SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2106.04530 · 2021

Learning from Multiple Noisy Partial Labelers

arXiv · PDF · Open in the Atlas

Code that ran

7 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; 0 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 6 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/2106.04530.json · badge: /api/ran/2106.04530.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
batsresearch/nplm canonical (classifier label)
classifier basis: classifier_2025
4 ran · 0 no recorded run
batsresearch/yu-aistats22-code canonical (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
FunctionStatusWhere it lives
evaluator Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
batsresearch/nplm/nplm/plf/exec.py
code served (permissive licence) · get_code("2e0222f1fd6d4471")
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 tuple, ndim 2, 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

class MockRule:
    def __init__(self, name, predictions):
        self.name = name
        self.predictions = predictions
    
    def eval(self, gold_labels, class_wise_acc=False, average='macro'):
        # Simple mock evaluation that returns accuracy, precision, recall
        # and optionally class-wise accuracy
        acc = np.random.rand()
        p = np.random.rand()
        r = np.random.rand()
        
        if class_wise_acc:
            # Return class-wise accuracy as a dict or array
            cw_acc = np.random.rand(3)  # 3 classes
            return acc, p, r, cw_acc
        else:
            return acc, p, r

torch.manual_seed(0)

# Create mock rules
rule1 = MockRule("Rule1", torch.randn(2, 4))
rule2 = MockRule("Rule2", torch.randn(2, 4))
rule3 = MockRule("Rule3", torch.randn(2, 4))

rules = [rule1, rule2, rule3]

# Create mock gold labels
gold_labels = torch.tensor([0, 1, 2, 0])

ARGS = {
    'rules': rules,
    'gold_labels': gold_labels,
    'pprint': False,
    'class_wise_acc': False,
    'average': 'macro'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 2, dtype family float — a tuple 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
executor Ran
ran; our contract draft was wrong, not the code
batsresearch/nplm/nplm/plf/exec.py
code served (permissive licence) · get_code("041afdc897dbb18a")
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 tuple. 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

class Rule:
    def __init__(self, name=None, label_maps=None):
        self.name = name
        self.label_maps = label_maps if label_maps is not None else {}
    
    def execute(self, data):
        # Return votes for each data instance
        return np.random.randint(0, 2, size=len(data)).astype(np.uint8)
    
    def set_name(self, name):
        self.name = name

# Create sample rules
rule1 = Rule(name="Rule_1", label_maps={0: [1, 2], 1: [3, 4]})
rule2 = Rule(name="Rule_2", label_maps={0: [0, 1], 1: [2]})
rule3 = Rule(name=None, label_maps={0: [5], 1: [6, 7, 8]})

rules = [rule1, rule2, rule3]

# Create sample data (can be any objects, function just iterates over them)
data = [
    {"feature": 1},
    {"feature": 2}
]

ARGS = {
    "rules": rules,
    "data": data,
    "cpu_parallelism": False,
    "one_indexed": True
}
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 tuple — a tuple output has no array shape of its own; none is 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
extract_inst_label_pair Ran
ran; our contract draft was wrong, not the code
batsresearch/yu-aistats22-code/run_end_model_ag.py
pointer only (licence: NONE) · get_code("c9dac2329ec2721b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, batsresearch/yu-aistats22-code when it was checked. 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.

intercect Ran
ran, violated the contract we drafted
batsresearch/nplm/nplm/lm/labelmodel.py
code served (permissive licence) · get_code("48a7c7193f6372d0")
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 1, dtype family int; 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

torch.manual_seed(0)

# The function intercect takes two lists and returns values from l1 that are also in l2
# It performs a simple intersection operation on two lists

l1 = [1, 2, 3, 4, 5]
l2 = [3, 4, 5, 6, 7]

ARGS = {
    "l1": l1,
    "l2": l2,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family int — 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
load_jkl Ran
ran; our contract draft was wrong, not the code
batsresearch/yu-aistats22-code/run_end_model_ag.py
pointer only (licence: NONE) · get_code("6e31f0275ba89ef3")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, batsresearch/yu-aistats22-code when it was checked. 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.

proc_inst Ran
ran; our contract draft was wrong, not the code
batsresearch/yu-aistats22-code/run_end_model_ag.py
pointer only (licence: NONE) · get_code("fcc62568f492b055")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, batsresearch/yu-aistats22-code when it was checked. 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.

union Ran
ran; our contract draft was wrong, not the code
batsresearch/nplm/nplm/lm/labelmodel.py
code served (permissive licence) · get_code("841a5403e54db840")
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 1, dtype family int; 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

torch.manual_seed(0)

# The function union(l1, l2) takes two lists and returns their set union as a list
# It uses set() on both inputs and the | operator for set union

l1 = [1, 2, 3, 4]
l2 = [3, 4, 5, 6]

ARGS = {
    "l1": l1,
    "l2": l2,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family int — 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

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

Programmatic weak supervision creates models without hand-labeled training data by combining the outputs of heuristic labelers. Existing frameworks make the restrictive assumption that labelers output a single class label. Enabling users to create partial labelers that output subsets of possible class labels would greatly expand the expressivity of programmatic weak supervision. We introduce this capability by defining a probabilistic generative model that can estimate the underlying accuracies of multiple noisy partial labelers without ground truth labels. We show how to scale up learning, for example learning on 100k examples in one minute, a 300x speed up compared to a naive implementation. We also prove that this class of models is generically identifiable up to label swapping under mild conditions. We evaluate our framework on three text classification and six object classification tasks. On text tasks, adding partial labels increases average accuracy by 8.6 percentage points. On image tasks, we show that partial labels allow us to approach some zero-shot object classification problems with programmatic weak supervision by using class attributes as partial labelers. On these tasks, our framework has accuracy comparable to recent embedding-based zero-shot learning methods, while using only pre-trained attribute detectors.

For agents

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

get_harvested_code_for_paper("2106.04530")
get_code_for_paper("2106.04530")
have("2106.04530")

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

curl https://syntology.ai/api/ran/2106.04530.json

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

[![Syntology run record](https://syntology.ai/api/ran/2106.04530.svg)](https://syntology.ai/paper/2106.04530)

Connect an agent — have() is free.