SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1912.00682 · 2019

GeoTrackNet-A Maritime Anomaly Detector using Probabilistic Neural Network Representation of AIS Tracks and A Contrario Detection

arXiv · PDF · Open in the Atlas

Code that ran

10 of 14 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; 1 honoured the contract we drafted, 0 violated it and 9 have no contract verdict; 0 failures of our instrument; 4 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/1912.00682.json · badge: /api/ran/1912.00682.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
dnguyengithub/MultitaskAIS canonical (classifier label)
classifier basis: census
10 ran · 4 no recorded run
FunctionStatusWhere it lives
gather_tensors Ran
ran on a synthesized input
dnguyengithub/MultitaskAIS/nested_utils.py
code served (permissive licence) · get_code("a2b7bc574c08232a")
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 tensorflow as tf
import numpy as np

# Create sample tensors with distinct shapes
tensor_1 = tf.constant(np.random.randn(5, 4), dtype=tf.float32)
tensor_2 = tf.constant(np.random.randn(5, 3), dtype=tf.float32)
tensor_3 = tf.constant(np.random.randn(5, 2), dtype=tf.float32)

# Create indices for gathering
indices = tf.constant([0, 2, 4], dtype=tf.int32)

# Test with different nested structures
tensors_list = [tensor_1, tensor_2]
tensors_nested = (tensor_1, [tensor_2, tensor_3])

ARGS = {
    "tensors": tensors_nested,
    "indices": indices,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — 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
map_nested Ran
ran on a synthesized input
dnguyengithub/MultitaskAIS/nested_utils.py
code served (permissive licence) · get_code("e873390031b7a546")
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
from tensorflow.python.util import nest

# Define a simple map function that will be applied to each element
def simple_map_fn(x):
    """A simple function that adds 1 to numeric values or uppercases strings."""
    if isinstance(x, (int, float)):
        return x + 1
    elif isinstance(x, str):
        return x.upper()
    else:
        return x

# Create a nested structure with various types of elements
# This demonstrates a deeply nested combination of lists, tuples, and values
nested_structure = [
    1,
    (2, 'hello'),
    [3.5, (4, 'world')],
    {'key': 5}
]

# Construct ARGS dictionary with all required parameters
ARGS = {
    'map_fn': simple_map_fn,
    'nested': nested_structure
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow, 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
nCr Ran
ran, honoured the contract we drafted
dnguyengithub/MultitaskAIS/contrario_utils.py
code served (permissive licence) · get_code("666ef01e4aba55f6")
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 operator as op
from functools import reduce

# Define ARGS dict with valid arguments for nCr function
ARGS = {
    "n": 10,
    "r": 3
}
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 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
never_resample_criterion Ran
ran on a synthesized input
dnguyengithub/MultitaskAIS/bounds.py
code served (permissive licence) · get_code("70da44a4051c6a4d")
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 EagerTensor, ndim 2, dtype family float, shape (2, 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 tensorflow as tf
import numpy as np

# Set seed for reproducibility
tf.random.set_seed(0)
np.random.seed(0)

# Create small tensors for the function parameters
batch_size = 2
log_ess_tensor = tf.constant([[0.5, 0.3], [0.4, 0.6]], dtype=tf.float32)

unused_num_samples = 10
log_ess = log_ess_tensor
unused_t = 0

ARGS = {
    "unused_num_samples": unused_num_samples,
    "log_ess": log_ess,
    "unused_t": unused_t,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 2, dtype family float, shape (2, 2) — shape (2, 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
nonzero_segments Ran
ran on a synthesized input
dnguyengithub/MultitaskAIS/contrario_utils.py
code served (permissive licence) · get_code("ffd1c1c61eab5ced")
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 numpy as np

torch.manual_seed(0)

# The function nonzero_segments takes a single parameter x_ which should be
# an indexable sequence (list, tuple, or array-like) that supports:
# - len(x_) to get length
# - x_[d_i] to access elements by index
# - comparison with 0 (x_[d_i] != 0)
# The function identifies consecutive runs of nonzero values and returns their indices

# Create a simple list with some zero and nonzero elements
x_ = [0, 1, 2, 0, 0, 3, 4, 5, 0, 6, 0]

ARGS = {
    "x_": x_
}
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 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
sample_from_max_logits Ran
ran on a synthesized input
dnguyengithub/MultitaskAIS/distribution_utils.py
code served (permissive licence) · get_code("47eb9ac7bd14d87a")
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 EagerTensor, ndim 2, dtype family float, shape (2, 19) 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 tensorflow as tf

# Set random seed for reproducibility
tf.random.set_seed(0)

# Define bin sizes for each dimension
lat_bins = 5
lon_bins = 7
sog_bins = 3
cog_bins = 4

# Create a batch of logits
# Batch size: 2
# Total logit dimensions: lat_bins + lon_bins + sog_bins + cog_bins = 5 + 7 + 3 + 4 = 19
batch_size = 2
total_logit_dim = lat_bins + lon_bins + sog_bins + cog_bins

logit = tf.random.normal([batch_size, total_logit_dim], dtype=tf.float32)

ARGS = {
    "logit": logit,
    "lat_bins": lat_bins,
    "lon_bins": lon_bins,
    "sog_bins": sog_bins,
    "cog_bins": cog_bins,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 2, dtype family float, shape (2, 19) — shape (2, 19) 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
sublist Ran
ran on a synthesized input
dnguyengithub/MultitaskAIS/generate_vessel_type_dataset.py
code served (permissive licence) · get_code("c1f80016dbb7f740")
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 bool, 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)

# The function sublist takes two parameters: lst1 and lst2
# It filters elements from lst1 that are in lst2, and vice versa
# Then checks if both filtered lists are non-empty and equal
# So we need two lists with some overlapping elements

lst1 = [1, 2, 3, 4]
lst2 = [2, 3, 5, 6]

ARGS = {
    "lst1": lst1,
    "lst2": lst2
}
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 bool, 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
tile_tensors Ran
ran on a synthesized input
dnguyengithub/MultitaskAIS/nested_utils.py
code served (permissive licence) · get_code("bc0273e28849cdd8")
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 tensorflow as tf
import numpy as np

# Create test tensors with distinct dimensions
tensor1 = tf.constant(np.random.randn(2, 3, 4), dtype=tf.float32)
tensor2 = tf.constant(np.random.randn(2, 5), dtype=tf.float32)
tensor3 = tf.constant(np.random.randn(2, 3, 4, 6), dtype=tf.float32)

# Test case 1: flat list of tensors
tensors_flat = [tensor1, tensor2]

# Test case 2: nested structure with tuples and lists
tensors_nested = [tensor1, (tensor2, tensor3)]

# Test case 3: single tensor
tensors_single = tensor1

# Test multiples - should be shorter than tensor ranks
multiples_2d = [2, 3]
multiples_1d = [2]

ARGS = {
    "tensors": tensors_nested,
    "multiples": multiples_2d
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — 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
trackOutlier Ran
ran on a synthesized input
dnguyengithub/MultitaskAIS/utils.py
code served (permissive licence) · get_code("448021fccec60c3d")
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 (5,) 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

# Create a small symmetric binary anomaly matrix
# This is a 5x5 matrix where 1 indicates anomaly between observations
A = np.zeros((5, 5))
idx = np.array([[0, 2], [1, 2], [1, 3], [0, 3], [2, 4], [3, 4]])
A[idx[:, 0], idx[:, 1]] = 1
A[idx[:, 1], idx[:, 0]] = 1

ARGS = {
    "A": A
}
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 ndarray, ndim 1, dtype family float, shape (5,) — shape (5,) 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
zero_segments Ran
ran on a synthesized input
dnguyengithub/MultitaskAIS/contrario_utils.py
code served (permissive licence) · get_code("1bfd90225b2bfd86")
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 numpy as np

torch.manual_seed(0)

# Analyzing the function:
# - Takes parameter x_ which is indexed with x_[d_i] and has len(x_)
# - Compares elements to 0 (x_[d_i] == 0)
# - Appends indices to lists
# - Returns a list of lists
# So x_ should be a list or array-like of numbers

# Simple test case: a list with some zeros and non-zeros
x_ = [1, 0, 0, 3, 0, 5, 0, 0, 0, 2]

ARGS = {
    "x_": x_
}
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 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
ess_criterion Not yet run
unverified (harvested, no recorded run)
dnguyengithub/MultitaskAIS/bounds.py
code served (permissive licence) · get_code("7f8bece65b2cbe84")
How it ran

no invocation recorded.

restore_checkpoint_if_exists Not yet run
unverified (harvested, no recorded run)
dnguyengithub/MultitaskAIS/runners.py
code served (permissive licence) · get_code("df0d2e93c521f8ce")
How it ran

no invocation recorded.

sample_from_logits Not yet run
unverified (harvested, no recorded run)
dnguyengithub/MultitaskAIS/distribution_utils.py
code served (permissive licence) · get_code("c3c0833e08e9626a")
How it ran

no invocation recorded.

sample_one_hot Not yet run
unverified (harvested, no recorded run)
dnguyengithub/MultitaskAIS/distribution_utils.py
code served (permissive licence) · get_code("0d4537caed34aa77")
How it ran

no invocation recorded.

Also present in other papers' code

Of 14 samples lifted from this paper's linked repositories, 14 carry a function identity and 0 do not. Of the 14 distinct functions with an identity, 1 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 1 later) and 13 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
nCr 1 other paper
1 later
2209.05559 (2022-09)
Deep Reinforcement Learning for Cryptocurrency Trading: Practical Approach to Address Backtest Overfitting

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

Representing maritime traffic patterns and detecting anomalies from them are key to vessel monitoring and maritime situational awareness. We propose a novel approach -- referred to as GeoTrackNet -- for maritime anomaly detection from AIS data streams. Our model exploits state-of-the-art neural network schemes to learn a probabilistic representation of AIS tracks and a contrario detection to detect abnormal events. The neural network provides a new means to capture complex and heterogeneous patterns in vessels' behaviours, while the \textit{a contrario} detector takes into account the fact that the learnt distribution may be location-dependent. Experiments on a real AIS dataset comprising more than 4.2 million AIS messages demonstrate the relevance of the proposed method compared with state-of-the-art schemes.

For agents

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

get_harvested_code_for_paper("1912.00682")
get_code_for_paper("1912.00682")
have("1912.00682")

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

curl https://syntology.ai/api/ran/1912.00682.json

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

[![Syntology run record](https://syntology.ai/api/ran/1912.00682.svg)](https://syntology.ai/paper/1912.00682)

Connect an agent — have() is free.