Pith. Sign in

RECORD

Nearest Neighbor Non-autoregressive Text Generation

PaperPDF from arXiv

data files

Paper breakdownThe authors' claims, conditions and measurements, as JSON Signed receiptProof this review has not been altered
X Bluesky LinkedIn Reddit Hacker News
Ayana Niwa, Sho Takase, Naoaki Okazaki

Computation and Language · Submitted 26 Aug 2022

abstract compiled · 2026-09-23 · no language-model review

The authors' main claim Experimental results show that the proposed method (NeighborEdit) achieves higher translation quality (1.69 points higher than the vanilla Transformer) with fewer decoding iterations (one-eighteenth fewer iterations) on the JRC-Acquis En-De dataset, the common benchmark dataset for machine translation using nearest neighbors.

how it is stated · asserted 1.00 read the record →

Comment (0) Read on arXiv ↗
Cite ▾
arXiv ID2208.12496 (version 1)
Pith IDM7H5ZRT3, Pith's permanent record of this paper and its review
CitationAyana Niwa, Naoaki Okazaki, Sho Takase. "Nearest Neighbor Non-autoregressive Text Generation." arXiv:2208.12496 (2022). https://pith.science/paper/M7H5ZRT3
Claim paper

The compiled record

The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.

Open the compiled record

Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-09-23: 1 request, 6,109 input tokens.

What the abstract itself states, as probabilities: claims first 0.09 · states limitation 0.11 · human participants 0.03 · controlled comparison 0.96 · randomized assignment 0.01 · states quantitative result 0.99 · compares with prior results 0.97 · mentions code or data availability 0.02.

The abstract, sentence by sentence

arXiv listing abstract v1. Results and the main claim are underlined; .

Non-autoregressive (NAR) models can generate sentences with less computation than autoregressive models but sacrifice generation quality. Previous studies addressed this issue through iterative decoding. This study proposes using nearest neighbors as the initial state of an NAR decoder and editing them iteratively. We present a novel training strategy to learn the edit operations on neighbors to improve NAR text generation. Experimental results show that the proposed method (NeighborEdit) achieves higher translation quality (1.69 points higher than the vanilla Transformer) with fewer decoding iterations (one-eighteenth fewer iterations) on the JRC-Acquis En-De dataset, the common benchmark dataset for machine translation using nearest neighbors. We also confirm the effectiveness of the proposed method on a data-to-text task (WikiBio). In addition, the proposed method outperforms an NAR baseline on the WMT'14 En-De dataset. We also report analysis on neighbor examples used in the proposed method.

background contribution method result interpretation main claim

What the classifier reports for each sentence. The number is the probability it gave the label.
#sentencerolethe authors' own workstates the main conclusionstrength
s0 Non-autoregressive (NAR) models can generate sentences with less computation than autoregressive models but sacrifice generation quality. background 1.00 0.07 0.03 asserted 0.99
s1 Previous studies addressed this issue through iterative decoding. background 1.00 0.03 0.02 asserted 1.00
s2 This study proposes using nearest neighbors as the initial state of an NAR decoder and editing them iteratively. contribution 0.50 0.68 0.07 asserted 0.98
s3 We present a novel training strategy to learn the edit operations on neighbors to improve NAR text generation. contribution 0.62 0.88 0.06 asserted 0.93
s4 Experimental results show that the proposed method (NeighborEdit) achieves higher translation quality (1.69 points higher than the vanilla Transformer) with fewer decoding iterations (one-eighteenth fewer iterations) on the JRC-Acquis En-De dataset, the common benchmark dataset for machine translation using nearest neighbors. result 1.00 0.96 0.75 asserted 1.00
s5 We also confirm the effectiveness of the proposed method on a data-to-text task (WikiBio). result 0.99 0.95 0.08 asserted 1.00
s6 In addition, the proposed method outperforms an NAR baseline on the WMT'14 En-De dataset. result 1.00 0.95 0.15 asserted 1.00
s7 We also report analysis on neighbor examples used in the proposed method. contribution 0.95 0.88 0.02 unclear 0.60

What kind of paper the classifier reports this is: method or tool 1.00 · evidence from observation or experiment 0.99.

What was examined, and what was not

sources read
  • arXiv listing abstract v1
not examined
  • full text
  • figures and tables
  • references and citing sentences
  • relations to other papers
levels run
  • P abstract
not run
  • S sections
  • T body sentences
  • C citations
  • R cross-paper relations

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. Every number in the measurements table is the authors', checked by code. Record as JSON

The paper
Original abstract

Non-autoregressive (NAR) models can generate sentences with less computation than autoregressive models but sacrifice generation quality. Previous studies addressed this issue through iterative decoding. This study proposes using nearest neighbors as the initial state of an NAR decoder and editing them iteratively. We present a novel training strategy to learn the edit operations on neighbors to improve NAR text generation. Experimental results show that the proposed method (NeighborEdit) achieves higher translation quality (1.69 points higher than the vanilla Transformer) with fewer decoding iterations (one-eighteenth fewer iterations) on the JRC-Acquis En-De dataset, the common benchmark dataset for machine translation using nearest neighbors. We also confirm the effectiveness of the proposed method on a data-to-text task (WikiBio). In addition, the proposed method outperforms an NAR baseline on the WMT'14 En-De dataset. We also report analysis on neighbor examples used in the proposed method.

Human review

How Pith checks papers → · Full audit of this paper →

References
How the literature uses it

What this paper does with the arXiv papers it cites

Each citing sentence is taken from the citing paper's LaTeX source and typed by Pith's classifier. A label appears at probability 0.8 or more (0.95 for support), a contest at 0.7 or more; the quoted words are the citing authors'. How each label was checked · JSON

What it cites

Cites as background 1 paper

Discussion

Know something about this paper? Leave a comment, or highlight any passage above to comment on it.