SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2410.01259 · 2024

Revisiting Optimism and Model Complexity in the Wake of Overparameterized Machine Learning

arXiv · PDF · Open in the Atlas

Code that ran

3 of 5 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; 3 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 0 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/2410.01259.json · badge: /api/ran/2410.01259.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
jaydu1/model-complexity canonical (classifier label)
classifier basis: classifier_2025
3 ran · 2 no recorded run
FunctionStatusWhere it lives
v_phi_lam Ran
ran, honoured the contract we drafted
jaydu1/model-complexity/compute_risk.py
code served (permissive licence) · get_code("308492dbffcd594a")
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 numpy as np

# Construct valid arguments for v_phi_lam function
# The function takes phi, lam, and optional a parameter

# Test case 1: lam > 0, phi > 0 (normal case)
phi_val = 2.5
lam_val = 0.5
a_val = 1.0

ARGS = {
    "phi": phi_val,
    "lam": lam_val,
    "a": a_val
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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
vb_phi_lam Ran
ran, honoured the contract we drafted
jaydu1/model-complexity/compute_risk.py
code served (permissive licence) · get_code("2fd9ecc03ae15f9a")
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 numpy as np

def v_phi_lam(phi, lam, a=1):
    """
    The unique solution v for fixed-point equation
        1 / v(-lam;phi) = lam + phi * int r / 1 + r * v(-lam;phi) dH(r)
    where H is the distribution of eigenvalues of Sigma.
    For isotopic features Sigma = a*I, the solution has a closed form, which reads that
        lam>0:
            v(-lam;phi) = (-(phi+lam/a-1)+np.sqrt((phi+lam/a-1)**2+4*lam/a))/(2*lam)
        lam=0, phi>1
            v(-lam;phi) = 1/(a*(phi-1))
    and undefined otherwise.
    """
    assert a > 0
    min_lam = -(1 - np.sqrt(phi)) ** 2 * a
    if phi <= 0.0 or lam < min_lam:
        raise ValueError('The input parameters should satisfy phi>0 and lam>=min_lam.')
    if phi == np.inf:
        return 0
    elif lam != 0:
        return (-(phi + lam / a - 1) + np.sqrt((phi + lam / a - 1) ** 2 + 4 * lam / a)) / (2 * lam)
    elif phi <= 1.0:
        return np.inf
    else:
        return 1 / (a * (phi - 1))

def vb_phi_lam(phi, lam, a=1, v=None):
    if lam == 0:
        if phi > 1:
            return 1 / (phi - 1)
        else:
            return phi / (phi - 1)
    else:
        if v is None:
            v = v_phi_lam(phi, lam, a)
        return phi / (1 / a + v) ** 2 / (1 / v ** 2 - phi / (1 / a + v) ** 2)

ARGS = {
    'phi': 2.5,
    'lam': 0.1,
    'a': 1.0,
    'v': None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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
vv_phi_lam Ran
ran, honoured the contract we drafted
jaydu1/model-complexity/compute_risk.py
code served (permissive licence) · get_code("4cbfeff98fca978b")
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 numpy as np

# Construct valid arguments for vv_phi_lam function
# The function takes parameters: phi, lam, a=1, v=None

# Test case 1: lam != 0, v is None (will be computed internally)
phi_1 = 2.5
lam_1 = 0.5
a_1 = 1.0
v_1 = None

# Test case 2: lam = 0, phi > 1, v is None
phi_2 = 2.0
lam_2 = 0
a_2 = 1.0
v_2 = None

# Test case 3: lam != 0, v is provided explicitly
phi_3 = 3.0
lam_3 = 1.0
a_3 = 2.0
v_3 = 0.5

# Use the first test case for ARGS
ARGS = {
    'phi': 2.5,
    'lam': 0.5,
    'a': 1.0,
    'v': None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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
run_one_simulation Not yet run
unverified (harvested, no recorded run)
jaydu1/model-complexity/run_simu_mn2ls.py
code served (permissive licence) · get_code("b1c4a197b74335cc")
How it ran

no invocation recorded.

run_one_simulation_theory Not yet run
unverified (harvested, no recorded run)
jaydu1/model-complexity/run_simu_mn2ls.py
code served (permissive licence) · get_code("1d1de2012bc7d34b")
How it ran

no invocation recorded.

Also present in other papers' code

Of 5 samples lifted from this paper's linked repositories, 5 carry a function identity and 0 do not. Of the 5 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 5 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

Common practice in modern machine learning involves fitting a large number of parameters relative to the number of observations. These overparameterized models can exhibit surprising generalization behavior, e.g., ``double descent'' in the prediction error curve when plotted against the raw number of model parameters, or another simplistic notion of complexity. In this paper, we revisit model complexity from first principles, by first reinterpreting and then extending the classical statistical concept of (effective) degrees of freedom. Whereas the classical definition is connected to fixed-X prediction error (in which prediction error is defined by averaging over the same, nonrandom covariate points as those used during training), our extension of degrees of freedom is connected to random-X prediction error (in which prediction error is averaged over a new, random sample from the covariate distribution). The random-X setting more naturally embodies modern machine learning problems, where highly complex models, even those complex enough to interpolate the training data, can still lead to desirable generalization performance under appropriate conditions. We demonstrate the utility of our proposed complexity measures through a mix of conceptual arguments, theory, and experiments, and illustrate how they can be used to interpret and compare arbitrary prediction models.

For agents

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

get_harvested_code_for_paper("2410.01259")
get_code_for_paper("2410.01259")
have("2410.01259")

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

curl https://syntology.ai/api/ran/2410.01259.json

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

[![Syntology run record](https://syntology.ai/api/ran/2410.01259.svg)](https://syntology.ai/paper/2410.01259)

Connect an agent — have() is free.