SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2203.04940 · 2022

Data-Efficient Structured Pruning via Submodular Optimization

arXiv · PDF · Open in the Atlas

Code that ran

11 of 11 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 10 have no contract verdict; 1 failure 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/2203.04940.json · badge: /api/ran/2203.04940.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
marwash25/subpruning canonical (classifier label)
classifier basis not recorded on the link
11 ran · 0 no recorded run
FunctionStatusWhere it lives
accuracy Ran
ran; our contract draft was wrong, not the code
marwash25/subpruning/shrinkbench/metrics/accuracy.py
code served (permissive licence) · get_code("8ac1d792ff6911f1")
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 ndarray, ndim 1, dtype family float, shape (2,) 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 torch.nn as nn
import numpy as np

torch.manual_seed(0)

# Create a minimal model that takes input and outputs logits
class MinimalModel(nn.Module):
    def __init__(self, input_dim, num_classes):
        super().__init__()
        self.linear = nn.Linear(input_dim, num_classes)
    
    def forward(self, x):
        return self.linear(x)

# Create model with small dimensions
model = MinimalModel(input_dim=4, num_classes=3)

# Create a minimal dataloader as a list of (input, target) tuples
# Batch size 2, input dimension 4, num_classes 3
batch1_input = torch.randn(2, 4)
batch1_target = torch.tensor([0, 2])

batch2_input = torch.randn(2, 4)
batch2_target = torch.tensor([1, 0])

dataloader = [(batch1_input, batch1_target), (batch2_input, batch2_target)]

# Add dataset attribute for len() call in accuracy function
class SimpleDataset:
    def __init__(self, size):
        self.size = size
    def __len__(self):
        return self.size

dataloader_with_dataset = dataloader.copy()
# Wrap dataloader with a dataset attribute
class DataLoaderWrapper:
    def __init__(self, data):
        self.data = data
        self.dataset = SimpleDataset(sum(len(batch[0]) for batch in data))
    
    def __iter__(self):
        return iter(self.data)

dataloader_wrapped = DataLoaderWrapper(dataloader)

ARGS = {
    "model": model,
    "dataloader": dataloader_wrapped,
    "topk": (1, 2)
}
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 ndarray, ndim 1, dtype family float, shape (2,) — shape (2,) 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
conv2d_flops Ran
ran on a synthesized input
marwash25/subpruning/shrinkbench/metrics/abstract_flops.py
code served (permissive licence) · get_code("ea28692522bc6071")
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 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.

import numpy as np

# Set up arguments for conv2d_flops function
ARGS = {
    'in_channels': 10,
    'out_channels': 15,
    'input_shape': (4, 5),
    'kernel_shape': (3, 2),
    'padding': 'same',
    'strides': (1, 2),
    'dilation': 1
}
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 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
correct Ran
ran on a synthesized input
marwash25/subpruning/shrinkbench/metrics/accuracy.py
code served (permissive licence) · get_code("b8471510e1111074")
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 list, ndim 1, 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

torch.manual_seed(0)

# Create small tensors for the function
batch_size = 2
num_classes = 4

# output: shape (batch_size, num_classes) - model predictions/logits
output = torch.randn(batch_size, num_classes)

# target: shape (batch_size,) - ground truth class indices
target = torch.tensor([1, 3], dtype=torch.long)

# topk: tuple of k values to check
topk = (1, 2)

ARGS = {
    "output": output,
    "target": target,
    "topk": topk,
}
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 list, ndim 1, 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
dataset_path Ran
ran on a synthesized input
marwash25/subpruning/shrinkbench/datasets/datasets.py
code served (permissive licence) · get_code("1b4a7337eb8e904c")
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 PosixPath. 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 pathlib
import os
import tempfile

# Create a temporary directory structure to simulate dataset locations
temp_dir = tempfile.TemporaryDirectory()
temp_path = pathlib.Path(temp_dir.name)

# Create a fake dataset directory
dataset_dir = temp_path / "MNIST"
dataset_dir.mkdir(parents=True, exist_ok=True)

# Create a fake file in the dataset directory to make it "exist"
(dataset_dir / "data.txt").touch()

ARGS = {
    "dataset": "MNIST",
    "path": str(temp_path)
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
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 PosixPath — a PosixPath output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a PosixPath 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
dense_flops Ran
ran on a synthesized input
marwash25/subpruning/shrinkbench/metrics/abstract_flops.py
code served (permissive licence) · get_code("6339375cb9607ec8")
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 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.

import torch
import numpy as np

torch.manual_seed(0)

# The function dense_flops takes two integer parameters: in_neurons and out_neurons
# It simply multiplies them together to compute FLOPs for a dense layer

ARGS = {
    "in_neurons": 128,
    "out_neurons": 256,
}
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 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
get_classifier_module Ran
ran on a synthesized input
marwash25/subpruning/shrinkbench/models/head.py
code served (permissive licence) · get_code("e5e70e38817af64f")
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 Linear. 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 torch
import torch.nn as nn
import torchvision.models

torch.manual_seed(0)

# Create a ResNet model (one of the MODELS_WITH_FC)
model = torchvision.models.resnet18(pretrained=False)

ARGS = {
    "model": model
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torchvision — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Linear — a Linear output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Linear 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
mark_classifier Ran
ran on a synthesized input
marwash25/subpruning/shrinkbench/models/head.py
code served (permissive licence) · get_code("d40ad3113b27ca66")
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 ResNet. 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 torch
import torch.nn as nn
import torchvision.models

torch.manual_seed(0)

# Create a ResNet model (one of MODELS_WITH_FC)
model = torchvision.models.resnet18(pretrained=False)

# Verify the model has the expected 'fc' attribute
assert hasattr(model, 'fc'), "ResNet should have 'fc' attribute"

ARGS = {
    'model': model
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torchvision — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ResNet — a ResNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ResNet 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
reduce_linear_layer Ran
ran on a synthesized input
marwash25/subpruning/shrinkbench/models/head.py
code served (permissive licence) · get_code("3a12702023a2fdf5")
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 Linear. 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 torch
from torch import nn

torch.manual_seed(0)

# Create a linear layer with in_features=4, out_features=10 (original number of classes)
layer = nn.Linear(in_features=4, out_features=10)

# Define the number of classes to reduce to
n_classes = 3

# Define whether to keep weights
keep_weights = True

ARGS = {
    "layer": layer,
    "n_classes": n_classes,
    "keep_weights": keep_weights,
}
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 Linear — a Linear output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Linear 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
resnet_factory Ran
ran on a synthesized input
marwash25/subpruning/shrinkbench/models/cifar_resnet.py
code served (permissive licence) · get_code("ce91586f8f908872")
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 function. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 torch
import torch.nn as nn

torch.manual_seed(0)

def weights_path(weight_file):
    """Stand-in for weights_path function that the factory calls"""
    return weight_file

# Create a minimal valid weight file structure
# The function expects torch.load() to return a dict with 'state_dict' key
weight_dict = {
    'state_dict': {
        'conv1.weight': torch.randn(16, 3, 3, 3),
        'bn1.weight': torch.randn(16),
        'bn1.bias': torch.randn(16),
        'bn1.running_mean': torch.randn(16),
        'bn1.running_var': torch.randn(16),
        'layer1.0.conv1.weight': torch.randn(16, 16, 3, 3),
        'layer1.0.bn1.weight': torch.randn(16),
        'layer1.0.bn1.bias': torch.randn(16),
        'layer1.0.bn1.running_mean': torch.randn(16),
        'layer1.0.bn1.running_var': torch.randn(16),
        'layer1.0.conv2.weight': torch.randn(16, 16, 3, 3),
        'layer1.0.bn2.weight': torch.randn(16),
        'layer1.0.bn2.bias': torch.randn(16),
        'layer1.0.bn2.running_mean': torch.randn(16),
        'layer1.0.bn2.running_var': torch.randn(16),
        'layer2.0.conv1.weight': torch.randn(32, 16, 3, 3),
        'layer2.0.bn1.weight': torch.randn(32),
        'layer2.0.bn1.bias': torch.randn(32),
        'layer2.0.bn1.running_mean': torch.randn(32),
        'layer2.0.bn1.running_var': torch.randn(32),
        'layer2.0.conv2.weight': torch.randn(32, 32, 3, 3),
        'layer2.0.bn2.weight': torch.randn(32),
        'layer2.0.bn2.bias': torch.randn(32),
        'layer2.0.bn2.running_mean': torch.randn(32),
        'layer2.0.bn2.running_var': torch.randn(32),
        'layer3.0.conv1.weight': torch.randn(64, 32, 3, 3),
        'layer3.0.bn1.weight': torch.randn(64),
        'layer3.0.bn1.bias': torch.randn(64),
        'layer3.0.bn1.running_mean': torch.randn(64),
        'layer3.0.bn1.running_var': torch.randn(64),
        'layer3.0.conv2.weight': torch.randn(64, 64, 3, 3),
        'layer3.0.bn2.weight': torch.randn(64),
        'layer3.0.bn2.bias': torch.randn(64),
        'layer3.0.bn2.running_mean': torch.randn(64),
        'layer3.0.bn2.running_var': torch.randn(64),
        'linear.weight': torch.randn(10, 64),
        'linear.bias': torch.randn(10),
    }
}

# Mock torch.load to return our weight dict
original_load = torch.load
def mock_load(path, map_location=None):
    return weight_dict
torch.load = mock_load

ARGS = {
    'filters': [1, 1, 1],
    'num_classes': 10,
    'weight_file': 'mock_weights.pt'
}
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 function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
vgg_bn_drop Ran
ran on a synthesized input
marwash25/subpruning/shrinkbench/models/cifar_vgg.py
code served (permissive licence) · get_code("7317a582e094448e")
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 VGGBnDrop. 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 torch
import torch.nn as nn
import math
import torch.nn.init as init

torch.manual_seed(0)

# Mock the weights_path function since it's called in vgg_bn_drop
def weights_path(name):
    return f"/tmp/{name}"

# Mock torch.load to avoid actual file I/O
original_torch_load = torch.load
def mock_torch_load(path):
    # Return a mock state dict that matches VGGBnDrop structure
    model = VGGBnDrop(num_classes=10)
    return model.state_dict()

torch.load = mock_torch_load

ARGS = {
    'pretrained': False
}
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 VGGBnDrop — a VGGBnDrop output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a VGGBnDrop 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
vgg_bn_drop_100 Ran
ran on a synthesized input
marwash25/subpruning/shrinkbench/models/cifar_vgg.py
code served (permissive licence) · get_code("19ff4803e41e3e8e")
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 VGGBnDrop. 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 torch
import torch.nn as nn
import math

torch.manual_seed(0)

# Create a mock weights_path function and a mock weights file
def weights_path(filename):
    return filename

# Create a mock torch.load that returns a valid state dict
original_torch_load = torch.load
def mock_torch_load(path):
    if path == 'vgg_bn_drop_100.pt':
        model = VGGBnDrop(num_classes=100)
        return model.state_dict()
    return original_torch_load(path)

torch.load = mock_torch_load

ARGS = {
    'pretrained': False
}
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 VGGBnDrop — a VGGBnDrop output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a VGGBnDrop 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

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 distinct functions with an identity, 7 are also present in other papers whose code we hold (7 with at least one paper dated earlier, 0 the same month, 0 later) and 4 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
accuracy 1 other paper
1 dated earlier
2003.03033 (2020-03)
What is the State of Neural Network Pruning?
conv2d_flops 1 other paper
1 dated earlier
2003.03033 (2020-03)
What is the State of Neural Network Pruning?
dataset_path 1 other paper
1 dated earlier
2003.03033 (2020-03)
What is the State of Neural Network Pruning?
dense_flops 1 other paper
1 dated earlier
2003.03033 (2020-03)
What is the State of Neural Network Pruning?
get_classifier_module 1 other paper
1 dated earlier
2003.03033 (2020-03)
What is the State of Neural Network Pruning?
mark_classifier 1 other paper
1 dated earlier
2003.03033 (2020-03)
What is the State of Neural Network Pruning?
reduce_linear_layer 1 other paper
1 dated earlier
2003.03033 (2020-03)
What is the State of Neural Network Pruning?

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

Structured pruning is an effective approach for compressing large pre-trained neural networks without significantly affecting their performance. However, most current structured pruning methods do not provide any performance guarantees, and often require fine-tuning, which makes them inapplicable in the limited-data regime. We propose a principled data-efficient structured pruning method based on submodular optimization. In particular, for a given layer, we select neurons/channels to prune and corresponding new weights for the next layer, that minimize the change in the next layer's input induced by pruning. We show that this selection problem is a weakly submodular maximization problem, thus it can be provably approximated using an efficient greedy algorithm. Our method is guaranteed to have an exponentially decreasing error between the original model and the pruned model outputs w.r.t the pruned size, under reasonable assumptions. It is also one of the few methods in the literature that uses only a limited-number of training data and no labels. Our experimental results demonstrate that our method outperforms state-of-the-art methods in the limited-data regime.

For agents

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

get_harvested_code_for_paper("2203.04940")
get_code_for_paper("2203.04940")
have("2203.04940")

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

curl https://syntology.ai/api/ran/2203.04940.json

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

[![Syntology run record](https://syntology.ai/api/ran/2203.04940.svg)](https://syntology.ai/paper/2203.04940)

Connect an agent — have() is free.