems and leverages graph-based denoising diffusion models to generate high-quality solutions. We investigate two types of diffusion models with Gaussian and Bernoulli noise, respectively, and devise an effective inference schedule to enhance the solution quality. We evaluate our methods on two well-studied NPC combinatorial optimization problems: Traveling Salesman Problem (TSP) and Maximal Independent Set (MIS). Experimental results show that DIFUSCO strongly outperforms the previous state-of-the-art neural solvers, improving the performance gap between ground-truth and neural solvers from 1.76% to 0.46% on TSP-500, from 2.46% to 1.17% on TSP-1000, and from 3.19% to 2.58% on TSP-10000. For the MIS problem, DIFUSCO outperforms the previous state-of-the-art neural solver on the challenging SATLIB benchmark.

For agents

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

get_harvested_code_for_paper("2302.08224")
get_code_for_paper("2302.08224")
have("2302.08224")

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

curl https://syntology.ai/api/ran/2302.08224.json

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

[![Syntology run record](https://syntology.ai/api/ran/2302.08224.svg)](https://syntology.ai/paper/2302.08224)

Connect an agent — have() is free.