Papers › VarGrad: A Low-Variance Gradient Estimator for Variational Inference

VarGrad: A Low-Variance Gradient Estimator for Variational Inference

20 Oct 2020NeurIPS 2020 12arXiv:2010.10436archive 2025-07-28

Lorenz Richter, Ayman Boustati, Nikolas Nüsken, Francisco J. R. Ruiz, Ömer Deniz Akyildiz

We analyse the properties of an unbiased gradient estimator of the ELBO for variational inference, based on the score function method with leave-one-out control variates. We show that this gradient estimator can be obtained using a new loss, defined as the variance of the log-ratio between the exact posterior and the variational approximation, which we call the log-variance loss. Under certain conditions, the gradient of the log-variance loss equals the gradient of the (negative) ELBO. We show theoretically that this gradient estimator, which we call VarGrad due to its connection to the log-variance loss, exhibits lower variance than the score function method in certain settings, and that the leave-one-out control variate coefficients are close to the optimal ones. We empirically demonstrate that VarGrad offers a favourable variance versus computation trade-off compared to other state-of-the-art estimators on a discrete VAE.

PaperPDFConference PDFCodeCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 19 code samples harvested from 1 repository linked to this paper: 15 ran (of which 0 constructed an object rather than computing a result; 15 with no instrument failure: 0 honoured, 0 violated, 15 with no contract checked; 0 where Syntology's instrument failed) · 4 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.

By repository: official repository: 19 samples from 1 repository, 15 ran (15 with no instrument failure; 0 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written on 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

aboustati/vargrad officialmentioned in paperjaxApache-2.0Syntology: 19 samples, 15 ran (0 instrument-failed) · 4 unverified report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

19 samples harvested: 15 ran (of which 0 constructed an object rather than computing a result; 15 with no instrument failure: 0 honoured, 0 violated, 15 with no contract checked; 0 where Syntology's instrument failed) · 4 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

15ran
4unverified

Licence: 0 of the 19 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from aboustati/vargrad. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

aboustati/vargrad official repository 19 samples, 15 ran (0 instrument-failed) · 4 unverified
change_learning_rate experiments/vargrad_experiments/discrete_vae/config.py ran Apache-2.0 (permissive) · fe9915d4c3e50a53 · report fixture: signature config_dict:dict/0/?, new_learning_rate:float/0/float · returned dict · level recorded 2026-09-25
change_path experiments/vargrad_experiments/discrete_vae/config.py ran Apache-2.0 (permissive) · 5bf6ae65c27a84b7 · report fixture: signature config_dict:dict/0/?, new_path:str/0/str · returned dict · level recorded 2026-09-25
change_training_seed experiments/vargrad_experiments/discrete_vae/config.py ran Apache-2.0 (permissive) · 11564269193fd678 · report fixture: signature config_dict:dict/0/?, new_seed:int/0/int · returned dict · level recorded 2026-09-25
conditional_sample vargrad/distributions/binary_concrete.py ran Apache-2.0 (permissive) · eaea1bbb60313d70 · report fixture: signature p:ArrayImpl/2/float, y:ArrayImpl/2/float, temperature:float/0/float, key:ArrayImpl/1/int · returned ArrayImpl · level recorded 2026-09-25
conditional_sample vargrad/distributions/binary_relaxed.py ran Apache-2.0 (permissive) · 4fc51fcaff50f27e · report fixture: signature p:ArrayImpl/2/float, y:ArrayImpl/2/float, key:ArrayImpl/1/int · returned ArrayImpl · level recorded 2026-09-25
fori_loop vargrad/discrete_vae/utilities.py ran Apache-2.0 (permissive) · 9a3faf80f939fb16 · report fixture: signature lower:int/0/int, upper:int/0/int, body_fun:callable/0/?, init_val:Tensor/0/float · returned Tensor · level recorded 2026-09-25
generate_negative_elbo vargrad/discrete_vae/two_layer_objectives.py ran Apache-2.0 (permissive) · 9f917c4035c94cea · report fixture: signature encoder:tuple/1/callable, decoder:tuple/1/callable · returned function · level recorded 2026-09-25
generate_negative_elbo vargrad/discrete_vae/objectives.py ran Apache-2.0 (permissive) · f2ff37104e831e78 · report fixture: signature encoder:callable/0/?, decoder:callable/0/? · returned function · level recorded 2026-09-25
generate_score_function_objective vargrad/discrete_vae/two_layer_objectives.py ran Apache-2.0 (permissive) · 2493e2fccccb1611 · report fixture: signature encoder:tuple/1/callable, decoder:tuple/1/callable · returned function · level recorded 2026-09-25
generate_variance_loss_objective vargrad/discrete_vae/two_layer_objectives.py ran Apache-2.0 (permissive) · 27e122b570e527f9 · report fixture: signature encoder:tuple/1/callable, decoder:tuple/1/callable · returned function · level recorded 2026-09-25
logpdf vargrad/distributions/binary_concrete.py ran Apache-2.0 (permissive) · 624810d1981de662 · report fixture: signature x:ArrayImpl/2/float, p:ArrayImpl/2/float, temperature:float/0/float · returned ArrayImpl · level recorded 2026-09-25
logpmf vargrad/distributions/bernoulli.py ran fingerprinted Apache-2.0 (permissive) · b62664bcd7b69aeb · report fixture: signature x:ArrayImpl/2/float, p:ArrayImpl/2/float · returned ArrayImpl · level recorded 2026-09-25
sample in 167 papers
census: 8 implementations from 9 papers, 7 distinct outputs across 5 buckets, one shared input per bucket
vargrad/distributions/bernoulli.py ran Apache-2.0 (permissive) · ea712673f75d17f9 · report fixture: signature p:ArrayImpl/2/float, key:ArrayImpl/1/int, num_samples:int/0/int · returned ArrayImpl · level recorded 2026-09-25
sample in 167 papers
census: 8 implementations from 9 papers, 7 distinct outputs across 5 buckets, one shared input per bucket
vargrad/distributions/binary_concrete.py ran Apache-2.0 (permissive) · 829c661dd26bef5b · report fixture: signature p:ArrayImpl/1/float, temperature:float/0/float, key:ArrayImpl/1/int, num_samples:int/0/int · returned ArrayImpl · level recorded 2026-09-25
sample in 167 papers
census: 8 implementations from 9 papers, 7 distinct outputs across 5 buckets, one shared input per bucket
vargrad/distributions/binary_relaxed.py ran Apache-2.0 (permissive) · 369211dfef704bde · report fixture: signature p:ArrayImpl/1/float, key:ArrayImpl/1/int, num_samples:int/0/int · returned ArrayImpl · level recorded 2026-09-25
generate_score_function_objective vargrad/discrete_vae/objectives.py unverified Apache-2.0 (permissive) · dde442b8095695aa · report
generate_variance_loss_objective vargrad/discrete_vae/objectives.py unverified Apache-2.0 (permissive) · da1b7d08884188b5 · report
mnist in 14 papers vargrad/datasets.py unverified Apache-2.0 (permissive) · b08eaa978b0e8144 · report
omniglot vargrad/datasets.py unverified Apache-2.0 (permissive) · 7c4b2e4da8105181 · report

Tasks

Variational Inference

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections