SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2310.12952 · 2023

Cousins Of The Vendi Score: A Family Of Similarity-Based Diversity Metrics For Science And Machine Learning

arXiv · PDF · Open in the Atlas

Code that ran

15 of 16 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; 4 honoured the contract we drafted, 0 violated it and 1 has no contract verdict; 10 failures of our instrument; 1 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/2310.12952.json · badge: /api/ran/2310.12952.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
vertaix/vendi-sampling canonical (classifier label)
classifier basis: census
12 ran · 1 no recorded run
copy not recorded community
not in the paper's declared repositories
3 ran · 0 no recorded run
FunctionStatusWhere it lives
ReplicaSamp Ran
ran; our contract draft was wrong, not the code
vertaix/vendi-sampling/Model_Systems/samplers.py
code served (permissive licence) · get_code("348448d68d212df2")
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 3, dtype family float, shape (3, 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)

# E is a Callable that takes a tensor and returns a scalar energy value
# The function calls E(x_init) and then computes gradients w.r.t. x_init
class EnergyFunction(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 1)
    
    def forward(self, x):
        # x shape: (replicas, dim)
        # Return scalar energy
        return self.linear(x).sum()

energy_fn = EnergyFunction()

# x_init: tensor of shape (replicas, dim)
# The function accesses x_init.size()[0] for replicas and x_init.size()[1] for dim
x_init = torch.randn(2, 4, requires_grad=False)

# eta: learning rate (float)
eta = 0.001

# steps: number of sampling steps (int)
steps = 3

ARGS = {
    "E": energy_fn,
    "steps": steps,
    "x_init": x_init,
    "eta": eta
}
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 ndarray, ndim 3, dtype family float, shape (3, 2, 4) — shape (3, 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
VendiSamp Ran
ran; our contract draft was wrong, not the code
vertaix/vendi-sampling/Model_Systems/samplers.py
code served (permissive licence) · get_code("8a83db86ebf5b6a0")
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 torch
import torch.nn as nn
from typing import Callable

torch.manual_seed(0)

# Define a simple energy function that returns a scalar per replica
class EnergyFunction(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 1)
    
    def forward(self, x):
        # x shape: (replicas, dim)
        # return shape: (replicas,) - energy for each replica
        return self.linear(x).squeeze(-1)

# Define a simple score function that returns a scalar per replica
class ScoreFunction(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 1)
    
    def forward(self, x):
        # x shape: (replicas, dim)
        # return shape: scalar (sum of scores)
        return self.linear(x).sum()

# Create instances
E = EnergyFunction()
score = ScoreFunction()

# Initialize x_init with shape (replicas=2, dim=4)
x_init = torch.randn(2, 4)

# Define ARGS dictionary
ARGS = {
    'E': E,
    'score': score,
    'steps': 5,
    'x_init': x_init,
    'eta': 0.001,
    'nu': 0.004,
    'stop': 100
}
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 — 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
canvas Ran
ran; our contract draft was wrong, not the code
vertaix/vendi-sampling/Molecular/reform/reform.py
code served (permissive licence) · get_code("61b30a0cbdae38d6")
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 str. 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

torch.manual_seed(0)

ARGS = {
    'with_attribution': 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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
compile_parts Ran
ran, honoured the contract we drafted
vertaix/vendi-sampling/Molecular/AnalyzeAla2Run.py
code served (permissive licence) · get_code("bafc8f4d36a86808")
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 2, dtype family float, shape (2, 558) 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
import tempfile
import os

# Create temporary directory and mock .npy files
temp_dir = tempfile.mkdtemp()

# Create 31 small numpy arrays and save them
for i in range(31):
    arr = np.random.rand(2, 3 + i)  # Varying shapes to make horizontal stacking work
    filepath = os.path.join(temp_dir, f'{i}.npy')
    np.save(filepath, arr)

path_prefix = os.path.join(temp_dir, '')

ARGS = {
    'path_prefix': path_prefix,
    'parts': 31
}
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 2, dtype family float, shape (2, 558) — shape (2, 558) 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
compile_parts Ran
ran, honoured the contract we drafted
vertaix/vendi-sampling/Molecular/AnalyzeCLN025Run.py
code served (permissive licence) · get_code("a7a401df2a3e7c54")
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 2, dtype family float, shape (3, 204) 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
import tempfile
import os

# Create temporary directory and dummy .npy files
temp_dir = tempfile.mkdtemp()

# Create 51 small numpy arrays and save them
for i in range(51):
    arr = np.random.randn(3, 4).astype(np.float32)
    filepath = os.path.join(temp_dir, f'{i}.npy')
    np.save(filepath, arr)

path_prefix = os.path.join(temp_dir, '')

ARGS = {
    'path_prefix': path_prefix,
    'parts': 51
}
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 2, dtype family float, shape (3, 204) — shape (3, 204) 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
entropy_q Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
this paper's copy was not recorded; identical code first harvested from vertaix/vendi-score
pointer only · get_code("9b080afd41ab800a")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

entropy_q Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
vertaix/vendi-sampling/Model_Systems/vendi.py
code served (permissive licence) · get_code("975c21659e2b0515")
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)

# The function entropy_q takes:
# - p: a tensor of probabilities (used with indexing p[p > 0], torch.log, sum, max operations)
# - q: either 1, 'inf', or a numeric value (used in comparisons and arithmetic)

# Create a small probability distribution tensor
p = torch.tensor([0.1, 0.0, 0.3, 0.6], dtype=torch.float32)

# Define q as a numeric value (testing the general case, not q=1 or q='inf')
q = 2

ARGS = {
    'p': p,
    'q': q
}
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
normalize_K Ran
ran, honoured the contract we drafted
this paper's copy was not recorded; identical code first harvested from vertaix/vendi-score
pointer only · get_code("f8d0a9d83f3972e9")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

normalize_K Ran
ran, honoured the contract we drafted
vertaix/vendi-sampling/Model_Systems/vendi.py
code served (permissive licence) · get_code("1c12a347a8a9f81d")
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 (3, 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 a small positive semi-definite matrix K
# The function normalizes K by dividing by outer product of its diagonal square roots
# K should be square and symmetric for this to make sense
K = torch.tensor([
    [4.0, 1.0, 0.5],
    [1.0, 3.0, 0.2],
    [0.5, 0.2, 2.0]
])

ARGS = {
    "K": K
}
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 (3, 3) — shape (3, 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
register_vcs_handler Ran
ran; our contract draft was wrong, not the code
vertaix/vendi-sampling/Molecular/reform/_version.py
pointer only (licence: MIT); the stored copy get_code serves, from Hhuangsj/GREX, is pointer only (licence: NONE) · get_code("f12f4dabbeb242eb")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from Hhuangsj/GREX, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

run_command Ran
ran on a synthesized input
vertaix/vendi-sampling/Molecular/reform/_version.py
pointer only (licence: MIT); the stored copy get_code serves, from Hhuangsj/GREX, is pointer only (licence: NONE) · get_code("86d4ff45d6295e29")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from Hhuangsj/GREX, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

snapshotHeatMap Ran
ran; our contract draft was wrong, not the code
vertaix/vendi-sampling/Molecular/AnalyzeAla2Run.py
code served (permissive licence) · get_code("5823b8966b721f65")
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 AxesImage. 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 numpy as np
import matplotlib.pyplot as plt

# Create a mock axes object
fig, ax = plt.subplots()

# Generate sample data for phi and psi
# These are angle arrays that will be sliced based on replica, t, and t_per_step
np.random.seed(0)
phi = np.random.uniform(-np.pi, np.pi, size=10000)
psi = np.random.uniform(-np.pi, np.pi, size=10000)

# Optional weights for the histogram
weights = np.random.uniform(0.5, 1.5, size=10000)

ARGS = {
    'ax': ax,
    'phi': phi,
    'psi': psi,
    't': 0.1,
    't_per_step': 0.0001,
    'replica': 32,
    'weights': weights,
    'show_bar': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type AxesImage — a AxesImage output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a AxesImage 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
snapshotHeatMap Ran
ran; our contract draft was wrong, not the code
vertaix/vendi-sampling/Molecular/AnalyzeCLN025Run.py
code served (permissive licence) · get_code("9194cf144017b1a7")
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 AxesImage. 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 numpy as np
import matplotlib.pyplot as plt
from matplotlib.axes import Axes

# Create a mock axes object
fig, ax = plt.subplots()

# Generate synthetic phi and psi data (dihedral angles)
# These should be arrays of angles, typically in range [-180, 180] or [0, 360]
np.random.seed(0)
n_samples = 10000
phi = np.random.uniform(-180, 180, n_samples)
psi = np.random.uniform(-180, 180, n_samples)

# Create optional weights array
weights = np.random.uniform(0.5, 1.5, n_samples)

# Define parameters
t = 0.1  # time value
t_per_step = 0.002  # time per step
replica = 32  # number of replicas

ARGS = {
    'ax': ax,
    'phi': phi,
    'psi': psi,
    't': t,
    't_per_step': t_per_step,
    'replica': replica,
    'weights': weights
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type AxesImage — a AxesImage output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a AxesImage 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
versions_from_parentdir Ran
ran; our contract draft was wrong, not the code
vertaix/vendi-sampling/Molecular/reform/_version.py
pointer only (licence: MIT); the stored copy get_code serves, from Hhuangsj/GREX, is pointer only (licence: NONE) · get_code("b9a1535d27f2a9fb")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from Hhuangsj/GREX, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

weight_K Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
this paper's copy was not recorded; identical code first harvested from vertaix/vendi-score
pointer only · get_code("9c2e5c5c9d5b8e8e")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

weight_K Not yet run
unverified (harvested, no recorded run)
vertaix/vendi-sampling/Model_Systems/vendi.py
code served (permissive licence) · get_code("979e31e73f5dfa7c")
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 16 distinct functions with an identity, 7 are also present in other papers whose code we hold (7 with at least one paper dated earlier, 0 the same month, 6 later) and 9 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
register_vcs_handler 26 other papers
23 dated earlier, 3 later
1806.00550 (2018-06)
A Swiss Army Infinitesimal Jackknife
1806.03146 (2018-06)
Neural Message Passing with Edge Updates for Predicting Properties of Molecules and Materials
1806.03720 (2018-06)
Stochastic seismic waveform inversion using generative adversarial networks as a geological prior
and 23 more not named here
versions_from_parentdir 16 other papers
14 dated earlier, 2 later
1806.00550 (2018-06)
A Swiss Army Infinitesimal Jackknife
1806.03720 (2018-06)
Stochastic seismic waveform inversion using generative adversarial networks as a geological prior
1812.01719 (2018-12)
Knowing what you know in brain segmentation using Bayesian deep neural networks
and 13 more not named here
run_command 14 other papers
12 dated earlier, 2 later
1806.00550 (2018-06)
A Swiss Army Infinitesimal Jackknife
1806.03720 (2018-06)
Stochastic seismic waveform inversion using generative adversarial networks as a geological prior
1812.01719 (2018-12)
Knowing what you know in brain segmentation using Bayesian deep neural networks
and 11 more not named here
entropy_q 2 other papers
1 dated earlier, 1 later
2210.02410 (2022-10)
The Vendi Score: A Diversity Evaluation Metric for Machine Learning
2608.30214 (2026-08)
SPARK: Skeleton-Guided Reasoning Synthesis from Large-Scale Scientific Literature
normalize_K 2 other papers
1 dated earlier, 1 later
2210.02410 (2022-10)
The Vendi Score: A Diversity Evaluation Metric for Machine Learning
2608.30214 (2026-08)
SPARK: Skeleton-Guided Reasoning Synthesis from Large-Scale Scientific Literature
weight_K 2 other papers
1 dated earlier, 1 later
2210.02410 (2022-10)
The Vendi Score: A Diversity Evaluation Metric for Machine Learning
2608.30214 (2026-08)
SPARK: Skeleton-Guided Reasoning Synthesis from Large-Scale Scientific Literature
normalize_K 1 other paper
1 dated earlier
2304.11961 (2023-04)
Towards Mode Balancing of Generative Models via Diversity Weights

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

Measuring diversity accurately is important for many scientific fields, including machine learning (ML), ecology, and chemistry. The Vendi Score was introduced as a generic similarity-based diversity metric that extends the Hill number of order q=1 by leveraging ideas from quantum statistical mechanics. Contrary to many diversity metrics in ecology, the Vendi Score accounts for similarity and does not require knowledge of the prevalence of the categories in the collection to be evaluated for diversity. However, the Vendi Score treats each item in a given collection with a level of sensitivity proportional to the item's prevalence. This is undesirable in settings where there is a significant imbalance in item prevalence. In this paper, we extend the other Hill numbers using similarity to provide flexibility in allocating sensitivity to rare or common items. This leads to a family of diversity metrics -- Vendi scores with different levels of sensitivity -- that can be used in a variety of applications. We study the properties of the scores in a synthetic controlled setting where the ground truth diversity is known. We then test their utility in improving molecular simulations via Vendi Sampling. Finally, we use the Vendi scores to better understand the behavior of image generative models in terms of memorization, duplication, diversity, and sample quality.

For agents

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

get_harvested_code_for_paper("2310.12952")
get_code_for_paper("2310.12952")
have("2310.12952")

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

curl https://syntology.ai/api/ran/2310.12952.json

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

[![Syntology run record](https://syntology.ai/api/ran/2310.12952.svg)](https://syntology.ai/paper/2310.12952)

Connect an agent — have() is free.