SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1909.13846 · 2019

Universal Approximation with Certified Networks

arXiv · PDF · Open in the Atlas

Code that ran

3 of 3 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, 3 violated it and 0 have no contract verdict; 0 failures of our instrument; 0 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/1909.13846.json · badge: /api/ran/1909.13846.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/UniversalCertificationTheory canonical (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
FunctionStatusWhere it lives
R Ran
ran, violated the contract we drafted
eth-sri/UniversalCertificationTheory/PythonConstruction.py
code served (permissive licence) · get_code("214493b0384505b4")
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 int, ndim 0, dtype family int, 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.

from numpy import hstack, array, linspace, sqrt, ndarray
import numbers

class Interval(object):

    def __init__(self, l, u):
        assert l <= u
        self.l = l
        self.u = u

    def __add__(self, other):
        if isinstance(other, Interval):
            return Interval(self.l + other.l, self.u + other.u)
        elif isinstance(other, numbers.Number):
            return Interval(self.l + other, self.u + other)
        else:
            assert False

    def __radd__(self, other):
        return self + other

    def __neg__(self):
        return Interval(-self.u, -self.l)

    def __sub__(self, other):
        return self + -other

    def __rsub__(self, other):
        return -self + other

    def __mul__(self, other):
        assert not isinstance(other, Interval)
        if other >= 0:
            return Interval(self.l * other, self.u * other)
        elif isinstance(other, numbers.Number):
            return -(-other * self)
        else:
            assert False

    def __rmul__(self, other):
        return self * other

    def __repr__(self):
        return '[' + str(self.l) + ',' + str(self.u) + ']'

def R(x):
    if isinstance(x, numbers.Number):
        return max(x, 0)
    elif isinstance(x, Interval):
        return Interval(R(x.l), R(x.u))
    if isinstance(x, ndarray):
        return array(list(map(R, x)))
    else:
        assert False

# Test cases for R function
test_number = 5
test_negative_number = -3
test_interval = Interval(-2, 3)
test_array = array([-1, 0, 2, -5])

ARGS = {
    'x': test_number
}
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 int, ndim 0, dtype family int, 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
R1 Ran
ran, violated the contract we drafted
eth-sri/UniversalCertificationTheory/PythonConstruction.py
code served (permissive licence) · get_code("a490c2edc62865a7")
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 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 numbers
from numpy import hstack, array, linspace, sqrt, ndarray

class Interval(object):

    def __init__(self, l, u):
        assert l <= u
        self.l = l
        self.u = u

    def __add__(self, other):
        if isinstance(other, Interval):
            return Interval(self.l + other.l, self.u + other.u)
        elif isinstance(other, numbers.Number):
            return Interval(self.l + other, self.u + other)
        else:
            assert False

    def __radd__(self, other):
        return self + other

    def __neg__(self):
        return Interval(-self.u, -self.l)

    def __sub__(self, other):
        return self + -other

    def __rsub__(self, other):
        return -self + other

    def __mul__(self, other):
        assert not isinstance(other, Interval)
        if other >= 0:
            return Interval(self.l * other, self.u * other)
        elif isinstance(other, numbers.Number):
            return -(-other * self)
        else:
            assert False

    def __rmul__(self, other):
        return self * other

    def __repr__(self):
        return '[' + str(self.l) + ',' + str(self.u) + ']'

def R(x):
    if isinstance(x, numbers.Number):
        return max(x, 0)
    elif isinstance(x, Interval):
        return Interval(R(x.l), R(x.u))
    if isinstance(x, ndarray):
        return array(list(map(R, x)))
    else:
        assert False

def R1(x):
    return 1.0 - R(1.0 - x)

# Test with different input types
test_number = 0.5
test_interval = Interval(0.2, 0.8)
test_array = array([0.1, 0.5, 0.9, 0.3])

ARGS = {
    'x': test_number
}
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 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
nmin2 Ran
ran, violated the contract we drafted
eth-sri/UniversalCertificationTheory/PythonConstruction.py
code served (permissive licence) · get_code("f5cde5b43d9ab5b7")
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 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.

from numpy import hstack, array, linspace, sqrt, ndarray
import numbers

class Interval(object):

    def __init__(self, l, u):
        assert l <= u
        self.l = l
        self.u = u

    def __add__(self, other):
        if isinstance(other, Interval):
            return Interval(self.l + other.l, self.u + other.u)
        elif isinstance(other, numbers.Number):
            return Interval(self.l + other, self.u + other)
        else:
            assert False

    def __radd__(self, other):
        return self + other

    def __neg__(self):
        return Interval(-self.u, -self.l)

    def __sub__(self, other):
        return self + -other

    def __rsub__(self, other):
        return -self + other

    def __mul__(self, other):
        assert not isinstance(other, Interval)
        if other >= 0:
            return Interval(self.l * other, self.u * other)
        elif isinstance(other, numbers.Number):
            return -(-other * self)
        else:
            assert False

    def __rmul__(self, other):
        return self * other

    def __repr__(self):
        return '[' + str(self.l) + ',' + str(self.u) + ']'

def R(x):
    if isinstance(x, numbers.Number):
        return max(x, 0)
    elif isinstance(x, Interval):
        return Interval(R(x.l), R(x.u))
    if isinstance(x, ndarray):
        return array(list(map(R, x)))
    else:
        assert False

def nmin2(x, y):
    return 0.5 * (R(x + y) - R(-x - y) - R(x - y) - R(-x + y))

# Test with different input types
x_number = 2.5
y_number = 1.5

x_interval = Interval(1.0, 3.0)
y_interval = Interval(0.5, 2.5)

x_array = array([1.0, 2.0, 3.0, 4.0])
y_array = array([0.5, 1.5, 2.5, 3.5])

ARGS = {
    'x': x_number,
    'y': y_number
}
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 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

Also present in other papers' code

Of 3 samples lifted from this paper's linked repositories, 3 carry a function identity and 0 do not. Of the 3 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 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.

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

Training neural networks to be certifiably robust is critical to ensure their safety against adversarial attacks. However, it is currently very difficult to train a neural network that is both accurate and certifiably robust. In this work we take a step towards addressing this challenge. We prove that for every continuous function $f$, there exists a network $n$ such that: (i) $n$ approximates $f$ arbitrarily close, and (ii) simple interval bound propagation of a region $B$ through $n$ yields a result that is arbitrarily close to the optimal output of $f$ on $B$. Our result can be seen as a Universal Approximation Theorem for interval-certified ReLU networks. To the best of our knowledge, this is the first work to prove the existence of accurate, interval-certified networks.

For agents

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

get_harvested_code_for_paper("1909.13846")
get_code_for_paper("1909.13846")
have("1909.13846")

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

curl https://syntology.ai/api/ran/1909.13846.json

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

[![Syntology run record](https://syntology.ai/api/ran/1909.13846.svg)](https://syntology.ai/paper/1909.13846)

Connect an agent — have() is free.