SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2604.02472 · 2026

VALOR: Value-Aware Revenue Uplift Modeling with Treatment-Gated Representation for B2B Sales

Debanshu Das, Kavin Soni, Vamshi Guduguntla

arXiv · PDF · Open in the Atlas

Code that ran

9 of 10 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, 1 violated it and 0 have no contract verdict; 8 failures of our instrument; 1 no recorded run), measured 2026-11-21; 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/2604.02472.json · badge: /api/ran/2604.02472.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
vamshing/VALOR-2026 canonical (classifier label)
classifier basis: census
9 ran · 1 no recorded run
FunctionStatusWhere it lives
compute_mmd_loss Ran
ran; our contract draft was wrong, not the code
vamshing/VALOR-2026/training/losses.py
code served (permissive licence) · get_code("04712bbc90fec2f3")
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)

ARGS = {
    "rep_t": torch.randn(2, 4),
    "rep_c": torch.randn(2, 4),
}
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
compute_wasserstein_loss Ran
ran; our contract draft was wrong, not the code
vamshing/VALOR-2026/training/losses.py
code served (permissive licence) · get_code("6c8a24c48135b68f")
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)

# Create two small tensors with distinct shapes
# rep_t: batch_size=2, embedding_dim=4
rep_t = torch.randn(2, 4)

# rep_c: batch_size=2, embedding_dim=4
rep_c = torch.randn(2, 4)

ARGS = {
    "rep_t": rep_t,
    "rep_c": rep_c,
}
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
get_cumlift Ran
ran; our contract draft was wrong, not the code
vamshing/VALOR-2026/utils/metrics.py
code served (permissive licence) · get_code("6e73cce0707f79ab")
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 Series. 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 sample DataFrame with required columns
np.random.seed(0)
n_samples = 10

df = pd.DataFrame({
    'model': np.random.rand(n_samples),  # model predictions
    'y': np.random.randint(0, 2, n_samples),  # outcome (binary)
    'w': np.random.randint(0, 2, n_samples)   # treatment (binary)
})

ARGS = {
    'df': df
}
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 Series — a Series output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Series output the recorded ndim 1 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
get_data_with_treatment_type Ran
ran; our contract draft was wrong, not the code
vamshing/VALOR-2026/models/baselines.py
code served (permissive licence) · get_code("c1c1f3f53f2ff804")
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 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

torch.manual_seed(0)

# Create sample data tensor: shape (6, 4) - 6 samples, 4 features
data = torch.randn(6, 4)

# Create treatment tensor: shape (6,) or (6, 1) - binary treatment assignment
treatment = torch.tensor([1, 0, 1, 0, 1, 0], dtype=torch.long)

ARGS = {
    "data": data,
    "treatment": treatment,
}
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 tuple, ndim 3, 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
get_true_kendall Ran
ran, violated the contract we drafted
vamshing/VALOR-2026/utils/metrics.py
code served (permissive licence) · get_code("bf6d1b33d17b34c9")
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; 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 numpy arrays for predicted and true uplift values
# These represent predicted and ground truth Individual Treatment Effect (ITE) values
np.random.seed(0)

preds = np.random.randn(100)  # Predicted uplift values
true_tau = np.random.randn(100)  # Ground truth uplift (ITE) values

ARGS = {
    "preds": preds,
    "true_tau": true_tau,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — 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 — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
lift_h_metric Ran
ran; our contract draft was wrong, not the code
vamshing/VALOR-2026/utils/metrics.py
code served (permissive licence) · get_code("4a0331ce64a7a127")
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; 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 pandas as pd
import numpy as np

# Create a sample DataFrame with the required columns
np.random.seed(0)
n_samples = 100

df = pd.DataFrame({
    'label': np.random.randint(0, 2, n_samples),  # binary outcome
    'treatment': np.random.randint(0, 2, n_samples),  # binary treatment indicator
    '1': np.random.randn(n_samples)  # predicted uplift scores
})

ARGS = {
    'df': df,
    'h': 0.3,
    'outcome_col': 'label',
    'treatment_col': 'treatment',
    'predict_col': '1'
}
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 float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
maximum_mean_discrepancy_loss Ran
ran; our contract draft was wrong, not the code
vamshing/VALOR-2026/models/baselines.py
code served (permissive licence) · get_code("9a1a9e0086c05d1c")
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)

# Create small sample tensors
sample_1 = torch.randn(2, 4)  # batch_size=2, feature_dim=4
sample_2 = torch.randn(3, 4)  # batch_size=3, feature_dim=4

ARGS = {
    'sample_1': sample_1,
    'sample_2': sample_2,
    'scale': [0.1, 1.0, 10.0]
}
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
pdist Ran
ran; our contract draft was wrong, not the code
vamshing/VALOR-2026/models/baselines.py
code served (permissive licence) · get_code("9ec1888c5308b375")
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, 3) 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 two sample tensors with distinct sizes for different dimensions
# sample_1: (batch_size=2, feature_dim=4)
# sample_2: (batch_size=3, feature_dim=4)
sample_1 = torch.randn(2, 4)
sample_2 = torch.randn(3, 4)

ARGS = {
    "sample_1": sample_1,
    "sample_2": sample_2,
    "norm": 2,
    "eps": 1e-05
}
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, 3) — shape (2, 3) 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
valor_loss_function Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
vamshing/VALOR-2026/training/losses.py
code served (permissive licence) · get_code("0ab4d520dfed8099")
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
import torch.nn as nn
import torch.nn.functional as F

torch.manual_seed(0)

# Create tensors for the function
batch_size = 2
feature_dim = 4

# Labels: small positive values (used for log in ZILN)
labels = torch.tensor([0.5, 1.2], dtype=torch.float32)

# Treatment: binary indicator (0 or 1)
treatment = torch.tensor([0.0, 1.0], dtype=torch.float32)

# Control output: (logits, mu, sigma)
c_logits = torch.randn(batch_size, dtype=torch.float32)
c_mu = torch.randn(batch_size, dtype=torch.float32)
c_sigma = torch.abs(torch.randn(batch_size, dtype=torch.float32)) + 0.1  # ensure positive
c_out = (c_logits, c_mu, c_sigma)

# Treatment output: (logits, mu, sigma)
t_logits = torch.randn(batch_size, dtype=torch.float32)
t_mu = torch.randn(batch_size, dtype=torch.float32)
t_sigma = torch.abs(torch.randn(batch_size, dtype=torch.float32)) + 0.1  # ensure positive
t_out = (t_logits, t_mu, t_sigma)

# Construct ARGS dictionary with all parameters
ARGS = {
    'c_out': c_out,
    't_out': t_out,
    'labels': labels,
    'treatment': treatment,
    'lambda_rank': 0.1,
    'alpha': 0.75,
    'gamma': 2.0,
    'use_focal': True,
    'use_ranking': True,
    'use_ziln': True,
}
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
train_epoch Not yet run
unverified (harvested, no recorded run)
vamshing/VALOR-2026/training/trainers.py
code served (permissive licence) · get_code("bdbfb2db8e1a69ee")
How it ran

no invocation recorded.

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 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 10 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 2026-11-21.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

B2B sales organizations must identify "persuadable" accounts within zero-inflated revenue distributions to optimize expensive human resource allocation. Standard uplift frameworks struggle with treatment signal collapse in high-dimensional spaces and a misalignment between regression calibration and the ranking of high-value "whales." We introduce VALOR (Value Aware Learning of Optimized (B2B) Revenue), a unified framework featuring a Treatment-Gated Sparse-Revenue Network that uses bilinear interaction to prevent causal signal collapse. The framework is optimized via a novel Cost-Sensitive Focal-ZILN objective that combines a focal mechanism for distributional robustness with a value-weighted ranking loss that scales penalties based on financial magnitude. To provide interpretability for high-touch sales programs, we further derive Robust ZILN-GBDT, a tree based variant utilizing a custom splitting criterion for uplift heterogeneity. Extensive evaluations confirm VALOR's dominance, achieving a 20% improvement in rankability over state-of-the-art methods on public benchmarks and delivering a validated 2.7× increase in incremental revenue per account in a rigorous 4-month production A/B test.

For agents

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

get_harvested_code_for_paper("2604.02472")
get_code_for_paper("2604.02472")
have("2604.02472")

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

curl https://syntology.ai/api/ran/2604.02472.json

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

[![Syntology run record](https://syntology.ai/api/ran/2604.02472.svg)](https://syntology.ai/paper/2604.02472)

Connect an agent — have() is free.