PyTorchKR Papers

최신 논문

총 1,781편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2101.00234공식 코드 · Papers with Code

Subformer: Exploring Weight Sharing for Parameter Efficiency in Generative Transformers

Transformers have shown improved performance when compared to previous architectures for sequence processing such as RNNs. Despite their sizeable performance gains, as recently suggested, the model is computationally expensive to train and with a high parameter budget. In light of this, we explore parameter-sharing methods in Transformers with a specific focus on generative models. We perform an analysis of different parameter sharing/reduction methods and develop the Subformer. Our model combines sandwich-style parameter sharing, which overcomes naive cross-layer parameter sharing in generative models, and self-attentive embedding factorization (SAFE). Experiments on machine translation, abstractive summarization and language modeling show that the Subformer can outperform the Transformer even when using significantly fewer parameters.

arXiv:2012.15671공식 코드 · Papers with Code

Vocabulary Learning via Optimal Transport for Neural Machine Translation

The choice of token vocabulary affects the performance of machine translation. This paper aims to figure out what is a good vocabulary and whether one can find the optimal vocabulary without trial training. To answer these questions, we first provide an alternative understanding of the role of vocabulary from the perspective of information theory. Motivated by this, we formulate the quest of vocabularization -- finding the best token dictionary with a proper size -- as an optimal transport (OT) problem. We propose VOLT, a simple and efficient solution without trial training. Empirical results show that VOLT outperforms widely-used vocabularies in diverse scenarios, including WMT-14 English-German and TED's 52 translation directions. For example, VOLT achieves almost 70% vocabulary size reduction and 0.5 BLEU gain on English-German translation. Also, compared to BPE-search, VOLT reduces the search time from 384 GPU hours to 30 GPU hours on English-German translation. Codes are available at https://github.com/Jingjing-NLP/VOLT .

arXiv:2012.15150공식 코드 · Papers with Code

Improving BERT with Syntax-aware Local Attention

Pre-trained Transformer-based neural language models, such as BERT, have achieved remarkable results on varieties of NLP tasks. Recent works have shown that attention-based models can benefit from more focused attention over local regions. Most of them restrict the attention scope within a linear span, or confine to certain tasks such as machine translation and question answering. In this paper, we propose a syntax-aware local attention, where the attention scopes are restrained based on the distances in the syntactic structure. The proposed syntax-aware local attention can be integrated with pretrained language models, such as BERT, to render the model to focus on syntactically relevant words. We conduct experiments on various single-sentence benchmarks, including sentence classification and sequence labeling tasks. Experimental results show consistent gains over BERT on all benchmark datasets. The extensive studies verify that our model achieves better performance owing to more focused attention over syntactically relevant words.

arXiv:2012.15127공식 코드 · Papers with Code

Improving Zero-Shot Translation by Disentangling Positional Information

Multilingual neural machine translation has shown the capability of directly translating between language pairs unseen in training, i.e. zero-shot translation. Despite being conceptually attractive, it often suffers from low output quality. The difficulty of generalizing to new translation directions suggests the model representations are highly specific to those language pairs seen in training. We demonstrate that a main factor causing the language-specific representations is the positional correspondence to input tokens. We show that this can be easily alleviated by removing residual connections in an encoder layer. With this modification, we gain up to 18.5 BLEU points on zero-shot translation while retaining quality on supervised directions. The improvements are particularly prominent between related languages, where our proposed model outperforms pivot-based translation. Moreover, our approach allows easy integration of new languages, which substantially expands translation coverage. By thorough inspections of the hidden layer outputs, we show that our approach indeed leads to more language-independent representations.

arXiv:2012.14768공식 코드 · Papers with Code

Understanding and Improving Encoder Layer Fusion in Sequence-to-Sequence Learning

Encoder layer fusion (EncoderFusion) is a technique to fuse all the encoder layers (instead of the uppermost layer) for sequence-to-sequence (Seq2Seq) models, which has proven effective on various NLP tasks. However, it is still not entirely clear why and when EncoderFusion should work. In this paper, our main contribution is to take a step further in understanding EncoderFusion. Many of previous studies believe that the success of EncoderFusion comes from exploiting surface and syntactic information embedded in lower encoder layers. Unlike them, we find that the encoder embedding layer is more important than other intermediate encoder layers. In addition, the uppermost decoder layer consistently pays more attention to the encoder embedding layer across NLP tasks. Based on this observation, we propose a simple fusion method, SurfaceFusion, by fusing only the encoder embedding layer for the softmax layer. Experimental results show that SurfaceFusion outperforms EncoderFusion on several NLP benchmarks, including machine translation, text summarization, and grammatical error correction. It obtains the state-of-the-art performance on WMT16 Romanian-English and WMT14 English-French translation tasks. Extensive analyses reveal that SurfaceFusion learns more expressive bilingual word embeddings by building a closer relationship between relevant source and target embedding. Source code is freely available at https://github.com/SunbowLiu/SurfaceFusion.

arXiv:2012.15416공식 코드 · Papers with Code

Directed Beam Search: Plug-and-Play Lexically Constrained Language Generation

Large pre-trained language models are capable of generating realistic text. However, controlling these models so that the generated text satisfies lexical constraints, i.e., contains specific words, is a challenging problem. Given that state-of-the-art language models are too large to be trained from scratch in a manageable time, it is desirable to control these models without re-training them. Methods capable of doing this are called plug-and-play. Recent plug-and-play methods have been successful in constraining small bidirectional language models as well as forward models in tasks with a restricted search space, e.g., machine translation. However, controlling large transformer-based models to meet lexical constraints without re-training them remains a challenge. In this work, we propose Directed Beam Search (DBS), a plug-and-play method for lexically constrained language generation. Our method can be applied to any language model, is easy to implement and can be used for general language generation. In our experiments we use DBS to control GPT-2. We demonstrate its performance on keyword-to-phrase generation and we obtain comparable results as a state-of-the-art non-plug-and-play model for lexically constrained story generation.

arXiv:2012.13866공식 코드 · Papers with Code

Learning Light-Weight Translation Models from Deep Transformer

Recently, deep models have shown tremendous improvements in neural machine translation (NMT). However, systems of this kind are computationally expensive and memory intensive. In this paper, we take a natural step towards learning strong but light-weight NMT systems. We proposed a novel group-permutation based knowledge distillation approach to compressing the deep Transformer model into a shallow model. The experimental results on several benchmarks validate the effectiveness of our method. Our compressed model is 8X shallower than the deep model, with almost no loss in BLEU. To further enhance the teacher model, we present a Skipping Sub-Layer method to randomly omit sub-layers to introduce perturbation into training, which achieves a BLEU score of 30.63 on English-German newstest2014. The code is publicly available at https://github.com/libeineu/GPKD.

arXiv:2012.14271공식 코드 · Papers with Code

Towards Fully Automated Manga Translation

We tackle the problem of machine translation of manga, Japanese comics. Manga translation involves two important problems in machine translation: context-aware and multimodal translation. Since text and images are mixed up in an unstructured fashion in Manga, obtaining context from the image is essential for manga translation. However, it is still an open problem how to extract context from image and integrate into MT models. In addition, corpus and benchmarks to train and evaluate such model is currently unavailable. In this paper, we make the following four contributions that establishes the foundation of manga translation research. First, we propose multimodal context-aware translation framework. We are the first to incorporate context information obtained from manga image. It enables us to translate texts in speech bubbles that cannot be translated without using context information (e.g., texts in other speech bubbles, gender of speakers, etc.). Second, for training the model, we propose the approach to automatic corpus construction from pairs of original manga and their translations, by which large parallel corpus can be constructed without any manual labeling. Third, we created a new benchmark to evaluate manga translation. Finally, on top of our proposed methods, we devised a first comprehensive system for fully automated manga translation.

arXiv:2012.12366공식 코드 · Papers with Code

Multi-Head Self-Attention with Role-Guided Masks

The state of the art in learning meaningful semantic representations of words is the Transformer model and its attention mechanisms. Simply put, the attention mechanisms learn to attend to specific parts of the input dispensing recurrence and convolutions. While some of the learned attention heads have been found to play linguistically interpretable roles, they can be redundant or prone to errors. We propose a method to guide the attention heads towards roles identified in prior work as important. We do this by defining role-specific masks to constrain the heads to attend to specific parts of the input, such that different heads are designed to play different roles. Experiments on text classification and machine translation using 7 different datasets show that our method outperforms competitive attention-based, CNN, and RNN baselines.

arXiv:2012.11747공식 코드 · Papers with Code

RealFormer: Transformer Likes Residual Attention

Transformer is the backbone of modern NLP models. In this paper, we propose RealFormer, a simple and generic technique to create Residual Attention Layer Transformer networks that significantly outperform the canonical Transformer and its variants (BERT, ETC, etc.) on a wide spectrum of tasks including Masked Language Modeling, GLUE, SQuAD, Neural Machine Translation, WikiHop, HotpotQA, Natural Questions, and OpenKP. We also observe empirically that RealFormer stabilizes training and leads to models with sparser attention. Source code and pre-trained checkpoints for RealFormer can be found at https://github.com/google-research/google-research/tree/master/realformer.

arXiv:2012.10586공식 코드 · Papers with Code

Finding Sparse Structures for Domain Specific Neural Machine Translation

Neural machine translation often adopts the fine-tuning approach to adapt to specific domains. However, nonrestricted fine-tuning can easily degrade on the general domain and over-fit to the target domain. To mitigate the issue, we propose Prune-Tune, a novel domain adaptation method via gradual pruning. It learns tiny domain-specific sub-networks during fine-tuning on new domains. Prune-Tune alleviates the over-fitting and the degradation problem without model modification. Furthermore, Prune-Tune is able to sequentially learn a single network with multiple disjoint domain-specific sub-networks for multiple domains. Empirical experiment results show that Prune-Tune outperforms several strong competitors in the target domain test set without sacrificing the quality on the general domain in both single and multi-domain settings. The source code and data are available at https://github.com/ohlionel/Prune-Tune.

arXiv:2012.07280공식 코드 · Papers with Code

Contrastive Learning with Adversarial Perturbations for Conditional Text Generation

Recently, sequence-to-sequence (seq2seq) models with the Transformer architecture have achieved remarkable performance on various conditional text generation tasks, such as machine translation. However, most of them are trained with teacher forcing with the ground truth label given at each time step, without being exposed to incorrectly generated tokens during training, which hurts its generalization to unseen inputs, that is known as the "exposure bias" problem. In this work, we propose to mitigate the conditional text generation problem by contrasting positive pairs with negative pairs, such that the model is exposed to various valid or incorrect perturbations of the inputs, for improved generalization. However, training the model with naive contrastive learning framework using random non-target sequences as negative examples is suboptimal, since they are easily distinguishable from the correct output, especially so with models pretrained with large text corpora. Also, generating positive examples requires domain-specific augmentation heuristics which may not generalize over diverse domains. To tackle this problem, we propose a principled method to generate positive and negative samples for contrastive learning of seq2seq models. Specifically, we generate negative examples by adding small perturbations to the input sequence to minimize its conditional likelihood, and positive examples by adding large perturbations while enforcing it to have a high conditional likelihood. Such "hard" positive and negative pairs generated using our method guides the model to better distinguish correct outputs from incorrect ones. We empirically show that our proposed method significantly improves the generalization of the seq2seq on three text generation tasks - machine translation, text summarization, and question generation.

arXiv:2012.07162공식 코드 · Papers with Code

Mask-Align: Self-Supervised Neural Word Alignment

Word alignment, which aims to align translationally equivalent words between source and target sentences, plays an important role in many natural language processing tasks. Current unsupervised neural alignment methods focus on inducing alignments from neural machine translation models, which does not leverage the full context in the target sequence. In this paper, we propose Mask-Align, a self-supervised word alignment model that takes advantage of the full context on the target side. Our model masks out each target token and predicts it conditioned on both source and the remaining target tokens. This two-step process is based on the assumption that the source token contributing most to recovering the masked target token should be aligned. We also introduce an attention variant called leaky attention, which alleviates the problem of unexpected high cross-attention weights on special tokens such as periods. Experiments on four language pairs show that our model outperforms previous unsupervised neural aligners and obtains new state-of-the-art results.

arXiv:2012.06154공식 코드 · Papers with Code

ParsiNLU: A Suite of Language Understanding Challenges for Persian

Despite the progress made in recent years in addressing natural language understanding (NLU) challenges, the majority of this progress remains to be concentrated on resource-rich languages like English. This work focuses on Persian language, one of the widely spoken languages in the world, and yet there are few NLU datasets available for this rich language. The availability of high-quality evaluation datasets is a necessity for reliable assessment of the progress on different NLU tasks and domains. We introduce ParsiNLU, the first benchmark in Persian language that includes a range of high-level tasks -- Reading Comprehension, Textual Entailment, etc. These datasets are collected in a multitude of ways, often involving manual annotations by native speakers. This results in over 14.5$k$ new instances across 6 distinct NLU tasks. Besides, we present the first results on state-of-the-art monolingual and multi-lingual pre-trained language-models on this benchmark and compare them with human performance, which provides valuable insights into our ability to tackle natural language understanding challenges in Persian. We hope ParsiNLU fosters further research and advances in Persian language understanding.

arXiv:2012.06143공식 코드 · Papers with Code

Document-aligned Japanese-English Conversation Parallel Corpus

Sentence-level (SL) machine translation (MT) has reached acceptable quality for many high-resourced languages, but not document-level (DL) MT, which is difficult to 1) train with little amount of DL data; and 2) evaluate, as the main methods and data sets focus on SL evaluation. To address the first issue, we present a document-aligned Japanese-English conversation corpus, including balanced, high-quality business conversation data for tuning and testing. As for the second issue, we manually identify the main areas where SL MT fails to produce adequate translations in lack of context. We then create an evaluation set where these phenomena are annotated to alleviate automatic evaluation of DL systems. We train MT models using our corpus to demonstrate how using context leads to improvements.

arXiv:2012.05879공식 코드 · Papers with Code

Automatic Standardization of Colloquial Persian

The Iranian Persian language has two varieties: standard and colloquial. Most natural language processing tools for Persian assume that the text is in standard form: this assumption is wrong in many real applications especially web content. This paper describes a simple and effective standardization approach based on sequence-to-sequence translation. We design an algorithm for generating artificial parallel colloquial-to-standard data for learning a sequence-to-sequence model. Moreover, we annotate a publicly available evaluation data consisting of 1912 sentences from a diverse set of domains. Our intrinsic evaluation shows a higher BLEU score of 62.8 versus 61.7 compared to an off-the-shelf rule-based standardization model in which the original text has a BLEU score of 46.4. We also show that our model improves English-to-Persian machine translation in scenarios for which the training data is from colloquial Persian with 1.4 absolute BLEU score difference in the development data, and 0.8 in the test data.

arXiv:2012.04964공식 코드 · Papers with Code

On Knowledge Distillation for Direct Speech Translation

Direct speech translation (ST) has shown to be a complex task requiring knowledge transfer from its sub-tasks: automatic speech recognition (ASR) and machine translation (MT). For MT, one of the most promising techniques to transfer knowledge is knowledge distillation. In this paper, we compare the different solutions to distill knowledge in a sequence-to-sequence task like ST. Moreover, we analyze eventual drawbacks of this approach and how to alleviate them maintaining the benefits in terms of translation quality.

arXiv:2012.02975공식 코드 · Papers with Code

Reciprocal Supervised Learning Improves Neural Machine Translation

Despite the recent success on image classification, self-training has only achieved limited gains on structured prediction tasks such as neural machine translation (NMT). This is mainly due to the compositionality of the target space, where the far-away prediction hypotheses lead to the notorious reinforced mistake problem. In this paper, we revisit the utilization of multiple diverse models and present a simple yet effective approach named Reciprocal-Supervised Learning (RSL). RSL first exploits individual models to generate pseudo parallel data, and then cooperatively trains each model on the combined synthetic corpus. RSL leverages the fact that different parameterized models have different inductive biases, and better predictions can be made by jointly exploiting the agreement among each other. Unlike the previous knowledge distillation methods built upon a much stronger teacher, RSL is capable of boosting the accuracy of one model by introducing other comparable or even weaker models. RSL can also be viewed as a more efficient alternative to ensemble. Extensive experiments demonstrate the superior performance of RSL on several benchmarks with significant margins.

arXiv:2012.02420공식 코드 · Papers with Code

Benchmarking Automated Clinical Language Simplification: Dataset, Algorithm, and Evaluation

Patients with low health literacy usually have difficulty understanding medical jargon and the complex structure of professional medical language. Although some studies are proposed to automatically translate expert language into layperson-understandable language, only a few of them focus on both accuracy and readability aspects simultaneously in the clinical domain. Thus, simplification of the clinical language is still a challenging task, but unfortunately, it is not yet fully addressed in previous work. To benchmark this task, we construct a new dataset named MedLane to support the development and evaluation of automated clinical language simplification approaches. Besides, we propose a new model called DECLARE that follows the human annotation procedure and achieves state-of-the-art performance compared with eight strong baselines. To fairly evaluate the performance, we also propose three specific evaluation metrics. Experimental results demonstrate the utility of the annotated MedLane dataset and the effectiveness of the proposed model DECLARE.

arXiv:2012.02030공식 코드 · Papers with Code

Data-Informed Global Sparseness in Attention Mechanisms for Deep Neural Networks

Attention mechanisms play a crucial role in the neural revolution of Natural Language Processing (NLP). With the growth of attention-based models, several pruning techniques have been developed to identify and exploit sparseness, making these models more efficient. Most efforts focus on hard-coding attention patterns or pruning attention weights based on training data. We propose Attention Pruning (AP), a framework that observes attention patterns in a fixed dataset and generates a global sparseness mask. AP saves 90% of attention computation for language modeling and about 50% for machine translation and GLUE tasks, maintaining result quality. Our method reveals important distinctions between self- and cross-attention patterns, guiding future NLP research. Our framework can reduce both latency and memory requirements for any attention-based model, aiding in the development of improved models for existing or new NLP applications. We have demonstrated this with encoder and autoregressive transformer models using Triton GPU kernels and make our code publicly available at https://github.com/irugina/AP.