SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2510.21153 · NeurIPS · 2025

Uncertainty-Aware Multi-Objective Reinforcement Learning-Guided Diffusion Models for 3D De Novo Molecular Design

Pingzhao Hu, Dongkyu Kim, Lianghong Chen, Mike Domaratzki

arXiv · PDF · Open in the Atlas

Code that ran

5 of 7 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; 2 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 3 failures of our instrument; 2 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/2510.21153.json · badge: /api/ran/2510.21153.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
Kyle4490/RL-Diffusion community
Papers with Code archive
5 ran · 2 no recorded run
FunctionStatusWhere it lives
ei Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Kyle4490/RL-Diffusion/uncertainty/get_dist.py
code served (permissive licence) · get_code("548ebe17c3ad0c18")
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 ndarray, ndim 1, dtype family float, 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
import pandas as pd
from scipy.stats import norm

# Create a small DataFrame with the required columns
# columns_list[-2] is predictions, columns_list[-1] is variances
np.random.seed(0)
n_samples = 5

df = pd.DataFrame({
    'pred_col': np.random.randn(n_samples) * 2 - 1,  # predictions
    'var_col': np.abs(np.random.randn(n_samples)) + 0.1,  # variances
})

columns_list = ['pred_col', 'var_col']

cutoffs_dict = {
    'qm9': {'qed': 0.41, 'sas': 6.72, 'affinity': -4.73},
    'zinc15': {'qed': 0.4, 'sas': 7.326, 'affinity': -5.76},
    'geom': {'qed': 0.653, 'sas': 7.196, 'affinity': -6.73},
    'pubchem': {'qed': 0.651, 'sas': 7.183, 'affinity': -6.616}
}

objectives_dict = {'qed': 'max', 'sas': 'min', 'affinity': 'min'}

ARGS = {
    'df': df,
    'columns_list': columns_list,
    'cutoffs': cutoffs_dict,
    'dataset': 'qm9',
    'property': 'qed',
    'objectives': objectives_dict,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, pandas, 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 float, 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
expected_improvement Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Kyle4490/RL-Diffusion/uncertainty/get_dist.py
code served (permissive licence) · get_code("13b162ee180dfe7d")
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 ndarray, ndim 1, dtype family float, shape (4,) 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.stats import norm

# Create small test data for the ei function
np.random.seed(0)

# Create a simple dataframe-like object using a dict
df_data = {
    'pred_col': np.array([0.5, 0.6, 0.7, 0.8]),
    'var_col': np.array([0.1, 0.2, 0.15, 0.25])
}

class SimpleDataFrame:
    def __init__(self, data):
        self.data = data
    
    def __getitem__(self, key):
        return self.data[key]

df = SimpleDataFrame(df_data)

# Parameters for ei function
columns_list = ['pred_col', 'var_col']
dataset = 'qm9'
property = 'qed'
objectives = {'qed': 'max', 'sas': 'min', 'affinity': 'min'}

cutoffs_dict = {
    'qm9': {'qed': 0.41, 'sas': 6.72, 'affinity': -4.73},
    'zinc15': {'qed': 0.4, 'sas': 7.326, 'affinity': -5.76},
    'geom': {'qed': 0.653, 'sas': 7.196, 'affinity': -6.73},
    'pubchem': {'qed': 0.651, 'sas': 7.183, 'affinity': -6.616}
}

# Parameters for expected_improvement function
predictions = np.array([0.5, 0.6, 0.7, 0.8])
variances = np.array([0.1, 0.2, 0.15, 0.25])
cutoff = 0.41
minimize = False

ARGS = {
    'predictions': predictions,
    'variances': variances,
    'cutoff': cutoff,
    'minimize': minimize
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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 float, shape (4,) — shape (4,) 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
gaussian_cdf Ran
ran, honoured the contract we drafted
Kyle4490/RL-Diffusion/uncertainty/get_dist.py
code served (permissive licence) · get_code("76655e2a4abdcda7")
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 float64, ndim 0, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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
from scipy.stats import norm

torch.manual_seed(0)

# Analyze the function:
# - mean: used as first argument to norm(), expects a scalar or array-like
# - variance: used as base for ** 0.5 to compute std_dev, expects a numeric value
# - cutoff: used as argument to dist.cdf(), expects a scalar or array-like

ARGS = {
    "mean": 0.0,
    "variance": 1.0,
    "cutoff": 0.5
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, scipy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
gaussian_cdf_determined Ran
ran, honoured the contract we drafted
Kyle4490/RL-Diffusion/uncertainty/get_dist.py
code served (permissive licence) · get_code("3fc608f9e112f2f1")
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 float64, ndim 0, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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.

from scipy.stats import norm
import numpy as np

ARGS = {
    'mean': 0.5,
    'variance': 1.0,
    'cutoff': 1.5,
    'objective': 'max'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
pio Ran
ran; our contract draft was wrong, not the code
Kyle4490/RL-Diffusion/uncertainty/get_dist.py
code served (permissive licence) · get_code("1b729d0449239635")
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 ndarray, ndim 1, dtype family float, shape (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 pandas as pd
import numpy as np

# Create a small DataFrame with the required columns
# The function expects columns_list[-2] to be mean and columns_list[-1] to be variance
df = pd.DataFrame({
    'mean_col': [0.5, 0.3, 0.7],
    'variance_col': [0.1, 0.15, 0.2]
})

columns_list = ['mean_col', 'variance_col']
dataset = 'qed'
property = 'qed'

ARGS = {
    'df': df,
    'columns_list': columns_list,
    'cutoffs': {'qed': {'qed': 0.41, 'sas': 6.72, 'affinity': -4.73}, 'zinc15': {'qed': 0.4, 'sas': 7.326, 'affinity': -5.76}, 'geom': {'qed': 0.653, 'sas': 7.196, 'affinity': -6.73}, 'pubchem': {'qed': 0.651, 'sas': 7.183, 'affinity': -6.616}},
    'dataset': dataset,
    'property': property,
    'objectives': {'qed': 'max', 'sas': 'min', 'affinity': 'min'}
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, pandas, 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 float, shape (3,) — shape (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_pio_score Not yet run
unverified (harvested, no recorded run)
Kyle4490/RL-Diffusion/uncertainty/get_dist.py
code served (permissive licence) · get_code("8e25bd68f799db1a")
How it ran

no invocation recorded.

set_cutoff Not yet run
unverified (harvested, no recorded run)
Kyle4490/RL-Diffusion/uncertainty/get_dist.py
code served (permissive licence) · get_code("35125ab93ce63b54")
How it ran

no invocation recorded.

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 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 7 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

Designing de novo 3D molecules with desirable properties remains a fundamental challenge in drug discovery and molecular engineering. While diffusion models have demonstrated remarkable capabilities in generating high-quality 3D molecular structures, they often struggle to effectively control complex multi-objective constraints critical for real-world applications. In this study, we propose an uncertaintyaware Reinforcement Learning (RL) framework to guide the optimization of 3D molecular diffusion models toward multiple property objectives while enhancing the overall quality of the generated molecules. Our method leverages surrogate models with predictive uncertainty estimation to dynamically shape reward functions, facilitating balance across multiple optimization objectives. We comprehensively evaluate our framework across three benchmark datasets and multiple diffusion model architectures, consistently outperforming baselines for molecular quality and property optimization. Additionally, Molecular Dynamics (MD) simulations and ADMET profiling of top generated candidates indicate promising drug-like behavior and binding stability, comparable to known Epidermal Growth Factor Receptor (EGFR) inhibitors. Our results demonstrate the strong potential of RL-guided generative diffusion models for advancing automated molecular design. The implementation is available at https://github.com/Kyle4490/RL-Diffusion.

For agents

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

get_harvested_code_for_paper("2510.21153")
get_code_for_paper("2510.21153")
have("2510.21153")

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

curl https://syntology.ai/api/ran/2510.21153.json

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

[![Syntology run record](https://syntology.ai/api/ran/2510.21153.svg)](https://syntology.ai/paper/2510.21153)

Connect an agent — have() is free.