SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2401.01268 · ICML · 2024

f -Divergence Based Classification: Beyond the Use of Cross-Entropy

Nicola Novello, Andrea Tonello

arXiv · PDF · Open in the Atlas

Code that ran

3 of 3 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 0 have no contract verdict; 3 failures 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/2401.01268.json · badge: /api/ran/2401.01268.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
tonellolab/discriminative-classification-fdiv reimplementation (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
FunctionStatusWhere it lives
sl_cost_fcn Ran
ran; our contract draft was wrong, not the code
tonellolab/discriminative-classification-fdiv/utils.py
code served (permissive licence) · get_code("a26e2711d1526b8b")
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 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)

# Analyze the function sl_cost_fcn:
# - out_1: passed to sl_first after squeeze(), then matmul with data_tx.T
# - out_2: passed to sl_sec after squeeze(), used in log/exp operations
# - data_tx: transposed and used in matmul with out_1, needs .float() method
# - num_classes: passed to sl_first (not directly used in the function body)
# - alpha: scalar multiplier

# sl_first expects:
# - y_pred: tensor that can be matmul'd with transposed data_tx
# - data_tx: tensor with .float() method, will be transposed
# - num_classes: integer parameter

# sl_sec expects:
# - y_pred: tensor that supports log, exp, mean operations

# Create tensors with distinct dimensions for clarity
batch_size = 2
feature_dim = 4
num_classes_val = 3

# out_1: after squeeze should be (batch_size, feature_dim) for matmul with (feature_dim, batch_size)
out_1 = torch.randn(batch_size, feature_dim)

# out_2: after squeeze should be (batch_size, feature_dim) for log/mean operations
out_2 = torch.randn(batch_size, feature_dim)

# data_tx: (batch_size, feature_dim) so transpose gives (feature_dim, batch_size)
data_tx = torch.randn(batch_size, feature_dim)

# num_classes: integer
num_classes = num_classes_val

# alpha: scalar weight
alpha = 0.5

ARGS = {
    "out_1": out_1,
    "out_2": out_2,
    "data_tx": data_tx,
    "num_classes": num_classes,
    "alpha": alpha,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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 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
sl_first Ran
ran; our contract draft was wrong, not the code
tonellolab/discriminative-classification-fdiv/utils.py
code served (permissive licence) · get_code("baaab3a5508da164")
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 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)

# Analyzing the function:
# - y_pred: used with torch.matmul, so needs to be a tensor
# - data_tx: used with .float() and torch.transpose, so needs to be a tensor
# - num_classes: not used in the function body
# - t_tensor: not used in the function body

# y_pred is left operand of matmul with shape that produces valid matmul result
# data_tx is transposed (0, 1), so it's 2D, then used as right operand of matmul
# The result diagonal is taken, suggesting square result
# If data_tx is (batch, features), transpose gives (features, batch)
# y_pred @ (features, batch).T = y_pred @ (batch, features) requires y_pred to be (batch, batch)
# Or y_pred could be (batch, features) and data_tx (batch, features), giving (batch, batch)

# Let's make:
# y_pred: (2, 4) - batch size 2, features 4
# data_tx: (2, 4) - batch size 2, features 4
# After transpose: (4, 2)
# y_pred @ data_tx.T = (2, 4) @ (4, 2) = (2, 2)
# diagonal of (2, 2) works

y_pred = torch.randn(2, 4)
data_tx = torch.randn(2, 4)
num_classes = 10
t_tensor = True

ARGS = {
    "y_pred": y_pred,
    "data_tx": data_tx,
    "num_classes": num_classes,
    "t_tensor": t_tensor
}
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 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
sl_sec Ran
ran; our contract draft was wrong, not the code
tonellolab/discriminative-classification-fdiv/utils.py
code served (permissive licence) · get_code("ac4d31e440531907")
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 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)

# y_pred is used with:
# - torch.log(y_pred): needs to be a tensor of positive values
# - torch.mean(y_pred, dim=1): needs to support mean reduction along dim 1
# - torch.mean(...): final mean reduction
# So y_pred should be a 2D tensor with batch size 2 and some feature dimension

y_pred = torch.rand(2, 4) + 0.1  # Add 0.1 to ensure values are positive for log

ARGS = {
    "y_pred": y_pred
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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 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

Also present in other papers' code

Of 3 samples lifted from this paper's linked repositories, 3 carry a function identity and 0 do not. Of the 3 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 3 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

In deep learning, classification tasks are formalized as optimization problems often solved via the minimization of the cross-entropy. However, recent advancements in the design of objective functions allow the usage of the f -divergence to generalize the formulation of the optimization problem for classification. We adopt a Bayesian perspective and formulate the classification task as a maximum a posteriori probability problem. We propose a class of objective functions based on the variational representation of the f -divergence. Furthermore, driven by the challenge of improving the state-of-the-art approach, we propose a bottom-up method that leads us to the formulation of an objective function corresponding to a novel f -divergence referred to as shifted log (SL). We theoretically analyze the objective functions proposed and numerically test them in three application scenarios: toy examples, image datasets, and signal detection/decoding problems. The analyzed scenarios demonstrate the effectiveness of the proposed approach and that the SL divergence achieves the highest classification accuracy in almost all the considered cases.

For agents

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

get_harvested_code_for_paper("2401.01268")
get_code_for_paper("2401.01268")
have("2401.01268")

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

curl https://syntology.ai/api/ran/2401.01268.json

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

[![Syntology run record](https://syntology.ai/api/ran/2401.01268.svg)](https://syntology.ai/paper/2401.01268)

Connect an agent — have() is free.