SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2010.14860 · 2020

The ELBO of Variational Autoencoders Converges to a Sum of Three Entropies

arXiv · PDF · Open in the Atlas

Code that ran

5 of 6 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 5 have no contract verdict; 0 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/2010.14860.json · badge: /api/ran/2010.14860.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
learning-with-entropies/elbo-entropies-aistats23 canonical (classifier label)
classifier basis: census
5 ran · 1 no recorded run
FunctionStatusWhere it lives
ema Ran
ran on a synthesized input
learning-with-entropies/elbo-entropies-aistats23/utils.py
code served (permissive licence) · get_code("419e916c07567d5b")
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 1, dtype family float, shape (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 numpy as np

# Construct valid arguments for the ema function
values = np.array([1.0, 2.0, 3.0, 4.0, 5.0])
alpha = 0.3

ARGS = {
    "values": values,
    "alpha": alpha
}
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 1, dtype family float, shape (5,) — shape (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
logalpha2_opt Ran
ran on a synthesized input
learning-with-entropies/elbo-entropies-aistats23/elbo.py
code served (permissive licence) · get_code("208718b71db86a45")
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 (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

torch.manual_seed(0)

# Create small tensors for encoder means and log-variances
# batch_size=2, latent_dim=4
enc_mu = torch.randn(2, 4)
enc_logvar = torch.randn(2, 4)

ARGS = {
    "enc_mu": enc_mu,
    "enc_logvar": enc_logvar,
}
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 (4,) — shape (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
logsigma2_opt Ran
ran on a synthesized input
learning-with-entropies/elbo-entropies-aistats23/elbo.py
code served (permissive licence) · get_code("8e5836c9a5792e09")
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

torch.manual_seed(0)

# Create a minimal VAE instance
class MinimalVAE(nn.Module):
    def __init__(self, D):
        super().__init__()
        self.D = D
        self.fc = nn.Linear(10, 10)
    
    def forward(self, x):
        return self.fc(x)

# Create a minimal config
class Config:
    def __init__(self, dataset_type):
        self.dataset_type = dataset_type

# Test case 1: CELEBA dataset type (uses 4D tensors)
vae_celeba = MinimalVAE(D=48)  # D = 2 * 4 * 3 * 2 (batch * height * width * channels)
config_celeba = Config(dataset_type='CELEBA')
data_batch_celeba = torch.randn(2, 4, 3, 2)  # batch=2, height=4, width=3, channels=2
recon_batch_celeba = torch.randn(2, 4, 3, 2)

# Test case 2: Non-CELEBA dataset type (uses 2D tensors)
vae_other = MinimalVAE(D=8)  # D = 2 * 4 (batch * features)
config_other = Config(dataset_type='MNIST')
data_batch_other = torch.randn(2, 4)  # batch=2, features=4
recon_batch_other = torch.randn(2, 4)

# Use the CELEBA case as the primary test
ARGS = {
    'vae': vae_celeba,
    'data_batch': data_batch_celeba,
    'recon_batch': recon_batch_celeba,
    'config': config_celeba
}
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
read_files Ran
ran on a synthesized input
learning-with-entropies/elbo-entropies-aistats23/post_process/pp_plot.py
code served (permissive licence) · get_code("d1ffdcabfb14f976")
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 pandas as pd
import numpy as np
import tempfile
import os

# Create a temporary directory with sample CSV files
temp_dir = tempfile.mkdtemp()

# Define parameters
n_samples_list = [10, 20, 30]
n_runs_val = 5

# Create sample CSV files for each n_samples value
for n in n_samples_list:
    data = {
        'ELBO': np.random.randn(n_runs_val),
        'H_sum': np.random.randn(n_runs_val)
    }
    df = pd.DataFrame(data)
    csv_path = os.path.join(temp_dir, f'estimates_{n}.csv')
    df.to_csv(csv_path, index=False)

ARGS = {
    'PATH_TO_FOLDER': temp_dir,
    'n_samples': n_samples_list,
    'n_runs': n_runs_val
}
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 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
tensorlist_to_numpy Ran
ran on a synthesized input
learning-with-entropies/elbo-entropies-aistats23/utils.py
code served (permissive licence) · get_code("53e8a83b869472da")
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, 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
import numpy as np

torch.manual_seed(0)

# Create a list of tensors (tensorlist)
# The function expects:
# - len(tensorlist) to give the number of steps
# - len(tensorlist[0]) to give the number of latents
# - Each tensor should be convertible to numpy via .to('cpu').numpy()

tensor_list = [
    torch.randn(3),  # First timestep with 3 latent dimensions
    torch.randn(3),  # Second timestep with 3 latent dimensions
]

ARGS = {
    "tensorlist": tensor_list
}
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 ndarray, 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
lower_bound Not yet run
unverified (harvested, no recorded run)
learning-with-entropies/elbo-entropies-aistats23/elbo.py
code served (permissive licence) · get_code("d193a40b5e212dd4")
How it ran

no invocation recorded.

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 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 6 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

The central objective function of a variational autoencoder (VAE) is its variational lower bound (the ELBO). Here we show that for standard (i.e., Gaussian) VAEs the ELBO converges to a value given by the sum of three entropies: the (negative) entropy of the prior distribution, the expected (negative) entropy of the observable distribution, and the average entropy of the variational distributions (the latter is already part of the ELBO). Our derived analytical results are exact and apply for small as well as for intricate deep networks for encoder and decoder. Furthermore, they apply for finitely and infinitely many data points and at any stationary point (including local maxima and saddle points). The result implies that the ELBO can for standard VAEs often be computed in closed-form at stationary points while the original ELBO requires numerical approximations of integrals. As a main contribution, we provide the proof that the ELBO for VAEs is at stationary points equal to entropy sums. Numerical experiments then show that the obtained analytical results are sufficiently precise also in those vicinities of stationary points that are reached in practice. Furthermore, we discuss how the novel entropy form of the ELBO can be used to analyze and understand learning behavior. More generally, we believe that our contributions can be useful for future theoretical and practical studies on VAE learning as they provide novel information on those points in parameters space that optimization of VAEs converges to.

For agents

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

get_harvested_code_for_paper("2010.14860")
get_code_for_paper("2010.14860")
have("2010.14860")

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

curl https://syntology.ai/api/ran/2010.14860.json

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

[![Syntology run record](https://syntology.ai/api/ran/2010.14860.svg)](https://syntology.ai/paper/2010.14860)

Connect an agent — have() is free.