Papers › Cascaded Head-colliding Attention

Cascaded Head-colliding Attention

31 May 2021ACL 2021 5arXiv:2105.14850archive 2025-07-28

Lin Zheng, Zhiyong Wu, Lingpeng Kong

Transformers have advanced the field of natural language processing (NLP) on a variety of important tasks. At the cornerstone of the Transformer architecture is the multi-head attention (MHA) mechanism which models pairwise interactions between the elements of the sequence. Despite its massive success, the current framework ignores interactions among different heads, leading to the problem that many of the heads are redundant in practice, which greatly wastes the capacity of the model. To improve parameter efficiency, we re-formulate the MHA as a latent variable model from a probabilistic perspective. We present cascaded head-colliding attention (CODA) which explicitly models the interactions between attention heads through a hierarchical variational distribution. We conduct extensive experiments and demonstrate that CODA outperforms the transformer baseline, by 0.6 perplexity on Wikitext-103 in language modeling, and by 0.6 BLEU on WMT14 EN-DE in machine translation, due to its improvements on the parameter efficiency. (Our implementation is publicly available at https://github.com/LZhengisme/CODA.)

PaperPDFConference PDFCodeCode 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 10 code samples: 8 from 2 repositories linked to this paper, 2 identical to code harvested from another paper's repository. Across all 10: 9 ran (of which 4 constructed an object rather than computing a result; 4 with no instrument failure: 1 honoured, 1 violated, 2 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified. Where Syntology's instrument failed: 4 where our contract draft was wrong; 1 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: official repository: 3 samples from 1 repository, 3 ran (0 with no instrument failure; 3 where Syntology's instrument failed); found in paper text by Syntology: 5 samples from 1 repository, 4 ran (3 with no instrument failure; 1 where Syntology's instrument failed); 2 identical to code first harvested elsewhere. 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 between 2026-09-22 and 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 2 of the 9 samples that ran have no recorded date.

LZhengisme/CODA officialmentioned in papermentioned on GitHubpytorchMITSyntology: 3 samples, 3 ran (3 instrument-failed) · 0 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

10 samples harvested: 9 ran (of which 4 constructed an object rather than computing a result; 4 with no instrument failure: 1 honoured, 1 violated, 2 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

1ran · honoured contract
1ran · violated contract
4ran · our draft was wrong
1ran; the contract check could not run
2ran
1unverified

Licence: 5 of the 10 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 2 repositories linked to this paper, official or community; the list below is grouped by repository, each group headed by its name and what it is to the paper. 2 samples are listed last, in a group without a repository. “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.

LZhengisme/CODA official repository 3 samples, 3 ran (3 instrument-failed) · 0 unverified
Embedding in 90 papers fairseq/models/coda_transformer.py ran · our draft was wrong MIT recorded; this copy not marked cleared · pointer only · 96cbb5e9ca5b6be0 · report fixture: signature num_embeddings:int/0/int, embedding_dim:int/0/int, padding_idx:int/0/int · returned Embedding · level recorded, date not recorded
Linear in 127 papers fairseq/models/coda_transformer.py ran · our draft was wrong MIT recorded; this copy not marked cleared · pointer only · 8cd8cb0d1e9e63e4 · report fixture: signature in_features:int/0/int, out_features:int/0/int, bias:bool/0/bool · returned Linear · level recorded, date not recorded
safe_readline in 67 papers fairseq/binarizer.py ran · our draft was wrong MIT recorded; this copy not marked cleared · pointer only · 82bdf940665aeeb4 · report fixture: signature f:StringIO/0/? · returned str · level recorded 2026-09-24
pytorch/fairseq found in paper text by Syntology 5 samples, 4 ran (1 instrument-failed) · 1 unverified
Kmeans fairseq/modules/kmeans_attention.py ran MIT (permissive) · 9f73b0c0fce2870d · report fixture: returned Kmeans · level recorded 2026-09-22
KmeansAttention fairseq/modules/kmeans_attention.py ran MIT (permissive) · 70d8f18adccf84c1 · report fixture: returned KmeansAttention · level recorded 2026-09-22
ema in 15 papers
census: 4 implementations from 7 papers, 2 distinct outputs across 2 buckets, one shared input per bucket this sample: bucket 1 of 2, arg 1: rank 2, kind float, dtype float32 · arg 2: rank 2, kind float, dtype float32
fairseq/modules/kmeans_attention.py ran · honoured contract fingerprinted MIT (permissive) · 62b973828672478f · report fixture: signature old:Tensor/2/float, new:Tensor/2/float, decay:float/0/float · returned Tensor · level recorded 2026-09-22
kmeans_iter fairseq/modules/kmeans_attention.py ran; the contract check could not run (RAISES) MIT (permissive) · dfbee029313fac6a · report fixture: signature x:Tensor/4/float, means:Tensor/3/float, buckets:NoneType/0/? · returned Tensor · level recorded 2026-09-22
ema_inplace fairseq/modules/kmeans_attention.py unverified MIT (permissive) · b46ee69b75929ac8 · report
Repository not located 2 samples, 2 ran (1 instrument-failed) · 0 unverified
2 samples are identical to code Syntology harvested from another paper's repository; this paper's own copy was not located, so no repository is named and no licence is recorded.
default in 430 papers repository not located ran · violated contract fingerprinted licence of this copy not recorded · 27ab3e63c68fe18a · report fixture: signature x:int/0/int, d:str/0/str · returned int · level recorded 2026-09-22
is_empty in 15 papers repository not located ran · our draft was wrong fingerprinted licence of this copy not recorded · a91aa4d3b29e189e · report fixture: signature t:Tensor/2/float · returned bool · level recorded 2026-09-22

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 Cascaded Head-colliding Attention (CODA) (graph method: Cascaded Head-colliding Attention (CODA)). 17 bricks; individually on a fixture: 16 ran (of which 2 constructed an object rather than computing a result; 5 with no instrument failure: 1 honoured, 2 violated, 2 with no contract checked; 11 where Syntology's instrument failed) · 1 unverified. The assembly has not been run end-to-end.

  1. KmeansAttention entry · level 1 · ran · 70d8f18adccf84c1
  2. Kmeans helper · level 1 · ran · 9f73b0c0fce2870d
  3. batched_index_select helper · level 1 · ran · our draft was wrong · 098a5dd15c0b4167
  4. default helper · level 1 · ran · violated contract · 27ab3e63c68fe18a
  5. distribution helper · level 1 · ran · our draft was wrong · c88722608edfbb70
  6. exists helper · level 1 · ran · violated contract · aa5486a3650902d8
  7. expand_dim helper · level 1 · ran · our draft was wrong · 270b669b4662f5c1
  8. max_neg_value helper · level 1 · ran · our draft was wrong · b209878199e80c50
  9. scatter_mean helper · level 1 · ran · our draft was wrong · e437fa79b7c02bfb
  10. split_at_index helper · level 1 · ran · our draft was wrong · 79eb67938353bd45
  11. dists_and_buckets helper · level 1 · ran · our draft was wrong · 45afc9429eea1b63
  12. ema helper · level 1 · ran · honoured contract · 62b973828672478f
  13. ema_inplace helper · level 0 · unverified · b46ee69b75929ac8
  14. kmeans_iter helper · level 1 · ran; the contract check could not run (RAISES) · dfbee029313fac6a
  15. similarity helper · level 1 · ran · our draft was wrong · 4d588005f814f46d
  16. is_empty helper · level 1 · ran · our draft was wrong · a91aa4d3b29e189e
  17. batched_bincount helper · level 1 · ran · our draft was wrong · 3c9e316e182b4d7d

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="2105.14850") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Language ModelingLanguage ModellingMachine TranslationTranslation

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

Absolute Position EncodingsAdamAttentionBPEDense ConnectionsDropoutLabel SmoothingLayer NormalizationLinear LayerMulti-Head AttentionPosition-Wise Feed-Forward LayerResidual ConnectionSoftmaxTransformer

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