PyTorchKR Papers

최신 논문

총 5,206편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2004.01694공식 코드 · Papers with Code

A Set of Recommendations for Assessing Human-Machine Parity in Language Translation

The quality of machine translation has increased remarkably over the past years, to the degree that it was found to be indistinguishable from professional human translation in a number of empirical investigations. We reassess Hassan et al.'s 2018 investigation into Chinese to English news translation, showing that the finding of human-machine parity was owed to weaknesses in the evaluation design - which is currently considered best practice in the field. We show that the professional human translations contained significantly fewer errors, and that perceived quality in human evaluation depends on the choice of raters, the availability of linguistic context, and the creation of reference translations. Our results call for revisiting current best practices to assess strong machine translation systems in general and human-machine parity in particular, for which we offer a set of recommendations based on our empirical findings.

arXiv:2004.01655

Aligned Cross Entropy for Non-Autoregressive Machine Translation

Non-autoregressive machine translation models significantly speed up decoding by allowing for parallel prediction of the entire target sequence. However, modeling word order is more challenging due to the lack of autoregressive factors in the model. This difficultly is compounded during training with cross entropy loss, which can highly penalize small shifts in word order. In this paper, we propose aligned cross entropy (AXE) as an alternative loss function for training of non-autoregressive models. AXE uses a differentiable dynamic program to assign loss based on the best possible monotonic alignment between target tokens and model predictions. AXE-based training of conditional masked language models (CMLMs) substantially improves performance on major WMT benchmarks, while setting a new state of the art for non-autoregressive models.

arXiv:2004.00648

Igbo-English Machine Translation: An Evaluation Benchmark

Although researchers and practitioners are pushing the boundaries and enhancing the capacities of NLP tools and methods, works on African languages are lagging. A lot of focus on well resourced languages such as English, Japanese, German, French, Russian, Mandarin Chinese etc. Over 97% of the world's 7000 languages, including African languages, are low resourced for NLP i.e. they have little or no data, tools, and techniques for NLP research. For instance, only 5 out of 2965, 0.19% authors of full text papers in the ACL Anthology extracted from the 5 major conferences in 2018 ACL, NAACL, EMNLP, COLING and CoNLL, are affiliated to African institutions. In this work, we discuss our effort toward building a standard machine translation benchmark dataset for Igbo, one of the 3 major Nigerian languages. Igbo is spoken by more than 50 million people globally with over 50% of the speakers are in southeastern Nigeria. Igbo is low resourced although there have been some efforts toward developing IgboNLP such as part of speech tagging and diacritic restoration

arXiv:2004.00345공식 코드 · Papers with Code

Editable Neural Networks

These days deep neural networks are ubiquitously used in a wide range of tasks, from image classification and machine translation to face identification and self-driving cars. In many applications, a single model error can lead to devastating financial, reputational and even life-threatening consequences. Therefore, it is crucially important to correct model mistakes quickly as they appear. In this work, we investigate the problem of neural network editing $-$ how one can efficiently patch a mistake of the model on a particular sample, without influencing the model behavior on other samples. Namely, we propose Editable Training, a model-agnostic training technique that encourages fast editing of the trained model. We empirically demonstrate the effectiveness of this method on large-scale image classification and machine translation tasks.

arXiv:2004.00068

Assessing Human Translations from French to Bambara for Machine Learning: a Pilot Study

We present novel methods for assessing the quality of human-translated aligned texts for learning machine translation models of under-resourced languages. Malian university students translated French texts, producing either written or oral translations to Bambara. Our results suggest that similar quality can be obtained from either written or spoken translations for certain kinds of texts. They also suggest specific instructions that human translators should be given in order to improve the quality of their work.

arXiv:2003.14402공식 코드 · Papers with Code

Low Resource Neural Machine Translation: A Benchmark for Five African Languages

Recent advents in Neural Machine Translation (NMT) have shown improvements in low-resource language (LRL) translation tasks. In this work, we benchmark NMT between English and five African LRL pairs (Swahili, Amharic, Tigrigna, Oromo, Somali [SATOS]). We collected the available resources on the SATOS languages to evaluate the current state of NMT for LRLs. Our evaluation, comparing a baseline single language pair NMT model against semi-supervised learning, transfer learning, and multilingual modeling, shows significant performance improvements both in the En-LRL and LRL-En directions. In terms of averaged BLEU score, the multilingual approach shows the largest gains, up to +5 points, in six out of ten translation directions. To demonstrate the generalization capability of each model, we also report results on multi-domain test sets. We release the standardized experimental data and the test sets for future works addressing the challenges of NMT in under-resourced settings, in particular for the SATOS languages.

arXiv:2003.14386

Evaluating Amharic Machine Translation

Machine translation (MT) systems are now able to provide very accurate results for high resource language pairs. However, for many low resource languages, MT is still under active research. In this paper, we develop and share a dataset to automatically evaluate the quality of MT systems for Amharic. We compare two commercially available MT systems that support translation of Amharic to and from English to assess the current state of MT for Amharic. The BLEU score results show that the results for Amharic translation are promising but still low. We hope that this dataset will be useful to the research community both in academia and industry as a benchmark to evaluate Amharic MT systems.

arXiv:2003.14324

On the Integration of LinguisticFeatures into Statistical and Neural Machine Translation

New machine translations (MT) technologies are emerging rapidly and with them, bold claims of achieving human parity such as: (i) the results produced approach "accuracy achieved by average bilingual human translators" (Wu et al., 2017b) or (ii) the "translation quality is at human parity when compared to professional human translators" (Hassan et al., 2018) have seen the light of day (Laubli et al., 2018). Aside from the fact that many of these papers craft their own definition of human parity, these sensational claims are often not supported by a complete analysis of all aspects involved in translation. Establishing the discrepancies between the strengths of statistical approaches to MT and the way humans translate has been the starting point of our research. By looking at MT output and linguistic theory, we were able to identify some remaining issues. The problems range from simple number and gender agreement errors to more complex phenomena such as the correct translation of aspectual values and tenses. Our experiments confirm, along with other studies (Bentivogli et al., 2016), that neural MT has surpassed statistical MT in many aspects. However, some problems remain and others have emerged. We cover a series of problems related to the integration of specific linguistic features into statistical and neural MT, aiming to analyse and provide a solution to some of them. Our work focuses on addressing three main research questions that revolve around the complex relationship between linguistics and MT in general. We identify linguistic information that is lacking in order for automatic translation systems to produce more accurate translations and integrate additional features into the existing pipelines. We identify overgeneralization or 'algorithmic bias' as a potential drawback of neural MT and link it to many of the remaining linguistic issues.

arXiv:2003.12738공식 코드 · Papers with Code

Variational Transformers for Diverse Response Generation

Despite the great promise of Transformers in many sequence modeling tasks (e.g., machine translation), their deterministic nature hinders them from generalizing to high entropy tasks such as dialogue response generation. Previous work proposes to capture the variability of dialogue responses with a recurrent neural network (RNN)-based conditional variational autoencoder (CVAE). However, the autoregressive computation of the RNN limits the training efficiency. Therefore, we propose the Variational Transformer (VT), a variational self-attentive feed-forward sequence model. The VT combines the parallelizability and global receptive field of the Transformer with the variational nature of the CVAE by incorporating stochastic latent variables into Transformers. We explore two types of the VT: 1) modeling the discourse-level diversity with a global latent variable; and 2) augmenting the Transformer decoder with a sequence of fine-grained latent variables. Then, the proposed models are evaluated on three conversational datasets with both automatic metric and human evaluation. The experimental results show that our models improve standard Transformers and other baselines in terms of diversity, semantic relevance, and human judgment.

arXiv:2003.12660공식 코드 · Papers with Code

Towards Supervised and Unsupervised Neural Machine Translation Baselines for Nigerian Pidgin

Nigerian Pidgin is arguably the most widely spoken language in Nigeria. Variants of this language are also spoken across West and Central Africa, making it a very important language. This work aims to establish supervised and unsupervised neural machine translation (NMT) baselines between English and Nigerian Pidgin. We implement and compare NMT models with different tokenization methods, creating a solid foundation for future works.

arXiv:2003.13205

Learning Contextualized Sentence Representations for Document-Level Neural Machine Translation

Document-level machine translation incorporates inter-sentential dependencies into the translation of a source sentence. In this paper, we propose a new framework to model cross-sentence dependencies by training neural machine translation (NMT) to predict both the target translation and surrounding sentences of a source sentence. By enforcing the NMT model to predict source context, we want the model to learn "contextualized" source sentence representations that capture document-level dependencies on the source side. We further propose two different methods to learn and integrate such contextualized sentence embeddings into NMT: a joint training method that jointly trains an NMT model with the source context prediction model and a pre-training & fine-tuning method that pretrains the source context prediction model on a large-scale monolingual document corpus and then fine-tunes it with the NMT model. Experiments on Chinese-English and English-German translation show that both methods can substantially improve the translation quality over a strong document-level Transformer baseline.

arXiv:2003.13619

Repository for Reusing Artifacts of Artificial Neural Networks

Artificial Neural Networks (ANNs) replaced conventional software systems in various domains such as machine translation, natural language processing, and image processing. So, why do we need an repository for artificial neural networks? Those systems are developed with labeled data and we have strong dependencies between the data that is used for training and testing our network. Another challenge is the data quality as well as reuse-ability. There we are trying to apply concepts from classic software engineering that is not limited to the model, while data and code haven't been dealt with mostly in other projects. The first question that comes to mind might be, why don't we use GitHub, a well known widely spread tool for reuse, for our issue. And the reason why is that GitHub, although very good in its class is not developed for machine learning appliances and focuses more on software reuse. In addition to that GitHub does not allow to execute the code directly on the platform which would be very convenient for collaborative work on one project.

arXiv:2003.12111공식 코드 · Papers with Code

FFR V1.0: Fon-French Neural Machine Translation

Africa has the highest linguistic diversity in the world. On account of the importance of language to communication, and the importance of reliable, powerful and accurate machine translation models in modern inter-cultural communication, there have been (and still are) efforts to create state-of-the-art translation models for the many African languages. However, the low-resources, diacritical and tonal complexities of African languages are major issues facing African NLP today. The FFR is a major step towards creating a robust translation model from Fon, a very low-resource and tonal language, to French, for research and public use. In this paper, we describe our pilot project: the creation of a large growing corpora for Fon-to-French translations and our FFR v1.0 model, trained on this dataset. The dataset and model are made publicly available.

arXiv:2003.11529

Masakhane -- Machine Translation For Africa

Africa has over 2000 languages. Despite this, African languages account for a small portion of available resources and publications in Natural Language Processing (NLP). This is due to multiple factors, including: a lack of focus from government and funding, discoverability, a lack of community, sheer language complexity, difficulty in reproducing papers and no benchmarks to compare techniques. To begin to address the identified problems, MASAKHANE, an open-source, continent-wide, distributed, online research effort for machine translation for African languages, was founded. In this paper, we discuss our methodology for building the community and spurring research from the African continent, as well as outline the success of the community in terms of addressing the identified problems affecting African NLP.

arXiv:2003.11523공식 코드 · Papers with Code

Tigrinya Neural Machine Translation with Transfer Learning for Humanitarian Response

We report our experiments in building a domain-specific Tigrinya-to-English neural machine translation system. We use transfer learning from other Ge'ez script languages and report an improvement of 1.3 BLEU points over a classic neural baseline. We publish our development pipeline as an open-source library and also provide a demonstration application.

arXiv:2003.10704

Towards Neural Machine Translation for Edoid Languages

Many Nigerian languages have relinquished their previous prestige and purpose in modern society to English and Nigerian Pidgin. For the millions of L1 speakers of indigenous languages, there are inequalities that manifest themselves as unequal access to information, communications, health care, security as well as attenuated participation in political and civic life. To minimize exclusion and promote socio-linguistic and economic empowerment, this work explores the feasibility of Neural Machine Translation (NMT) for the Edoid language family of Southern Nigeria. Using the new JW300 public dataset, we trained and evaluated baseline translation models for four widely spoken languages in this group: \`Ed\'o, \'Es\'an, Urhobo and Isoko. Trained models, code and datasets have been open-sourced to advance future research efforts on Edoid language technology.

arXiv:2003.10687공식 코드 · Papers with Code

Felix: Flexible Text Editing Through Tagging and Insertion

We present Felix --- a flexible text-editing approach for generation, designed to derive the maximum benefit from the ideas of decoding with bi-directional contexts and self-supervised pre-training. In contrast to conventional sequence-to-sequence (seq2seq) models, Felix is efficient in low-resource settings and fast at inference time, while being capable of modeling flexible input-output transformations. We achieve this by decomposing the text-editing task into two sub-tasks: tagging to decide on the subset of input tokens and their order in the output text and insertion to in-fill the missing tokens in the output not present in the input. The tagging model employs a novel Pointer mechanism, while the insertion model is based on a Masked Language Model. Both of these models are chosen to be non-autoregressive to guarantee faster inference. Felix performs favourably when compared to recent text-editing methods and strong seq2seq baselines when evaluated on four NLG tasks: Sentence Fusion, Machine Translation Automatic Post-Editing, Summarization, and Text Simplification.

arXiv:2003.09833

SAC: Accelerating and Structuring Self-Attention via Sparse Adaptive Connection

While the self-attention mechanism has been widely used in a wide variety of tasks, it has the unfortunate property of a quadratic cost with respect to the input length, which makes it difficult to deal with long inputs. In this paper, we present a method for accelerating and structuring self-attentions: Sparse Adaptive Connection (SAC). In SAC, we regard the input sequence as a graph and attention operations are performed between linked nodes. In contrast with previous self-attention models with pre-defined structures (edges), the model learns to construct attention edges to improve task-specific performances. In this way, the model is able to select the most salient nodes and reduce the quadratic complexity regardless of the sequence length. Based on SAC, we show that previous variants of self-attention models are its special cases. Through extensive experiments on neural machine translation, language modeling, graph representation learning and image classification, we demonstrate SAC is competitive with state-of-the-art models while significantly reducing memory cost.

arXiv:2003.09606

A Joint Approach to Compound Splitting and Idiomatic Compound Detection

Applications such as machine translation, speech recognition, and information retrieval require efficient handling of noun compounds as they are one of the possible sources for out-of-vocabulary (OOV) words. In-depth processing of noun compounds requires not only splitting them into smaller components (or even roots) but also the identification of instances that should remain unsplitted as they are of idiomatic nature. We develop a two-fold deep learning-based approach of noun compound splitting and idiomatic compound detection for the German language that we train using a newly collected corpus of annotated German compounds. Our neural noun compound splitter operates on a sub-word level and outperforms the current state of the art by about 5%.

arXiv:2003.09229

Learning to Encode Position for Transformer with Continuous Dynamical Model

We introduce a new way of learning to encode position information for non-recurrent models, such as Transformer models. Unlike RNN and LSTM, which contain inductive bias by loading the input tokens sequentially, non-recurrent models are less sensitive to position. The main reason is that position information among input units is not inherently encoded, i.e., the models are permutation equivalent; this problem justifies why all of the existing models are accompanied by a sinusoidal encoding/embedding layer at the input. However, this solution has clear limitations: the sinusoidal encoding is not flexible enough as it is manually designed and does not contain any learnable parameters, whereas the position embedding restricts the maximum length of input sequences. It is thus desirable to design a new position layer that contains learnable parameters to adjust to different datasets and different architectures. At the same time, we would also like the encodings to extrapolate in accordance with the variable length of inputs. In our proposed solution, we borrow from the recent Neural ODE approach, which may be viewed as a versatile continuous version of a ResNet. This model is capable of modeling many kinds of dynamical systems. We model the evolution of encoded results along position index by such a dynamical system, thereby overcoming the above limitations of existing methods. We evaluate our new position layers on a variety of neural machine translation and language understanding tasks, the experimental results show consistent improvements over the baselines.