Papers › Simplified and Generalized Masked Diffusion for Discrete Data

Simplified and Generalized Masked Diffusion for Discrete Data

6 Jun 2024arXiv:2406.04329archive 2025-07-28

Jiaxin Shi, Kehang Han, Zhe Wang, Arnaud Doucet, Michalis K. Titsias

Masked (or absorbing) diffusion is actively explored as an alternative to autoregressive models for generative modeling of discrete data. However, existing work in this area has been hindered by unnecessarily complex model formulations and unclear relationships between different perspectives, leading to suboptimal parameterization, training objectives, and ad hoc adjustments to counteract these issues. In this work, we aim to provide a simple and general framework that unlocks the full potential of masked diffusion models. We show that the continuous-time variational objective of masked diffusion models is a simple weighted integral of cross-entropy losses. Our framework also enables training generalized masked diffusion models with state-dependent masking schedules. When evaluated by perplexity, our models trained on OpenWebText surpass prior diffusion language models at GPT-2 scale and demonstrate superior performance on 4 out of 5 zero-shot language modeling tasks. Furthermore, our models vastly outperform previous discrete diffusion models on pixel-level image modeling, achieving 2.75 (CIFAR-10) and 3.40 (ImageNet 64x64) bits per dimension that are better than autoregressive models of similar sizes. Our code is available at https://github.com/google-deepmind/md4.

PaperPDFCodeCode Syntology ranMethod as code

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 21 code samples harvested from 1 repository linked to this paper: 6 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 1 honoured, 0 violated, 1 with no contract checked; 4 where Syntology's instrument failed) · 15 unverified. Where Syntology's instrument failed: 1 where our contract draft was wrong; 3 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: official repository: 21 samples from 1 repository, 6 ran (2 with no instrument failure; 4 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 2026-09-22 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

google-deepmind/md4 officialmentioned in paperjaxSyntology: 21 samples, 6 ran (4 instrument-failed) · 15 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

21 samples harvested: 6 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 1 honoured, 0 violated, 1 with no contract checked; 4 where Syntology's instrument failed) · 15 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.

1ran · honoured contract
1ran · our draft was wrong
3ran; the contract check could not run
1ran
15unverified

Licence: 0 of the 21 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 google-deepmind/md4. “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.

google-deepmind/md4 official repository 21 samples, 6 ran (4 instrument-failed) · 15 unverified
MaskingSchedule md4/models/diffusion/md4.py ran · metamorphic tier: invariant fingerprinted Apache-2.0 (permissive) · 66a81c44aa3c341b · report fixture: signature t:ArrayImpl/1/float · returned ArrayImpl · level recorded 2026-09-22
get_timestep_embedding in 138 papers
census: 11 implementations from 15 papers, 5 distinct outputs across 2 buckets, one shared input per bucket
md4/models/diffusion/md4.py ran; the contract check could not run (RAISES) fingerprinted Apache-2.0 (permissive) · 84db74ee09e0bda4 · report fixture: signature timesteps:ArrayImpl/1/int, embedding_dim:int/0/int, dtype:str/0/str · returned ArrayImpl · level recorded 2026-09-22
jax_unstack md4/models/diffusion/md4.py ran · our draft was wrong fingerprinted Apache-2.0 (permissive) · b6960d9da81c946a · report fixture: signature x:ArrayImpl/3/int, axis:int/0/int · returned list · level recorded 2026-09-22
precompute_freqs_cis in 112 papers md4/models/diffusion/md4.py ran · honoured contract fingerprinted Apache-2.0 (permissive) · 8873a5201399b10b · report fixture: signature dim:int/0/int, end:int/0/int, theta:float/0/float · returned tuple · level recorded 2026-09-22
repeat_kv in 244 papers
census: 15 implementations from 60 papers, 2 distinct outputs on one shared input
md4/models/diffusion/md4.py ran; the contract check could not run (RAISES) fingerprinted Apache-2.0 (permissive) · 4a1c75a9a5f46c4f · report fixture: signature x:ArrayImpl/4/float, n_rep:int/0/int · returned ArrayImpl · level recorded 2026-09-22
reshape_for_broadcast in 90 papers
census: 6 implementations from 10 papers, 1 distinct output on one shared input this sample: bucket 1 of 1, arg 1: rank 2, kind float, dtype float32 · arg 2: rank 3, kind float, dtype float32
md4/models/diffusion/md4.py ran; the contract check could not run (RAISES) fingerprinted Apache-2.0 (permissive) · 0dc13509bbb03398 · report fixture: signature freqs_cis:Tensor/2/float, x:Tensor/3/float · returned Tensor · level recorded 2026-09-22
CondEmbedding md4/models/diffusion/md4.py unverified Apache-2.0 (permissive) · e341090d86d921b9 · report
DiscreteClassifier md4/models/diffusion/md4.py unverified Apache-2.0 (permissive) · b74decb1e729d273 · report
Dropout1d md4/models/diffusion/md4.py unverified Apache-2.0 (permissive) · d8f8ff9226830df3 · report
FeedForward in 198 papers md4/models/diffusion/md4.py unverified Apache-2.0 (permissive) · 170abc34ff9d4f81 · report
MD4 md4/models/diffusion/md4.py unverified Apache-2.0 (permissive) · 4b01cbf6a3ec21a6 · report
ModelArgs in 13 papers md4/models/diffusion/md4.py unverified Apache-2.0 (permissive) · be5a00527b2e82e0 · report
RMSNorm in 61 papers md4/models/diffusion/md4.py unverified Apache-2.0 (permissive) · a75e172eea6d85f1 · report
SelfAttention in 66 papers md4/models/diffusion/md4.py unverified Apache-2.0 (permissive) · daf4ef755a1978a3 · report
Transformer in 173 papers md4/models/diffusion/md4.py unverified Apache-2.0 (permissive) · 87a69858dd77c308 · report
TransformerBlock in 73 papers md4/models/diffusion/md4.py unverified Apache-2.0 (permissive) · 7a351fb6b29e26ac · report
UNet in 40 papers md4/models/diffusion/md4.py unverified Apache-2.0 (permissive) · 35c08ff52d7e2f6b · report
UNet5DWrapper md4/models/diffusion/md4.py unverified Apache-2.0 (permissive) · a2d820814fa59452 · report
apply_rotary_emb in 121 papers md4/models/diffusion/md4.py unverified Apache-2.0 (permissive) · 582256db3844c9de · report

Method as code Syntology

What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it.

1 assembly on this paper: 1 harvested, 0 generated.

harvested Assembly for Simplified and Generalized Masked Diffusion (graph method: Simplified and Generalized Masked Diffusion). 21 bricks; individually on a fixture: 6 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 1 honoured, 0 violated, 1 with no contract checked; 4 where Syntology's instrument failed) · 15 unverified. The assembly has not been run end-to-end.

  1. MD4 entry · level 0 · unverified · 4b01cbf6a3ec21a6
  2. DiscreteClassifier helper · level 0 · unverified · b74decb1e729d273
  3. MaskingSchedule scheduler · level 1 · ran · 66a81c44aa3c341b
  4. CondEmbedding helper · level 0 · unverified · e341090d86d921b9
  5. ModelArgs helper · level 0 · unverified · be5a00527b2e82e0
  6. Transformer helper · level 0 · unverified · 87a69858dd77c308
  7. UNet5DWrapper model · level 0 · unverified · a2d820814fa59452
  8. get_timestep_embedding helper · level 1 · ran; the contract check could not run (RAISES) · 84db74ee09e0bda4
  9. RMSNorm helper · level 0 · unverified · a75e172eea6d85f1
  10. TransformerBlock helper · level 0 · unverified · 7a351fb6b29e26ac
  11. precompute_freqs_cis helper · level 1 · ran · honoured contract · 8873a5201399b10b
  12. UNet helper · level 0 · unverified · 35c08ff52d7e2f6b
  13. Attention helper · level 0 · unverified · b5d57c4d33c17740
  14. FeedForward helper · level 0 · unverified · 170abc34ff9d4f81
  15. ResBlock helper · level 0 · unverified · 2db631aa126ebf6e
  16. SelfAttention helper · level 0 · unverified · daf4ef755a1978a3
  17. Dropout1d helper · level 0 · unverified · d8f8ff9226830df3
  18. apply_rotary_emb helper · level 0 · unverified · 582256db3844c9de
  19. repeat_kv helper · level 1 · ran; the contract check could not run (RAISES) · 4a1c75a9a5f46c4f
  20. jax_unstack helper · level 1 · ran · our draft was wrong · b6960d9da81c946a
  21. reshape_for_broadcast helper · level 1 · ran; the contract check could not run (RAISES) · 0dc13509bbb03398

Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="2406.04329") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Language ModelingLanguage Modelling

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Language Modelling OpenWebText GenMD4 eval_perplexity 21.80 #10 of 12 Archive leaderboard report
Language Modelling OpenWebText GenMD4 parameters 131M #10 of 12 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

AdamAttentionAttention DropoutBPECosine AnnealingDense ConnectionsDiffusionDiscriminative Fine-TuningDropoutGPT-2HOCLayer NormalizationLinear LayerLinear Warmup With Cosine AnnealingMulti-Head AttentionResidual ConnectionSoftmaxWeight Decay

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