SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1608.00507 · 2016

Top-down Neural Attention by Excitation Backprop

arXiv · PDF · Open in the Atlas

Code that ran

3 of 5 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 3 have no contract verdict; 0 failures of our instrument; 2 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/1608.00507.json · badge: /api/ran/1608.00507.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
lassiraa/weighting-game community (classifier label)
classifier basis not recorded on the link
3 ran · 2 no recorded run
FunctionStatusWhere it lives
calculate_mass_within Ran
ran on a synthesized input
lassiraa/weighting-game/utils.py
code served (permissive licence) · get_code("7b8b37d5389eed41")
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 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 torch

torch.manual_seed(0)

# Create a small saliency map tensor
# Shape: (2, 4) - batch of 2, 4 dimensions
saliency_map = torch.randn(2, 4).abs()  # Use abs to ensure positive values

# Create a class mask tensor with same shape
# Values should be 0 or 1 to act as a mask
class_mask = torch.randint(0, 2, (2, 4)).float()

ARGS = {
    "saliency_map": saliency_map,
    "class_mask": class_mask,
}
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 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
fine_tune Ran
ran on a synthesized input
lassiraa/weighting-game/fine_tune.py
code served (permissive licence) · get_code("ec81f6de11ab0b12")
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 SimpleModel. 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 torch.optim as optim
from torch.optim.lr_scheduler import ExponentialLR
from torch.utils.data import DataLoader
import tempfile
import os

torch.manual_seed(0)

# Create a minimal model that outputs logits for binary classification
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def forward(self, x):
        return self.linear(x)

# Create tiny datasets: batch of 2, input dim 4, output dim 3 (for multi-label)
train_data = [
    (torch.randn(2, 4), torch.randint(0, 2, (2, 3)).float()),
    (torch.randn(2, 4), torch.randint(0, 2, (2, 3)).float()),
]

val_data = [
    (torch.randn(2, 4), torch.randint(0, 2, (2, 3)).float()),
]

coco_loader_train = train_data
coco_loader_val = val_data

model = SimpleModel()
optimizer = optim.Adam(model.parameters(), lr=0.001)
scheduler = ExponentialLR(optimizer, gamma=0.9)

params = {
    'epochs': 2,
}

model_name = 'test_model'
device = torch.device('cpu')
checkpoint_dir = tempfile.gettempdir() + '/'

ARGS = {
    'model': model,
    'optimizer': optimizer,
    'scheduler': scheduler,
    'coco_loader_train': coco_loader_train,
    'coco_loader_val': coco_loader_val,
    'params': params,
    'model_name': model_name,
    'device': device,
    'checkpoint_dir': checkpoint_dir,
}
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 SimpleModel — a SimpleModel output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a SimpleModel 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
scale_image Ran
ran on a synthesized input
lassiraa/weighting-game/utils.py
code served (permissive licence) · get_code("f7a8fd5143f99c4a")
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 Tensor, ndim 2, dtype family float, shape (2, 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 torch
import numpy as np

torch.manual_seed(0)

# The function expects:
# - img: Any object that supports .min(), .max(), and arithmetic operations
# - max: int with default value 1

# Create a small tensor to serve as the image
img = torch.randn(2, 4)

ARGS = {
    "img": img,
    "max": 1,
}
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 Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 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
get_model_to_fine_tune Not yet run
unverified (harvested, no recorded run)
lassiraa/weighting-game/fine_tune.py
code served (permissive licence) · get_code("9202adee113ee2da")
How it ran

no invocation recorded.

load_model_with_target_layers Not yet run
unverified (harvested, no recorded run)
lassiraa/weighting-game/utils.py
code served (permissive licence) · get_code("5a183dcbd14b1938")
How it ran

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, 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 5 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

We aim to model the top-down attention of a Convolutional Neural Network (CNN) classifier for generating task-specific attention maps. Inspired by a top-down human visual attention model, we propose a new backpropagation scheme, called Excitation Backprop, to pass along top-down signals downwards in the network hierarchy via a probabilistic Winner-Take-All process. Furthermore, we introduce the concept of contrastive attention to make the top-down attention maps more discriminative. In experiments, we demonstrate the accuracy and generalizability of our method in weakly supervised localization tasks on the MS COCO, PASCAL VOC07 and ImageNet datasets. The usefulness of our method is further validated in the text-to-region association task. On the Flickr30k Entities dataset, we achieve promising performance in phrase localization by leveraging the top-down attention of a CNN model that has been trained on weakly labeled web images.

For agents

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

get_harvested_code_for_paper("1608.00507")
get_code_for_paper("1608.00507")
have("1608.00507")

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

curl https://syntology.ai/api/ran/1608.00507.json

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

[![Syntology run record](https://syntology.ai/api/ran/1608.00507.svg)](https://syntology.ai/paper/1608.00507)

Connect an agent — have() is free.