SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · aaai_25639 · AAAI · 2023

MDM: Molecular Diffusion Model for 3D Molecule Generation

Lei Huang, Tingyang Xu, Ka-Chun Wong, Hengtong Zhang

Code that ran

6 of 10 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 5 have no contract verdict; 1 failure of our instrument; 4 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/aaai_25639.json · badge: /api/ran/aaai_25639.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
tencent-ailab/MDM canonical (classifier label)
classifier basis: census
6 ran · 4 no recorded run
FunctionStatusWhere it lives
RMSD Ran
ran on a synthesized input
tencent-ailab/MDM/eval_qm9_condition_quality.py
code served (permissive licence) · get_code("0895bc97d2c8733a")
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 float, ndim 0, dtype family float, shape () 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 math
import torch

# Construct valid arguments for RMSD function
# The function expects:
# - probe: a list of coordinate tuples/lists (one per atom)
# - ref: a list of coordinate tuples/lists (one per atom)
# Both must have the same length (atom_num)

# Create two lists of atomic coordinates
# Each atom has xyz coordinates (3D positions)
probe = [
    [1.0, 2.0, 3.0],
    [4.0, 5.0, 6.0],
]

ref = [
    [1.1, 2.1, 3.1],
    [4.1, 5.1, 6.1],
]

ARGS = {
    "probe": probe,
    "ref": ref,
}
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 float, ndim 0, dtype family float, shape () — shape () 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
assemble_atom_pair_feature Ran
ran on a synthesized input
tencent-ailab/MDM/models/common.py
code served (permissive licence) · get_code("3fe49c0a88c25587")
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 Tensor, ndim 2, dtype family float, shape (3, 7) 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 torch

torch.manual_seed(0)

# node_attr: tensor of shape (num_nodes, node_feature_dim)
# For 2 nodes with feature dimension 4
node_attr = torch.randn(2, 4)

# edge_index: tensor of shape (2, num_edges)
# Contains node indices for edges
# Let's create 3 edges: (0,1), (1,0), (0,0)
edge_index = torch.tensor([[0, 1, 0], [1, 0, 0]], dtype=torch.long)

# edge_attr: tensor of shape (num_edges, edge_feature_dim) or None
# With 3 edges and edge feature dimension 3
edge_attr = torch.randn(3, 3)

ARGS = {
    "node_attr": node_attr,
    "edge_index": edge_index,
    "edge_attr": edge_attr,
}
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 Tensor, ndim 2, dtype family float, shape (3, 7) — shape (3, 7) 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
dist_2 Ran
ran on a synthesized input
tencent-ailab/MDM/eval_qm9_condition_quality.py
code served (permissive licence) · get_code("fef578ae782e82b6")
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 float, ndim 0, dtype family float, shape () 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 torch
import numpy as np

torch.manual_seed(0)

# The function dist_2 takes two iterables (atoma_xyz, atomb_xyz) and computes
# the squared Euclidean distance by iterating through paired elements and
# summing squared differences.

# Create two simple iterables representing atomic coordinates
# Using lists of floats to match the zip() and iteration pattern
atoma_xyz = [1.0, 2.0, 3.0, 4.0]
atomb_xyz = [1.5, 2.5, 3.5, 4.5]

ARGS = {
    "atoma_xyz": atoma_xyz,
    "atomb_xyz": atomb_xyz,
}
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 float, ndim 0, dtype family float, shape () — shape () 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
generate_symmetric_edge_noise Ran
ran on a synthesized input
tencent-ailab/MDM/models/common.py
code served (permissive licence) · get_code("ab68381cba40254d")
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 Tensor, ndim 2, dtype family float, shape (5, 1) 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 torch

torch.manual_seed(0)

# num_nodes_per_graph: tensor of number of nodes in each graph
# We'll create 2 graphs with 3 and 4 nodes respectively
num_nodes_per_graph = torch.tensor([3, 4])

# edge_index: tensor of shape [2, num_edges] with node indices
# Create 5 edges total (some in graph 0, some in graph 1)
edge_index = torch.tensor([
    [0, 1, 2, 3, 4],  # source nodes
    [1, 2, 0, 4, 5]   # target nodes
])

# edge2graph: tensor mapping each edge to which graph it belongs to
# 2 edges in graph 0, 3 edges in graph 1
edge2graph = torch.tensor([0, 0, 1, 1, 1])

# device: CPU device
device = torch.device('cpu')

ARGS = {
    'num_nodes_per_graph': num_nodes_per_graph,
    'edge_index': edge_index,
    'edge2graph': edge2graph,
    'device': device
}
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 Tensor, ndim 2, dtype family float, shape (5, 1) — shape (5, 1) 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
get_distance Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
tencent-ailab/MDM/models/geometry.py
pointer only (licence: MIT); the stored copy get_code serves, from huang3170/MolForm, is pointer only (licence: NOASSERTION) · get_code("5befa7d29d90c70b")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from huang3170/MolForm, and GitHub could not classify the licence of that repository (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

padding Ran
ran on a synthesized input
tencent-ailab/MDM/eval_qm9_condition.py
code served (permissive licence) · get_code("ae350126a5ebe28f")
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 Tensor, ndim 3, dtype family float, shape (3, 5, 3) 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 torch

torch.manual_seed(0)

# Create a list of tensors with different sizes (all < max_nodes)
# Each tensor has shape (num_nodes, num_features)
tensor1 = torch.randn(2, 3)  # 2 nodes, 3 features
tensor2 = torch.randn(3, 3)  # 3 nodes, 3 features
tensor3 = torch.randn(1, 3)  # 1 node, 3 features

data_list = [tensor1, tensor2, tensor3]
max_nodes = 5

ARGS = {
    "data_list": data_list,
    "max_nodes": max_nodes,
}
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 Tensor, ndim 3, dtype family float, shape (3, 5, 3) — shape (3, 5, 3) 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
get_dataset_info Not yet run
unverified (harvested, no recorded run)
tencent-ailab/MDM/configs/datasets_config.py
code served (permissive licence) · get_code("f113f8cda579e6ff")
How it ran

no invocation recorded.

get_edge_encoder Not yet run
unverified (harvested, no recorded run)
tencent-ailab/MDM/models/encoder/edge.py
code served (permissive licence) · get_code("9709623b5759d7ad")
How it ran

no invocation recorded.

load_split_data Not yet run
unverified (harvested, no recorded run)
tencent-ailab/MDM/build_geom_dataset.py
code served (permissive licence) · get_code("edf83b952680cf55")
How it ran

no invocation recorded.

sample Not yet run
unverified (harvested, no recorded run)
tencent-ailab/MDM/eval_qm9_condition.py
code served (permissive licence) · get_code("5f22f11a47853b8d")
How it ran

no invocation recorded.

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 distinct functions with an identity, 4 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 0 the same month, 2 later) and 6 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
get_distance 3 other papers
1 dated earlier, 2 later
2203.02923 (2022-03)
GeoDiff: a Geometric Diffusion Model for Molecular Conformation Generation
2406.16821 (2024-06)
General Binding Affinity Guidance for Diffusion Models in Structure-Based Drug Design
2507.05503 (2025-07)
MolFORM: Multi-modal Flow Matching for Structure-Based Drug Design
load_split_data 3 other papers
1 dated earlier, 1 later, 1 same year, order not determinable
2203.17003 (2022-03)
Equivariant Diffusion for Molecule Generation in 3D
2305.01140 (2023-05)
Geometric Latent Diffusion Models for 3D Molecule Generation
2410.06502 (2024-10)
Chemistry-Inspired Diffusion with Non-Differentiable Guidance
generate_symmetric_edge_noise 1 other paper
1 dated earlier
2203.02923 (2022-03)
GeoDiff: a Geometric Diffusion Model for Molecular Conformation Generation
get_edge_encoder 1 other paper
1 dated earlier
2203.02923 (2022-03)
GeoDiff: a Geometric Diffusion Model for Molecular Conformation Generation

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

Molecule generation, especially generating 3D molecular geometries from scratch (i.e., 3D de novo generation), has become a fundamental task in drug design. Existing diffusion based 3D molecule generation methods could suffer from unsatisfactory performances, especially when generating large molecules. At the same time, the generated molecules lack enough diversity. This paper proposes a novel diffusion model to address those two challenges. First, interatomic relations are not included in molecules' 3D point cloud representations. Thus, it is difficult for existing generative models to capture the potential interatomic forces and abundant local constraints. To tackle this challenge, we propose to augment the potential interatomic forces and further involve dual equivariant encoders to encode interatomic forces of different strengths. Second, existing diffusion-based models essentially shift elements in geometry along the gradient of data density. Such a process lacks enough exploration in the intermediate steps of the Langevin dynamics. To address this issue, we introduce a distributional controlling variable in each diffusion/reverse step to enforce thorough explorations and further improve generation diversity. Extensive experiments on multiple benchmarks demonstrate that the proposed model significantly outperforms existing methods for both unconditional and conditional generation tasks. We also conduct case studies to help understand the physicochemical properties of the generated molecules. The codes are available at https://github.com/tencent-ailab/MDM * This work is done when Lei Huang works as an intern in Tencent AI Lab.

For agents

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

get_harvested_code_for_paper("aaai_25639")
get_code_for_paper("aaai_25639")
have("aaai_25639")

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

curl https://syntology.ai/api/ran/aaai_25639.json

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

[![Syntology run record](https://syntology.ai/api/ran/aaai_25639.svg)](https://syntology.ai/paper/aaai_25639)

Connect an agent — have() is free.