SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2103.06624 · NeurIPS · 2021

Beta-CROWN: Efficient Bound Propagation with Per-neuron Split Constraints for Neural Network Robustness Verification

Huan Zhang, Cho-Jui Hsieh, Shiqi Wang, Suman Jana, Xue Lin, Kaidi Xu, Zico Kolter

arXiv · PDF · Open in the Atlas

Code that ran

4 of 5 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; 1 honoured the contract we drafted, 1 violated it and 1 has no contract verdict; 1 failure 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/2103.06624.json · badge: /api/ran/2103.06624.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
eth-sri/eran canonical (classifier label)
classifier basis: classifier_2025
4 ran · 0 no recorded run
Verified-Intelligence/alpha-beta-CROWN_vnncomp2024 community
Papers with Code archive
0 ran · 1 no recorded run
FunctionStatusWhere it lives
Optimizer Ran
ran on a synthesized input · constructed an object; no result was computed
eth-sri/eran/tf_verify/optimizer.py
code served (permissive licence) · get_code("ed7f6744c1db4625")
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, Optimizer: 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 numpy as np
from functools import reduce

# Define stand-in classes for the execution nodes used in the function
class DeepzonoAffine:
    def __init__(self, matrix, bias, input_names, output_name, output_shape):
        self.matrix = matrix
        self.bias = bias
        self.input_names = input_names
        self.output_name = output_name
        self.output_shape = output_shape

class DeeppolyFCNode:
    def __init__(self, matrix, bias, input_names, output_name, output_shape):
        self.matrix = matrix
        self.bias = bias
        self.input_names = input_names
        self.output_name = output_name
        self.output_shape = output_shape

class DeepzonoConvbias:
    def __init__(self, image_shape, filters, bias, strides, pad_top, pad_left, pad_bottom, pad_right, input_names, output_name, output_shape):
        pass

class DeeppolyConv2dNode:
    def __init__(self, filters, strides, pad_top, pad_left, pad_bottom, pad_right, bias, image_shape, input_names, output_name, output_shape):
        pass

class MockNN:
    def __init__(self):
        self.layertypes = []
        self.weights = []
        self.biases = []
        self.numlayer = 0
        self.numfilters = []
        self.filter_size = []
        self.input_shape = []
        self.strides = []
        self.padding = []
        self.out_shapes = []
        self.filters = []

# Construct valid arguments
operations = ['Relu', 'MatMul', 'Sigmoid']

# Create resources for each operation
# For Relu (neuron-counted): needs 'deepzono' key with last element as shape tuple
resources = [
    {
        'deepzono': (None, None, None, (5, 4, 3, 2)),  # 4D shape for Relu
        'deeppoly': (None, None, None, (5, 4, 3, 2))
    },
    {
        'deepzono': (
            np.random.randn(3, 4),  # matrix
            ['input1'],  # input names
            'output1',  # output name
            (3,)  # output shape
        ),
        'deeppoly': (
            np.random.randn(3, 4),
            ['input1'],
            'output1',
            (3,)
        )
    },
    {
        'deepzono': (None, None, None, (6, 8)),  # 2D shape for Sigmoid
        'deeppoly': (None, None, None, (6, 8))
    }
]

ARGS = {
    'operations': operations,
    'resources': resources
}
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 Optimizer — 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 Optimizer 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
isnetworkfile Ran
ran; our contract draft was wrong, not the code
eth-sri/eran/tf_verify/__main__.py
code served (permissive licence) · get_code("0c0bc63904f889b0")
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 str. 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 os
import argparse
import tempfile

def isnetworkfile(fname):
    _, ext = os.path.splitext(fname)
    if ext not in ['.pyt', '.meta', '.tf', '.onnx', '.pb']:
        raise argparse.ArgumentTypeError('only .pyt, .tf, .onnx, .pb, and .meta formats supported')
    return fname

# Create a temporary file with a valid network extension
with tempfile.NamedTemporaryFile(suffix='.pyt', delete=False) as tmp:
    valid_network_file = tmp.name

ARGS = {
    'fname': valid_network_file
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
parse_input_box Ran
ran, honoured the contract we drafted
eth-sri/eran/tf_verify/__main__.py
code served (permissive licence) · get_code("473beca3aa6713e4")
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 list, ndim 3, 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 numpy as np
import re
import itertools

# Construct a valid text input with interval notation
text_input = """[-1.5, 2.3] [0.0, 1.0]
[-0.5, 0.5] [1.0, 2.0]
[0.1, 0.9]"""

ARGS = {
    "text": text_input
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 3, dtype family float — a list output has no array shape of its own; none is stated
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
str2bool Ran
ran, violated the contract we drafted
eth-sri/eran/tf_verify/__main__.py
code served (permissive licence) · get_code("f017532fc389cbfe")
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 bool, 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 argparse

def str2bool(v):
    if v.lower() in ('yes', 'true', 't', 'y', '1'):
        return True
    elif v.lower() in ('no', 'false', 'f', 'n', '0'):
        return False
    else:
        raise argparse.ArgumentTypeError('Boolean value expected.')

# Setup code to construct valid arguments
ARGS = {
    'v': 'true'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bool, 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
beta_crown_backward_bound Not yet run
unverified (harvested, no recorded run)
Verified-Intelligence/alpha-beta-CROWN_vnncomp2024/auto_LiRPA/beta_crown.py
pointer only (licence: NOASSERTION) · get_code("522597fedc59dae8")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, Verified-Intelligence/alpha-beta-CROWN_vnncomp2024 (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. No invocation recorded.

Also present in other papers' code

Of 5 samples lifted from this paper's linked repositories, 5 carry a function identity and 0 do not. Of the 5 distinct functions with an identity, 2 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 2 later) and 3 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
str2bool 82 other papers
42 dated earlier, 39 later, 1 same year, order not determinable
1510.03055 (2015-10)
A Diversity-Promoting Objective Function for Neural Conversation Models
1703.06490 (2017-03)
Generating Multi-label Discrete Patient Records using Generative Adversarial Networks
1710.10467 (2017-10)
Generalized End-to-End Loss for Speaker Verification
and 79 more not named here
beta_crown_backward_bound
brick of this paper's assembly: entry
1 other paper
1 later
2405.21063 (2024-05)
Neural Network Verification with Branch-and-Bound for General Nonlinearities

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

Bound propagation based incomplete neural network verifiers such as CROWN are very efficient and can significantly accelerate branch-and-bound (BaB) based complete verification of neural networks. However, bound propagation cannot fully handle the neuron split constraints introduced by BaB commonly handled by expensive linear programming (LP) solvers, leading to loose bounds and hurting verification efficiency. In this work, we develop β-CROWN, a new bound propagation based method that can fully encode neuron splits via optimizable parameters β constructed from either primal or dual space. When jointly optimized in intermediate layers, β-CROWN generally produces better bounds than typical LP verifiers with neuron split constraints, while being as efficient and parallelizable as CROWN on GPUs. Applied to complete robustness verification benchmarks, β-CROWN with BaB is up to three orders of magnitude faster than LP-based BaB methods, and is notably faster than all existing approaches while producing lower timeout rates. By terminating BaB early, our method can also be used for efficient incomplete verification. We consistently achieve higher verified accuracy in many settings compared to powerful incomplete verifiers, including those based on convex barrier breaking techniques. Compared to the typically tightest but very costly semidefinite programming (SDP) based incomplete verifiers, we obtain higher verified accuracy with three orders of magnitudes less verification time. Our algorithm empowered the α,β-CROWN (alpha-beta-CROWN) verifier, the winning tool in VNN-COMP 2021. Our code is available at http://PaperCode.cc/BetaCROWN.

For agents

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

get_harvested_code_for_paper("2103.06624")
get_code_for_paper("2103.06624")
have("2103.06624")

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

curl https://syntology.ai/api/ran/2103.06624.json

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

[![Syntology run record](https://syntology.ai/api/ran/2103.06624.svg)](https://syntology.ai/paper/2103.06624)

Connect an agent — have() is free.