SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2410.14970 · NeurIPS · 2024

Taming the Long Tail in Human Mobility Prediction

Renhe Jiang, Chuang Yang, Xiaohang Xu, Zipei Fan, Kaoru Sezaki

arXiv · PDF · Open in the Atlas

Code that ran

14 of 16 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (8 of them constructed an object rather than computing a result; 3 honoured the contract we drafted, 0 violated it and 9 have no contract verdict; 2 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/2410.14970.json · badge: /api/ran/2410.14970.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
yukayo/lotnext canonical (classifier label)
classifier basis: classifier_2025
14 ran · 2 no recorded run
FunctionStatusWhere it lives
AttentionLayer Ran
ran on a synthesized input · constructed an object; no result was computed
yukayo/lotnext/network.py
code served (permissive licence) · get_code("39ffd4de890cc527")
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 an instance of its own class, AttentionLayer: the run constructed the object and did not call it, so no result was computed. 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
from torch.nn import init

torch.manual_seed(0)

# Define dimensions
user_dim = 4
item_dim = 5

# Create dummy embeddings
# user_embeddings: [num_users, user_dim]
user_embeddings = torch.randn(3, user_dim)

# item_embeddings: [num_items, item_dim]
item_embeddings = torch.randn(2, item_dim)

# Create edge_index: [2, num_edges]
# First row: user indices, Second row: item indices
edge_index = torch.tensor([
    [0, 1, 2, 0],  # user indices
    [0, 1, 0, 1]   # item indices
], dtype=torch.long)

ARGS = {
    'user_dim': user_dim,
    'item_dim': item_dim
}
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 AttentionLayer — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a AttentionLayer 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
CosineActivation Ran
ran on a synthesized input
yukayo/lotnext/network.py
code served (permissive licence) · get_code("c5179dfbf964106b")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 3, 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

INIT_ARGS = {
    "in_features": 4,
    "out_features": 3,
}

# Create input tensor with batch size 2 and feature dimension matching in_features
ARGS = {
    "tau": torch.randn(2, 4),
}
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
DenoisingGCNNet Ran
ran on a synthesized input · constructed an object; no result was computed
yukayo/lotnext/network.py
code served (permissive licence) · get_code("af17f75e18c219e8")
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 an instance of its own class, DenoisingGCNNet: the run constructed the object and did not call it, so no result was computed. 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
from torch_geometric.nn import GCNConv
from torch.nn import init

torch.manual_seed(0)

# Define model parameters
user_dim = 8
item_dim = 12
out_channels = 16

# Create the model
model = nn.Module()

# Create sample tensors
# user_embeddings: (num_users, user_dim)
num_users = 2
user_embeddings = torch.randn(num_users, user_dim)

# item_embeddings: (num_items, item_dim)
num_items = 2
item_embeddings = torch.randn(num_items, item_dim)

# edge_index: (2, num_edges)
# Each edge connects a user to an item
num_edges = 3
edge_index = torch.tensor([
    [0, 1, 0],  # user indices
    [0, 1, 1]   # item indices
], dtype=torch.long)

ARGS = {
    'user_dim': user_dim,
    'item_dim': item_dim,
    'out_channels': out_channels
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torch_geometric — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DenoisingGCNNet — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a DenoisingGCNNet 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
EncoderLayer Ran
ran on a synthesized input · constructed an object; no result was computed
yukayo/lotnext/network.py
code served (permissive licence) · get_code("d7994a00f2625856")
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 an instance of its own class, EncoderLayer: the run constructed the object and did not call it, so no result was computed. 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 numpy as np

torch.manual_seed(0)

ARGS = {
    "hidden_size": 8,
    "ffn_size": 16,
    "dropout_rate": 0.1,
    "attention_dropout_rate": 0.1,
    "num_heads": 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 EncoderLayer — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a EncoderLayer 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
FuseEmbeddings Ran
ran on a synthesized input · constructed an object; no result was computed
yukayo/lotnext/network.py
code served (permissive licence) · get_code("7691514af8a9c61f")
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 an instance of its own class, FuseEmbeddings: the run constructed the object and did not call it, so no result was computed. 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)

# Define the FuseEmbeddings module
class FuseEmbeddings(nn.Module):
    def __init__(self, user_embed_dim, poi_embed_dim):
        super(FuseEmbeddings, self).__init__()
        embed_dim = user_embed_dim + poi_embed_dim
        self.fuse_embed = nn.Linear(embed_dim, embed_dim)
        self.leaky_relu = nn.LeakyReLU(0.2)

    def forward(self, user_embed, poi_embed):
        x = self.fuse_embed(torch.cat((user_embed, poi_embed), 2))
        x = self.leaky_relu(x)
        return x

# Define test arguments
user_embed_dim = 4
poi_embed_dim = 3

ARGS = {
    "user_embed_dim": user_embed_dim,
    "poi_embed_dim": poi_embed_dim,
}
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 FuseEmbeddings — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a FuseEmbeddings 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
GCNLayer Ran
ran on a synthesized input · constructed an object; no result was computed
yukayo/lotnext/network.py
code served (permissive licence) · get_code("0128b93f0fc0c806")
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 an instance of its own class, GCNLayer: the run constructed the object and did not call it, so no result was computed. 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
from torch_geometric.nn import GCNConv

torch.manual_seed(0)

# GCNLayer expects:
# - in_channels: int (input feature dimension)
# - out_channels: int (output feature dimension)

# For forward pass, it needs:
# - x: node feature tensor of shape [num_nodes, in_channels]
# - edge_index: edge connectivity tensor of shape [2, num_edges]

in_channels = 4
out_channels = 3

# Create a small node feature matrix: 2 nodes with 4 features each
x = torch.randn(2, in_channels)

# Create edge index: 2 edges connecting the nodes
# edge_index format is [source_nodes, target_nodes]
edge_index = torch.tensor([[0, 1], [1, 0]], dtype=torch.long)

ARGS = {
    "in_channels": in_channels,
    "out_channels": out_channels,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torch_geometric — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type GCNLayer — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a GCNLayer 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
MultiHeadAttention Ran
ran on a synthesized input · constructed an object; no result was computed
yukayo/lotnext/network.py
code served (permissive licence) · get_code("71ea19393168127b")
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 an instance of its own class, MultiHeadAttention: the run constructed the object and did not call it, so no result was computed. 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 numpy as np

torch.manual_seed(0)

# Create sample tensors
# batch_size=2, seq_len=3, hidden_size=8
batch_size = 2
seq_len = 3
hidden_size = 8
num_heads = 2

q = torch.randn(batch_size, seq_len, hidden_size)
k = torch.randn(batch_size, seq_len, hidden_size)
v = torch.randn(batch_size, seq_len, hidden_size)

# Optional attention bias
attn_bias = None

# Current epoch for masking schedule
current_epoch = 0

# Mask flag
mask = True

ARGS = {
    "hidden_size": hidden_size,
    "attention_dropout_rate": 0.1,
    "num_heads": num_heads,
}
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 MultiHeadAttention — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a MultiHeadAttention 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
SineActivation1 Ran
ran on a synthesized input · constructed an object; no result was computed
yukayo/lotnext/network.py
code served (permissive licence) · get_code("203edaf76afe2af6")
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 an instance of its own class, SineActivation1: the run constructed the object and did not call it, so no result was computed. 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)

# SineActivation1 takes seq_len and out_features as parameters
seq_len = 7
out_features = 5

ARGS = {
    'seq_len': seq_len,
    'out_features': out_features
}
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 SineActivation1 — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a SineActivation1 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
Time2Vec Ran
ran on a synthesized input · constructed an object; no result was computed
yukayo/lotnext/network.py
code served (permissive licence) · get_code("6e58ae86619a5ac0")
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 an instance of its own class, Time2Vec: the run constructed the object and did not call it, so no result was computed. 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)

ARGS = {
    'activation': 'sin',
    'batch_size': 2,
    'seq_len': 4,
    'out_dim': 3
}
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 Time2Vec — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Time2Vec 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
haversine Ran
ran, honoured the contract we drafted
yukayo/lotnext/network.py
code served (permissive licence) · get_code("870ae801eddcf615")
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 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 math

torch.manual_seed(0)

# Create two batches of geographic locations [batch_size, 2]
# Each row contains [latitude, longitude] in degrees
s1 = torch.tensor([
    [40.7128, -74.0060],  # New York City
    [51.5074, -0.1278]     # London
], dtype=torch.float32)

s2 = torch.tensor([
    [48.8566, 2.3522],     # Paris
    [35.6762, 139.6503]    # Tokyo
], dtype=torch.float32)

ARGS = {
    "s1": s1,
    "s2": s2
}
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 Tensor, 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
mask_value Ran
ran, honoured the contract we drafted
yukayo/lotnext/model.py
code served (permissive licence) · get_code("16dbf894494405ad")
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 float64, 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
import numpy as np

torch.manual_seed(0)

# Analyzing the function:
# - epoch: used in division (epoch / T), should be a number
# - T: used in division (epoch / T), should be a number (total epochs)
# - v_min: default -100, used in np.log10(-v_min), should be negative
# - v_max: default -1000000000.0, used in np.log10(-v_max), should be negative

ARGS = {
    "epoch": 5,
    "T": 10,
    "v_min": -100,
    "v_max": -1000000000.0
}
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 float64, 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
mask_value Ran
ran, honoured the contract we drafted
yukayo/lotnext/network.py
code served (permissive licence) · get_code("944d26dcc9252cd9")
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 float64, 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 numpy as np

# The function mask_value takes:
# - epoch: an integer representing current epoch
# - T: an integer representing total epochs
# - v_min: a float, default -100
# - v_max: a float, default -1000000000.0

# Looking at the function body:
# - epoch and T are used in division (epoch / T), so they should be numeric
# - v_min and v_max are used with np.log10 on their negations
# - The function computes a logarithmic schedule between v_min and v_max

ARGS = {
    "epoch": 5,
    "T": 10,
    "v_min": -100,
    "v_max": -1000000000.0,
}
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 float64, 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
t2v Ran
ran; our contract draft was wrong, not the code
yukayo/lotnext/model.py
code served (permissive licence) · get_code("56be6093d5ee62df")
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 Tensor, ndim 2, dtype family float, shape (2, 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 torch
import torch.nn as nn
import torch.nn.functional as F
import math
import numpy as np

torch.manual_seed(0)

# tau: input tensor with time values
tau = torch.randn(2)

# f: a function that takes input and applies some transformation
f = torch.sin

# out_features: output feature dimension
out_features = 5

# w: weight tensor for the first part (tau * w + b)
# tau is shape (2,), after unsqueeze(-1) becomes (2, 1)
# so w should be shape (1, 4) to produce (2, 4) after multiplication
w = torch.randn(1, 4)

# b: bias tensor for the first part
b = torch.randn(1, 4)

# w0: weight tensor for the second part (tau * w0 + b0)
# should produce a single feature per sample
w0 = torch.randn(1, 1)

# b0: bias tensor for the second part
b0 = torch.randn(1, 1)

ARGS = {
    "tau": tau,
    "f": f,
    "out_features": out_features,
    "w": w,
    "b": b,
    "w0": w0,
    "b0": b0,
}
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 Tensor, ndim 2, dtype family float, shape (2, 5) — shape (2, 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
t2v Ran
ran; our contract draft was wrong, not the code
yukayo/lotnext/network.py
code served (permissive licence) · get_code("bc78b05475081c66")
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 torch.nn as nn

torch.manual_seed(0)

# tau: a tensor that will be unsqueezed and used in arithmetic
tau = torch.randn(2)

# f: a function/module that takes a tensor and returns a tensor
# It's called with (tau.unsqueeze(-1) * w + b), so it should accept a tensor
f = nn.ReLU()

# out_features: an integer specifying output features
out_features = 4

# w: a weight tensor used in multiplication with tau.unsqueeze(-1)
# tau is shape (2,), unsqueezed to (2, 1), so w should be shape (1, 3)
w = torch.randn(1, 3)

# b: a bias tensor added after w multiplication
# After tau.unsqueeze(-1) * w, shape is (2, 3), so b should be (1, 3) or (3,)
b = torch.randn(3)

# w0: a weight tensor used in multiplication with tau.unsqueeze(-1)
# Same as w, shape (1, 1) to give different dimension
w0 = torch.randn(1, 1)

# b0: a bias tensor added after w0 multiplication
# After tau.unsqueeze(-1) * w0, shape is (2, 1), so b0 should be (1, 1) or (1,)
b0 = torch.randn(1)

ARGS = {
    "tau": tau,
    "f": f,
    "out_features": out_features,
    "w": w,
    "b": b,
    "w0": w0,
    "b0": b0,
}
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 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
DenoisingLayer Not yet run
unverified (harvested, no recorded run)
yukayo/lotnext/network.py
code served (permissive licence) · get_code("8d2f0cf9459eb579")
How it ran

no invocation recorded.

Flashback Not yet run
unverified (harvested, no recorded run)
yukayo/lotnext/network.py
code served (permissive licence) · get_code("938100a84fe0e515")
How it ran

no invocation recorded.

Also present in other papers' code

Of 16 samples lifted from this paper's linked repositories, 16 carry a function identity and 0 do not. Of the 14 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 14 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

With the popularity of location-based services, human mobility prediction plays a key role in enhancing personalized navigation, optimizing recommendation systems, and facilitating urban mobility and planning. This involves predicting a user's next POI (point-of-interest) visit using their past visit history. However, the uneven distribution of visitations over time and space, namely the long-tail problem in spatial distribution, makes it difficult for AI models to predict those POIs that are less visited by humans. In light of this issue, we propose the Long-Tail Adjusted Next POI Prediction (LoTNext) framework for mobility prediction, combining a Long-Tailed Graph Adjustment module to reduce the impact of the long-tailed nodes in the user-POI interaction graph and a novel Long-Tailed Loss Adjustment module to adjust loss by logit score and sample weight adjustment strategy. Also, we employ the auxiliary prediction task to enhance generalization and accuracy. Our experiments with two real-world trajectory datasets demonstrate that LoTNext significantly surpasses existing state-of-the-art works.

For agents

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

get_harvested_code_for_paper("2410.14970")
get_code_for_paper("2410.14970")
have("2410.14970")

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

curl https://syntology.ai/api/ran/2410.14970.json

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

[![Syntology run record](https://syntology.ai/api/ran/2410.14970.svg)](https://syntology.ai/paper/2410.14970)

Connect an agent — have() is free.