SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2110.07607 · NeurIPS · 2021

HumBugDB: A Large-scale Acoustic Mosquito Dataset

Ivan Kiskin, Adam Cobb, Lawrence Wang, Yunpeng Li, Davide Zilli, Marianne Sinka, Kathy Willis, Stephen Roberts, Waqas Rafique, ¶ Mind, Benjamin Gutteridge, Rinita Dam, and 5 more

arXiv · PDF · Open in the Atlas

Code that ran

9 of 12 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 7 have no contract verdict; 2 failures of our instrument; 3 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/2110.07607.json · badge: /api/ran/2110.07607.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
HumBug-Mosquito/MozzBNN canonical (classifier label)
classifier basis not recorded on the link
5 ran · 1 no recorded run
HumBug-Mosquito/HumBugDB canonical (classifier label)
classifier basis not recorded on the link
4 ran · 2 no recorded run
FunctionStatusWhere it lives
active_BALD Ran
ran on a synthesized input
HumBug-Mosquito/HumBugDB/lib/evaluate.py
code served (permissive licence) · get_code("c4e9ed0a19b6ac98")
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 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

# Construct arguments for active_BALD function
# out: shape (num_samples, num_data, num_classes) - log probabilities from multiple forward passes
# X: shape (num_data, num_features) - input data
# n_classes: int - number of classes

num_forward_passes = 3  # out.shape[0]
num_data = 5            # X.shape[0]
num_classes = 4         # n_classes
num_features = 2        # X.shape[1]

# Create log probabilities output from multiple forward passes
# Shape: (num_forward_passes, num_data, num_classes)
out = np.random.randn(num_forward_passes, num_data, num_classes)

# Create input data
# Shape: (num_data, num_features)
X = np.random.randn(num_data, num_features)

# Number of classes
n_classes = num_classes

ARGS = {
    "out": out,
    "X": X,
    "n_classes": n_classes
}
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 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
conv1x1 Ran
ran; our contract draft was wrong, not the code
HumBug-Mosquito/HumBugDB/lib/PyTorch/ResNetDropoutSource.py
pointer only (licence: MIT); the stored copy get_code serves, from medaitech/scrd4ad, is pointer only (licence: NONE) · get_code("2a80220dabcb742a")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from medaitech/scrd4ad, and GitHub detected no licence in that repository 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.

conv3x3 Ran
ran; our contract draft was wrong, not the code
HumBug-Mosquito/HumBugDB/lib/PyTorch/ResNetDropoutSource.py
pointer only (licence: MIT); the stored copy get_code serves, from medaitech/scrd4ad, is pointer only (licence: NONE) · get_code("600ff2c45e0de056")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from medaitech/scrd4ad, and GitHub detected no licence in that repository 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.

evaluate_model Ran
ran on a synthesized input
HumBug-Mosquito/HumBugDB/lib/Keras/runKeras.py
code served (permissive licence) · get_code("fffbb3c08c0354be")
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. 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

torch.manual_seed(0)

class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def predict(self, X):
        return self.linear(X)

model = SimpleModel()

X_test = torch.randn(2, 4)
y_test = torch.randn(2, 3)
n_samples = 3

ARGS = {
    "model": model,
    "X_test": X_test,
    "y_test": y_test,
    "n_samples": n_samples
}
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 — a list 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
plot_mozz_MI Ran
ran on a synthesized input
HumBug-Mosquito/MozzBNN/Code/lib/util_dashboard.py
code served (permissive licence) · get_code("acf9ba7f8d4247c8")
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 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 numpy as np
import tempfile
import os

# Create temporary directory for output
temp_dir = tempfile.mkdtemp()

# X_CNN: list of CNN feature inputs (spectrograms)
# The function does: np.hstack(X_CNN).squeeze().T[:, pos_pred_idx]
# So X_CNN should be a list of arrays that can be hstacked
# Each element should be a 2D array (frequency x time)
X_CNN = [
    np.random.randn(128, 50),  # 128 frequency bins, 50 time steps
    np.random.randn(128, 50),  # another spectrogram chunk
]

# y: 1D array of prediction probabilities (0 to 1)
# The function filters by p_threshold and accesses y[pos_pred_idx]
y = np.array([0.95, 0.05, 0.88, 0.12, 0.91, 0.03, 0.87, 0.15])

# MI: 1D array of mutual information values (uncertainty)
# Must be same length as y
MI = np.array([0.1, 0.8, 0.15, 0.75, 0.12, 0.85, 0.18, 0.72])

# p_threshold: probability threshold for filtering detections
p_threshold = 0.5

# root_out: output directory
root_out = temp_dir

# filename: base filename for output (without extension)
filename = "mozz_detection"

# out_format: file format extension
out_format = '.png'

ARGS = {
    'X_CNN': X_CNN,
    'y': y,
    'MI': MI,
    'p_threshold': p_threshold,
    'root_out': root_out,
    'filename': filename,
    'out_format': out_format,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, numpy — 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
remove_label_bug Ran
ran on a synthesized input
HumBug-Mosquito/MozzBNN/Code/lib/util.py
code served (permissive licence) · get_code("89f0bf9e7cdce5c9")
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 DataFrame. 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 pandas as pd
import numpy as np

# Create a small dataframe with audio file labels
df = pd.DataFrame({
    'fine_start_time': [0.0, 1.0, 2.0, 3.0, 4.0],
    'fine_end_time': [0.5, 1.5, 2.0, 4.0, 5.0],  # Note: index 2 has duration 0, index 3 will be kept
    'path': [
        '/path/to/audio1.wav',
        '/path/to/audio2.wav',
        '/path/to/audio3.wav',
        '/path/to/audio4.wav',
        '/path/to/audio5.wav'
    ]
})

ARGS = {
    'df': df,
    'verbose': 0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 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
speech_stats Ran
ran on a synthesized input
HumBug-Mosquito/MozzBNN/Code/lib/VAD-Humbug/webrtc_utils.py
code served (permissive licence) · get_code("5a9320a5331e48be")
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 tuple, 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 numpy as np

# Create small sample data for speech timestamps
# speech_pred_labels: predicted speech segments (start, end) in seconds
speech_pred_labels = np.array([
    [0.5, 2.0],
    [3.0, 5.5],
    [7.0, 8.5]
], dtype=np.float32)

# speech_labels: true speech segments (start, end) in seconds
speech_labels = np.array([
    [0.3, 2.2],
    [3.1, 5.3],
    [6.9, 9.0]
], dtype=np.float32)

# file_length: total duration of audio file in seconds
file_length = 10.0

ARGS = {
    "speech_pred_labels": speech_pred_labels,
    "speech_labels": speech_labels,
    "file_length": file_length
}
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 tuple, ndim 1, 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
speech_stripper Ran
ran on a synthesized input
HumBug-Mosquito/MozzBNN/Code/lib/VAD-Humbug/webrtc_utils.py
code served (permissive licence) · get_code("ccfe9f740451af06")
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 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

# Create a simple audio time series (1D numpy array)
# Let's create 2 seconds of audio at 8000 Hz
sr = 8000
duration = 2.0
t_series = np.random.randn(int(sr * duration)).astype(np.float32)

# Create labels: n*2 numpy array with speech regions in seconds
# Two speech regions: [0.2-0.5 seconds] and [1.0-1.5 seconds]
labels = np.array([
    [0.2, 0.5],
    [1.0, 1.5]
], dtype=np.float32)

# Define buffer times in seconds
Lbuffer = 0.1
Rbuffer = 0.1

ARGS = {
    "t_series": t_series,
    "labels": labels,
    "sr": sr,
    "Lbuffer": Lbuffer,
    "Rbuffer": Rbuffer
}
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 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
vad_collector Ran
ran on a synthesized input
HumBug-Mosquito/MozzBNN/Code/lib/VAD-Humbug/webrtc_vad.py
code served (permissive licence) · get_code("62c58c40551afebf")
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 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 collections
import numpy as np

class MockFrame:
    def __init__(self, bytes_data, timestamp):
        self.bytes = bytes_data
        self.timestamp = timestamp

class MockVAD:
    def __init__(self):
        self.call_count = 0
    
    def is_speech(self, audio_bytes, sample_rate):
        self.call_count += 1
        return self.call_count % 2 == 0

sample_rate = 16000
frame_duration_ms = 30
padding_duration_ms = 300

vad = MockVAD()

frames = [
    MockFrame(b'\x00\x01\x02\x03', 0.0),
    MockFrame(b'\x04\x05\x06\x07', 0.03),
    MockFrame(b'\x08\x09\x0a\x0b', 0.06),
    MockFrame(b'\x0c\x0d\x0e\x0f', 0.09),
    MockFrame(b'\x10\x11\x12\x13', 0.12),
    MockFrame(b'\x14\x15\x16\x17', 0.15),
    MockFrame(b'\x18\x19\x1a\x1b', 0.18),
    MockFrame(b'\x1c\x1d\x1e\x1f', 0.21),
    MockFrame(b'\x20\x21\x22\x23', 0.24),
    MockFrame(b'\x24\x25\x26\x27', 0.27),
    MockFrame(b'\x28\x29\x2a\x2b', 0.30),
]

ARGS = {
    'sample_rate': sample_rate,
    'frame_duration_ms': frame_duration_ms,
    'padding_duration_ms': padding_duration_ms,
    'vad': vad,
    'frames': frames,
}
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 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
read_wave Not yet run
unverified (harvested, no recorded run)
HumBug-Mosquito/MozzBNN/Code/lib/VAD-Humbug/webrtc_vad.py
pointer only (licence: MIT); the stored copy get_code serves, from wiseman/py-webrtcvad, is pointer only (licence: NOASSERTION) · get_code("2c4ba80fd768dc88")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from wiseman/py-webrtcvad, and GitHub could not classify the licence of that repository (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.

resnet18 Not yet run
unverified (harvested, no recorded run)
HumBug-Mosquito/HumBugDB/lib/PyTorch/ResNetSource.py
code served (permissive licence) · get_code("bc99cd552cb919bf")
How it ran

no invocation recorded.

test_model Not yet run
unverified (harvested, no recorded run)
HumBug-Mosquito/HumBugDB/lib/PyTorch/runTorch.py
code served (permissive licence) · get_code("34e93ffee7856b18")
How it ran

no invocation recorded.

Also present in other papers' code

Of 12 samples lifted from this paper's linked repositories, 12 carry a function identity and 0 do not. Of the 12 distinct functions with an identity, 4 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 3 the same month, 4 later) and 8 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
conv1x1 112 other papers
14 dated earlier, 2 the same month, 96 later · 6 only as an assembly brick
1512.03385 (2015-12)
Deep Residual Learning for Image Recognition
1806.01531 (2018-06)
Deep Mixture of Experts via Shallow Embedding
1905.12794 (2019-05)
Fashion IQ: A New Dataset Towards Retrieving Images by Natural Language Feedback
and 109 more not named here
conv3x3 111 other papers
12 dated earlier, 2 the same month, 97 later · 6 only as an assembly brick
1512.03385 (2015-12)
Deep Residual Learning for Image Recognition
1806.01531 (2018-06)
Deep Mixture of Experts via Shallow Embedding
1905.12794 (2019-05)
Fashion IQ: A New Dataset Towards Retrieving Images by Natural Language Feedback
and 108 more not named here
resnet18 36 other papers
6 dated earlier, 2 the same month, 28 later
1905.12794 (2019-05)
Fashion IQ: A New Dataset Towards Retrieving Images by Natural Language Feedback
2002.09564 (2020-02)
Towards Robust and Reproducible Active Learning Using Neural Networks
2003.04297 (2020-03)
Improved Baselines with Momentum Contrastive Learning
and 33 more not named here
read_wave 2 other papers
1 dated earlier, 1 later
1710.10467 (2017-10)
Generalized End-to-End Loss for Speaker Verification
2507.00472 (2025-07)
ARIG: Autoregressive Interactive Head Generation for Real-time Conversations

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

This paper presents the first large-scale multi-species dataset of acoustic recordings of mosquitoes tracked continuously in free flight. We present 20 hours of audio recordings that we have expertly labelled and tagged precisely in time. Significantly, 18 hours of recordings contain annotations from 36 different species. Mosquitoes are well-known carriers of diseases such as malaria, dengue and yellow fever. Collecting this dataset is motivated by the need to assist applications which utilise mosquito acoustics to conduct surveys to help predict outbreaks and inform intervention policy. The task of detecting mosquitoes from the sound of their wingbeats is challenging due to the difficulty in collecting recordings from realistic scenarios. To address this, as part of the HumBug project, we conducted global experiments to record mosquitoes ranging from those bred in culture cages to mosquitoes captured in the wild. Consequently, the audio recordings vary in signal-to-noise ratio and contain a broad range of indoor and outdoor background environments from Tanzania, Thailand, Kenya, the USA and the UK. In this paper we describe in detail how we collected, labelled and curated the data. The data is provided from a PostgreSQL database, which contains important metadata such as the capture method, age, feeding status and gender of the mosquitoes. Additionally, we provide code to extract features and train Bayesian convolutional neural networks for two key tasks: the identification of mosquitoes from their corresponding background environments, and the classification of detected mosquitoes into species. Our extensive dataset is both challenging to machine learning researchers focusing on acoustic identification, and critical to entomologists, geo-spatial modellers and other domain experts to understand mosquito behaviour, model their distribution, and manage the threat they pose to humans.

For agents

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

get_harvested_code_for_paper("2110.07607")
get_code_for_paper("2110.07607")
have("2110.07607")

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

curl https://syntology.ai/api/ran/2110.07607.json

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

[![Syntology run record](https://syntology.ai/api/ran/2110.07607.svg)](https://syntology.ai/paper/2110.07607)

Connect an agent — have() is free.