SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2406.02740 · ICML · 2024

Long Range Propagation on Continuous-Time Dynamic Graphs

Davide Bacciu, Alessio Gravina, Claudio Gallicchio, Giulio Lovisotto, Claas Grohnfeldt

arXiv · PDF · Open in the Atlas

Code that ran

3 of 4 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (1 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 1 has no contract verdict; 2 failures of our instrument; 1 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/2406.02740.json · badge: /api/ran/2406.02740.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
gravins/non-dissipative-propagation-ctdgs canonical (classifier label)
classifier basis: classifier_2025
2 ran · 0 no recorded run
gravins/non-dissipative-propagation-CTDGs marked official in Papers with Code
Papers with Code archive
1 ran · 1 no recorded run
FunctionStatusWhere it lives
GenericModel Ran
ran on a synthesized input · constructed an object; no result was computed
gravins/non-dissipative-propagation-CTDGs/models/ctdg_models.py
code served (permissive licence) · get_code("e95a36ecfd64604b")
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 an instance of its own class, GenericModel: the run constructed the object and did not call it, so no result was computed. 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 a minimal memory module
class MockMemory(nn.Module):
    def __init__(self, num_nodes, memory_dim):
        super().__init__()
        self.num_nodes = num_nodes
        self.memory_dim = memory_dim
        self.memory = nn.Parameter(torch.randn(num_nodes, memory_dim))
        self.last_update = nn.Parameter(torch.zeros(num_nodes))
    
    def forward(self, n_id):
        return self.memory[n_id], self.last_update[n_id]
    
    def reset_state(self):
        pass
    
    def zero_grad_memory(self):
        if self.memory.grad is not None:
            self.memory.grad.zero_()
    
    def update_state(self, src, pos_dst, t, msg):
        pass
    
    def detach(self):
        self.memory.detach_()
    
    def reset_parameters(self):
        nn.init.normal_(self.memory)

# Define a minimal GNN layer
class MockGNNLayer(nn.Module):
    def __init__(self, in_channels, out_channels):
        super().__init__()
        self.linear = nn.Linear(in_channels, out_channels)
    
    def forward(self, z, last_update, edge_index, t, msg):
        return self.linear(z)
    
    def reset_parameters(self):
        self.linear.reset_parameters()

# Define activation function
gnn_act = nn.ReLU()

# Define a minimal readout module
class MockReadout(nn.Module):
    def __init__(self, in_channels, out_channels=1):
        super().__init__()
        self.linear1 = nn.Linear(in_channels, out_channels)
        self.linear2 = nn.Linear(2 * in_channels, out_channels)
    
    def forward(self, *args):
        if len(args) == 1:
            return self.linear1(args[0])
        else:
            return self.linear2(torch.cat(args, dim=-1))
    
    def reset_parameters(self):
        self.linear1.reset_parameters()
        self.linear2.reset_parameters()

# Set up model components
num_nodes = 8
memory_dim = 4
node_feat_dim = 3
gnn_in_dim = memory_dim + node_feat_dim
gnn_out_dim = 5

memory = MockMemory(num_nodes, memory_dim)
gnn_layer = MockGNNLayer(gnn_in_dim, gnn_out_dim)
gnn = [gnn_layer]
readout = MockReadout(gnn_out_dim, out_channels=1)

ARGS = {
    'num_nodes': num_nodes,
    'memory': memory,
    'gnn': gnn,
    'gnn_act': gnn_act,
    'readout': readout,
    'predict_dst': False,
}
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 GenericModel — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a GenericModel 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
aggregate_res Ran
ran; our contract draft was wrong, not the code
gravins/non-dissipative-propagation-ctdgs/main_tgb_ctan.py
code served (permissive licence) · get_code("c21c2f155065170c")
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 DataFrame. 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 pandas as pd
import numpy as np

class Args:
    def __init__(self):
        self.num_runs = 2
        self.num_final_runs = 2
        self.metric = 'accuracy'

# Create sample dataframe with multiple runs
data = {
    'conf_id': [1, 1, 2, 2],
    'seed': [42, 43, 44, 45],
    'test_accuracy': [0.85, 0.87, 0.90, 0.88],
    'val_accuracy': [0.82, 0.84, 0.88, 0.86],
    'train_accuracy': [0.95, 0.96, 0.97, 0.96],
    'best_epoch': [10, 12, 8, 9],
    'model_name': ['model_a', 'model_a', 'model_b', 'model_b'],
    'learning_rate': [0.001, 0.001, 0.0005, 0.0005]
}

df = pd.DataFrame(data)

args = Args()

ARGS = {
    'df': df,
    'args': args,
    'is_assessment': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 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
compute_row Ran
ran; our contract draft was wrong, not the code
gravins/non-dissipative-propagation-ctdgs/main_tgb_ctan.py
code served (permissive licence) · get_code("3506c74436204d6e")
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 dict. 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

torch.manual_seed(0)

# test_score: a dict with string keys and numeric values
test_score = {
    'accuracy': 0.92,
    'loss': 0.15,
    'f1': 0.89
}

# val_score: a dict with string keys and numeric values
val_score = {
    'accuracy': 0.88,
    'loss': 0.22,
    'f1': 0.85
}

# train_score: a dict with string keys and numeric values
train_score = {
    'accuracy': 0.95,
    'loss': 0.08,
    'f1': 0.93
}

# best_epoch: an integer
best_epoch = 42

# res_conf: a dict that can contain both nested dicts and scalar values
res_conf = {
    'learning_rate': 0.001,
    'batch_size': 32,
    'model_params': {
        'hidden_dim': 128,
        'num_layers': 3,
        'dropout': 0.5
    },
    'optimizer_params': {
        'weight_decay': 1e-5,
        'momentum': 0.9
    },
    'epochs': 100
}

ARGS = {
    'test_score': test_score,
    'val_score': val_score,
    'train_score': train_score,
    'best_epoch': best_epoch,
    'res_conf': res_conf
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
CTAN Not yet run
unverified (harvested, no recorded run)
gravins/non-dissipative-propagation-CTDGs/models/ctdg_models.py
code served (permissive licence) · get_code("b1d782a58a37a3a7")
How it ran

no invocation recorded.

Also present in other papers' code

Of 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 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 4 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

Learning Continuous-Time Dynamic Graphs (C-TDGs) requires accurately modeling spatiotemporal information on streams of irregularly sampled events. While many methods have been proposed recently, we find that most message passing-, recurrent-or self-attention-based methods perform poorly on long-range tasks. These tasks require correlating information that occurred "far" away from the current event, either spatially (higher-order node information) or along the time dimension (events occurred in the past). To address long-range dependencies, we introduce continuous-time graph anti-symmetric network (CTAN). Grounded within the ordinary differential equations framework, our method is designed for efficient propagation of information. In this paper, we show how CTAN's (i) long-range modeling capabilities are substantiated by theoretical findings and how (ii) its empirical performance on synthetic long-range benchmarks and real-world benchmarks is superior to other methods. Our results motivate CTAN's ability to propagate longrange information in C-TDGs as well as the inclusion of long-range tasks as part of temporal graph models evaluation.

For agents

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

get_harvested_code_for_paper("2406.02740")
get_code_for_paper("2406.02740")
have("2406.02740")

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

curl https://syntology.ai/api/ran/2406.02740.json

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

[![Syntology run record](https://syntology.ai/api/ran/2406.02740.svg)](https://syntology.ai/paper/2406.02740)

Connect an agent — have() is free.