Greedy Search with Probabilistic N-gram Matching for Neural Machine Translation

Chenze Shao; Xilin Chen; Yang Feng

RECORD

Greedy Search with Probabilistic N-gram Matching for Neural Machine Translation

Chenze Shao, Yang Feng, Xilin Chen

Computation and Language · Submitted 10 Sep 2018

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

The authors' main claim Experiment results on the NIST Chinese-to-English translation tasks show that our method significantly outperforms the reinforcement-based algorithms and achieves an improvement of 1.5 BLEU points on average over a strong baseline system.

how it is stated · asserted 1.00 read the record →

Comment (0) Read on arXiv ↗

Cite ▾

arXiv ID1809.03132 (version 1)

Pith IDQRQKHGEV, Pith's permanent record of this paper and its review

CitationChenze Shao, Xilin Chen, Yang Feng. "Greedy Search with Probabilistic N-gram Matching for Neural Machine Translation." arXiv:1809.03132 (2018). https://pith.science/paper/QRQKHGEV

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, 4,338 input tokens.

What the abstract itself states, as probabilities: claims first 0.05 · states limitation 0.10 · human participants 0.03 · controlled comparison 0.94 · randomized assignment 0.01 · states quantitative result 0.99 · compares with prior results 0.98 · mentions code or data availability 0.02.

The abstract, sentence by sentence

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

Neural machine translation (NMT) models are usually trained with the word-level loss using the teacher forcing algorithm, which not only evaluates the translation improperly but also suffers from exposure bias. Sequence-level training under the reinforcement framework can mitigate the problems of the word-level loss, but its performance is unstable due to the high variance of the gradient estimation. On these grounds, we present a method with a differentiable sequence-level training objective based on probabilistic n-gram matching which can avoid the reinforcement framework. In addition, this method performs greedy search in the training which uses the predicted words as context just as at inference to alleviate the problem of exposure bias. Experiment results on the NIST Chinese-to-English translation tasks show that our method significantly outperforms the reinforcement-based algorithms and achieves an improvement of 1.5 BLEU points on average over a strong baseline system.

background contribution method result interpretation main claim

What the classifier reports for each sentence. The number is the probability it gave the label.
#	sentence	role	the authors' own work	states the main conclusion	strength
s0	Neural machine translation (NMT) models are usually trained with the word-level loss using the teacher forcing algorithm, which not only evaluates the translation improperly but also suffers from exposure bias.	background 0.99	0.06	0.02	asserted 0.99
s1	Sequence-level training under the reinforcement framework can mitigate the problems of the word-level loss, but its performance is unstable due to the high variance of the gradient estimation.	background 0.92	0.08	0.02	asserted 0.99
s2	On these grounds, we present a method with a differentiable sequence-level training objective based on probabilistic n-gram matching which can avoid the reinforcement framework.	contribution 0.91	0.90	0.20	asserted 0.97
s3	In addition, this method performs greedy search in the training which uses the predicted words as context just as at inference to alleviate the problem of exposure bias.	method 0.94	0.87	0.05	asserted 0.99
s4	Experiment results on the NIST Chinese-to-English translation tasks show that our method significantly outperforms the reinforcement-based algorithms and achieves an improvement of 1.5 BLEU points on average over a strong baseline system.	result 1.00	0.97	0.83	asserted 1.00

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

Neural machine translation (NMT) models are usually trained with the word-level loss using the teacher forcing algorithm, which not only evaluates the translation improperly but also suffers from exposure bias. Sequence-level training under the reinforcement framework can mitigate the problems of the word-level loss, but its performance is unstable due to the high variance of the gradient estimation. On these grounds, we present a method with a differentiable sequence-level training objective based on probabilistic n-gram matching which can avoid the reinforcement framework. In addition, this method performs greedy search in the training which uses the predicted words as context just as at inference to alleviate the problem of exposure bias. Experiment results on the NIST Chinese-to-English translation tasks show that our method significantly outperforms the reinforcement-based algorithms and achieves an improvement of 1.5 BLEU points on average over a strong baseline system.

Human review

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

How the literature uses it

3 arXiv papers cite this in 3 sentences

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

Papers citing it · 2 sentences without a confident label

Cites it as background 1 paper

$N$-gram based methods have been widely utilized in natural language processing tasks, including language modeling [ref] , text classification [ref] and scoring [ref] , statistical machine translation [ref] and neural machine translation [ref] , etc.

N-Gram Nearest Neighbor Machine Translation · N-Gram Methods in Machine Translation · 0.99

What it cites · 10 sentences without a confident label

Uses the method or tool of 2 papers

We apply our method to an attention-based NMT system [ref] implemented by Pytorch.

Neural Machine Translation by Jointly Learning to Align and Translate · Settings · 1.00
In addition, the learning rate is adjusted by adadelta optimizer [ref] with $ρ=0.95$ and $ϵ=1e-6$.

ADADELTA: An Adaptive Learning Rate Method · Settings · 1.00

Cites as background 7 papers

Neural machine translation (NMT) [ref] has now achieved impressive performance [ref] and draws more attention.

Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation · Introduction · 0.98
Neural machine translation (NMT) [ref] has now achieved impressive performance [ref] and draws more attention.

Neural Machine Translation by Jointly Learning to Align and Translate · Introduction · 0.98
Neural machine translation (NMT) [ref] has now achieved impressive performance [ref] and draws more attention.

Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation · Introduction · 0.98

Discussion

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