PyTorchKR Papers

최신 논문

총 5,206편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:1908.10118

Multi-Layer Softmaxing during Training Neural Machine Translation for Flexible Decoding with Fewer Layers

This paper proposes a novel procedure for training an encoder-decoder based deep neural network which compresses NxM models into a single model enabling us to dynamically choose the number of encoder and decoder layers for decoding. Usually, the output of the last layer of the N-layer encoder is fed to the M-layer decoder, and the output of the last decoder layer is used to compute softmax loss. Instead, our method computes a single loss consisting of NxM losses: the softmax loss for the output of each of the M decoder layers derived using the output of each of the N encoder layers. A single model trained by our method can be used for decoding with an arbitrary fewer number of encoder and decoder layers. In practical scenarios, this (a) enables faster decoding with insignificant losses in translation quality and (b) alleviates the need to train NxM models, thereby saving space. We take a case study of neural machine translation and show the advantage and give a cost-benefit analysis of our approach.

arXiv:1908.10090

On NMT Search Errors and Model Errors: Cat Got Your Tongue?

We report on search errors and model errors in neural machine translation (NMT). We present an exact inference procedure for neural sequence models based on a combination of beam search and depth-first search. We use our exact search to find the global best model scores under a Transformer base model for the entire WMT15 English-German test set. Surprisingly, beam search fails to find these global best model scores in most cases, even with a very large beam size of 100. For more than 50% of the sentences, the model in fact assigns its global best score to the empty translation, revealing a massive failure of neural models in properly accounting for adequacy. We show by constraining search with a minimum translation length that at the root of the problem of empty translations lies an inherent bias towards shorter translations. We conclude that vanilla NMT in its current form requires just the right amount of beam search errors, which, from a modelling perspective, is a highly unsatisfactory conclusion indeed, as the model often prefers an empty translation.

arXiv:1908.09324

Multilingual Neural Machine Translation with Language Clustering

Multilingual neural machine translation (NMT), which translates multiple languages using a single model, is of great practical importance due to its advantages in simplifying the training process, reducing online maintenance costs, and enhancing low-resource and zero-shot translation. Given there are thousands of languages in the world and some of them are very different, it is extremely burdensome to handle them all in a single model or use a separate model for each language pair. Therefore, given a fixed resource budget, e.g., the number of models, how to determine which languages should be supported by one model is critical to multilingual NMT, which, unfortunately, has been ignored by previous work. In this work, we develop a framework that clusters languages into different groups and trains one multilingual model for each cluster. We study two methods for language clustering: (1) using prior knowledge, where we cluster languages according to language family, and (2) using language embedding, in which we represent each language by an embedding vector and cluster them in the embedding space. In particular, we obtain the embedding vectors of all the languages by training a universal neural machine translation model. Our experiments on 23 languages show that the first clustering method is simple and easy to understand but leading to suboptimal translation accuracy, while the second method sufficiently captures the relationship among languages well and improves the translation accuracy for almost all the languages over baseline methods

arXiv:1908.08917

A Lost Croatian Cybernetic Machine Translation Program

We are exploring the historical significance of research in the field of machine translation conducted by Bulcsu Laszlo, Croatian linguist, who was a pioneer in machine translation in Yugoslavia during the 1950s. We are focused on two important seminal papers written by members of his research group from 1959 and 1962, as well as their legacy in establishing a Croatian machine translation program based around the Faculty of Humanities and Social Sciences of the University of Zagreb in the late 1950s and early 1960s. We are exploring their work in connection with the beginnings of machine translation in the USA and USSR, motivated by the Cold War and the intelligence needs of the period. We also present the approach to machine translation advocated by the Croatian group in Yugoslavia, which is different from the usual logical approaches of the period, and his advocacy of cybernetic methods, which would be adopted as a canon by the mainstream AI community only decades later.

arXiv:1908.08566

Unsupervised Text Summarization via Mixed Model Back-Translation

Back-translation based approaches have recently lead to significant progress in unsupervised sequence-to-sequence tasks such as machine translation or style transfer. In this work, we extend the paradigm to the problem of learning a sentence summarization system from unaligned data. We present several initial models which rely on the asymmetrical nature of the task to perform the first back-translation step, and demonstrate the value of combining the data created by these diverse initialization methods. Our system outperforms the current state-of-the-art for unsupervised sentence summarization from fully unaligned data by over 2 ROUGE, and matches the performance of recent semi-supervised approaches.

arXiv:1908.08835공식 코드 · Papers with Code

Deep Learning Based Chatbot Models

A conversational agent (chatbot) is a piece of software that is able to communicate with humans using natural language. Modeling conversation is an important task in natural language processing and artificial intelligence. While chatbots can be used for various tasks, in general they have to understand users' utterances and provide responses that are relevant to the problem at hand. In my work, I conduct an in-depth survey of recent literature, examining over 70 publications related to chatbots published in the last 3 years. Then, I proceed to make the argument that the very nature of the general conversation domain demands approaches that are different from current state-of-of-the-art architectures. Based on several examples from the literature I show why current chatbot models fail to take into account enough priors when generating responses and how this affects the quality of the conversation. In the case of chatbots, these priors can be outside sources of information that the conversation is conditioned on like the persona or mood of the conversers. In addition to presenting the reasons behind this problem, I propose several ideas on how it could be remedied. The next section focuses on adapting the very recent Transformer model to the chatbot domain, which is currently state-of-the-art in neural machine translation. I first present experiments with the vanilla model, using conversations extracted from the Cornell Movie-Dialog Corpus. Secondly, I augment the model with some of my ideas regarding the issues of encoder-decoder architectures. More specifically, I feed additional features into the model like mood or persona together with the raw conversation data. Finally, I conduct a detailed analysis of how the vanilla model performs on conversational data by comparing it to previous chatbot models and how the additional features affect the quality of the generated responses.

arXiv:1908.08399

Controllable Dual Skew Divergence Loss for Neural Machine Translation

In sequence prediction tasks like neural machine translation, training with cross-entropy loss often leads to models that overgeneralize and plunge into local optima. In this paper, we propose an extended loss function called \emph{dual skew divergence} (DSD) that integrates two symmetric terms on KL divergences with a balanced weight. We empirically discovered that such a balanced weight plays a crucial role in applying the proposed DSD loss into deep models. Thus we eventually develop a controllable DSD loss for general-purpose scenarios. Our experiments indicate that switching to the DSD loss after the convergence of ML training helps models escape local optima and stimulates stable performance improvements. Our evaluations on the WMT 2014 English-German and English-French translation tasks demonstrate that the proposed loss as a general and convenient mean for NMT training indeed brings performance improvement in comparison to strong baselines.

arXiv:1908.07688

Improving Neural Machine Translation with Pre-trained Representation

Monolingual data has been demonstrated to be helpful in improving the translation quality of neural machine translation (NMT). The current methods stay at the usage of word-level knowledge, such as generating synthetic parallel data or extracting information from word embedding. In contrast, the power of sentence-level contextual knowledge which is more complex and diverse, playing an important role in natural language generation, has not been fully exploited. In this paper, we propose a novel structure which could leverage monolingual data to acquire sentence-level contextual representations. Then, we design a framework for integrating both source and target sentence-level representations into NMT model to improve the translation quality. Experimental results on Chinese-English, German-English machine translation tasks show that our proposed model achieves improvement over strong Transformer baselines, while experiments on English-Turkish further demonstrate the effectiveness of our approach in the low-resource scenario.

arXiv:1908.07226공식 코드 · Papers with Code

Prosodic Phrase Alignment for Machine Dubbing

Dubbing is a type of audiovisual translation where dialogues are translated and enacted so that they give the impression that the media is in the target language. It requires a careful alignment of dubbed recordings with the lip movements of performers in order to achieve visual coherence. In this paper, we deal with the specific problem of prosodic phrase synchronization within the framework of machine dubbing. Our methodology exploits the attention mechanism output in neural machine translation to find plausible phrasing for the translated dialogue lines and then uses them to condition their synthesis. Our initial work in this field records comparable speech rate ratio to professional dubbing translation, and improvement in terms of lip-syncing of long dialogue lines.

arXiv:1908.07181공식 코드 · Papers with Code

Latent-Variable Non-Autoregressive Neural Machine Translation with Deterministic Inference Using a Delta Posterior

Although neural machine translation models reached high translation quality, the autoregressive nature makes inference difficult to parallelize and leads to high translation latency. Inspired by recent refinement-based approaches, we propose LaNMT, a latent-variable non-autoregressive model with continuous latent variables and deterministic inference procedure. In contrast to existing approaches, we use a deterministic inference algorithm to find the target sequence that maximizes the lowerbound to the log-probability. During inference, the length of translation automatically adapts itself. Our experiments show that the lowerbound can be greatly increased by running the inference algorithm, resulting in significantly improved translation quality. Our proposed model closes the performance gap between non-autoregressive and autoregressive approaches on ASPEC Ja-En dataset with 8.6x faster decoding. On WMT'14 En-De dataset, our model narrows the gap with autoregressive baseline to 2.0 BLEU points with 12.5x speedup. By decoding multiple initial latent variables in parallel and rescore using a teacher model, the proposed model further brings the gap down to 1.0 BLEU point on WMT'14 En-De task with 6.8x speedup.

arXiv:1908.06748

Adabot: Fault-Tolerant Java Decompiler

Reverse Engineering(RE) has been a fundamental task in software engineering. However, most of the traditional Java reverse engineering tools are strictly rule defined, thus are not fault-tolerant, which pose serious problem when noise and interference were introduced into the system. In this paper, we view reverse engineering as a statistical machine translation task instead of rule-based task, and propose a fault-tolerant Java decompiler based on machine translation models. Our model is based on attention-based Neural Machine Translation (NMT) and Transformer architectures. First, we measure the translation quality on both the redundant and purified datasets. Next, we evaluate the fault-tolerance(anti-noise ability) of our framework on test sets with different unit error probability (UEP). In addition, we compare the suitability of different word segmentation algorithms for decompilation task. Experimental results demonstrate that our model is more robust and fault-tolerant compared to traditional Abstract Syntax Tree (AST) based decompilers. Specifically, in terms of BLEU-4 and Word Error Rate (WER), our performance has reached 94.50% and 2.65% on the redundant test set; 92.30% and 3.48% on the purified test set.

arXiv:1908.06288

Leveraging Sentence Similarity in Natural Language Generation: Improving Beam Search using Range Voting

We propose a method for natural language generation, choosing the most representative output rather than the most likely output. By viewing the language generation process from the voting theory perspective, we define representativeness using range voting and a similarity measure. The proposed method can be applied when generating from any probabilistic language model, including n-gram models and neural network models. We evaluate different similarity measures on an image captioning task and a machine translation task, and show that our method generates longer and more diverse sentences, providing a solution to the common problem of short outputs being preferred over longer and more informative ones. The generated sentences obtain higher BLEU scores, particularly when the beam size is large. We also perform a human evaluation on both tasks and find that the outputs generated using our method are rated higher.

arXiv:1908.06259

Hard but Robust, Easy but Sensitive: How Encoder and Decoder Perform in Neural Machine Translation

Neural machine translation (NMT) typically adopts the encoder-decoder framework. A good understanding of the characteristics and functionalities of the encoder and decoder can help to explain the pros and cons of the framework, and design better models for NMT. In this work, we conduct an empirical study on the encoder and the decoder in NMT, taking Transformer as an example. We find that 1) the decoder handles an easier task than the encoder in NMT, 2) the decoder is more sensitive to the input noise than the encoder, and 3) the preceding words/tokens in the decoder provide strong conditional information, which accounts for the two observations above. We hope those observations can shed light on the characteristics of the encoder and decoder and inspire future research on NMT.

arXiv:1908.06258

Language Graph Distillation for Low-Resource Machine Translation

Neural machine translation on low-resource language is challenging due to the lack of bilingual sentence pairs. Previous works usually solve the low-resource translation problem with knowledge transfer in a multilingual setting. In this paper, we propose the concept of Language Graph and further design a novel graph distillation algorithm that boosts the accuracy of low-resource translations in the graph with forward and backward knowledge distillation. Preliminary experiments on the TED talks multilingual dataset demonstrate the effectiveness of our proposed method. Specifically, we improve the low-resource translation pair by more than 3.13 points in terms of BLEU score.

arXiv:1908.06140

Improving CAT Tools in the Translation Workflow: New Approaches and Evaluation

This paper describes strategies to improve an existing web-based computer-aided translation (CAT) tool entitled CATaLog Online. CATaLog Online provides a post-editing environment with simple yet helpful project management tools. It offers translation suggestions from translation memories (TM), machine translation (MT), and automatic post-editing (APE) and records detailed logs of post-editing activities. To test the new approaches proposed in this paper, we carried out a user study on an English--German translation task using CATaLog Online. User feedback revealed that the users preferred using CATaLog Online over existing CAT tools in some respects, especially by selecting the output of the MT system and taking advantage of the color scheme for TM suggestions.

arXiv:1908.06132

Learning Representations and Agents for Information Retrieval

A goal shared by artificial intelligence and information retrieval is to create an oracle, that is, a machine that can answer our questions, no matter how difficult they are. A more limited, but still instrumental, version of this oracle is a question-answering system, in which an open-ended question is given to the machine, and an answer is produced based on the knowledge it has access to. Such systems already exist and are increasingly capable of answering complicated questions. This progress can be partially attributed to the recent success of machine learning and to the efficient methods for storing and retrieving information, most notably through web search engines. One can imagine that this general-purpose question-answering system can be built as a billion-parameters neural network trained end-to-end with a large number of pairs of questions and answers. We argue, however, that although this approach has been very successful for tasks such as machine translation, storing the world's knowledge as parameters of a learning machine can be very hard. A more efficient way is to train an artificial agent on how to use an external retrieval system to collect relevant information. This agent can leverage the effort that has been put into designing and running efficient storage and retrieval systems by learning how to best utilize them to accomplish a task. ...

arXiv:1908.06559

Recurrent Graph Syntax Encoder for Neural Machine Translation

Syntax-incorporated machine translation models have been proven successful in improving the model's reasoning and meaning preservation ability. In this paper, we propose a simple yet effective graph-structured encoder, the Recurrent Graph Syntax Encoder, dubbed \textbf{RGSE}, which enhances the ability to capture useful syntactic information. The RGSE is done over a standard encoder (recurrent or self-attention encoder), regarding recurrent network units as graph nodes and injects syntactic dependencies as edges, such that RGSE models syntactic dependencies and sequential information (\textit{i.e.}, word order) simultaneously. Our approach achieves considerable improvements over several syntax-aware NMT models in English$\Rightarrow$German and English$\Rightarrow$Czech translation tasks. And RGSE-equipped big model obtains competitive result compared with the state-of-the-art model in WMT14 En-De task. Extensive analysis further verifies that RGSE could benefit long sentence modeling, and produces better translations.

arXiv:1908.05731공식 코드 · Papers with Code

Simple and Effective Noisy Channel Modeling for Neural Machine Translation

Previous work on neural noisy channel modeling relied on latent variable models that incrementally process the source and target sentence. This makes decoding decisions based on partial source prefixes even though the full source is available. We pursue an alternative approach based on standard sequence to sequence models which utilize the entire source. These models perform remarkably well as channel models, even though they have neither been trained on, nor designed to factor over incomplete target sentences. Experiments with neural language models trained on billions of words show that noisy channel models can outperform a direct model by up to 3.2 BLEU on WMT'17 German-English translation. We evaluate on four language-pairs and our channel models consistently outperform strong alternatives such right-to-left reranking models and ensembles of direct models.

arXiv:1908.05679

Transformer-based Automatic Post-Editing with a Context-Aware Encoding Approach for Multi-Source Inputs

Recent approaches to the Automatic Post-Editing (APE) research have shown that better results are obtained by multi-source models, which jointly encode both source (src) and machine translation output (mt) to produce post-edited sentence (pe). Along this trend, we present a new multi-source APE model based on the Transformer. To construct effective joint representations, our model internally learns to incorporate src context into mt representation. With this approach, we achieve a significant improvement over baseline systems, as well as the state-of-the-art multi-source APE model. Moreover, to demonstrate the capability of our model to incorporate src context, we show that the word alignment of the unknown MT system is successfully captured in our encoding results.

arXiv:1908.05672공식 코드 · Papers with Code

Towards Making the Most of BERT in Neural Machine Translation

GPT-2 and BERT demonstrate the effectiveness of using pre-trained language models (LMs) on various natural language processing tasks. However, LM fine-tuning often suffers from catastrophic forgetting when applied to resource-rich tasks. In this work, we introduce a concerted training framework (CTNMT) that is the key to integrate the pre-trained LMs to neural machine translation (NMT). Our proposed CTNMT consists of three techniques: a) asymptotic distillation to ensure that the NMT model can retain the previous pre-trained knowledge; b) a dynamic switching gate to avoid catastrophic forgetting of pre-trained knowledge; and c) a strategy to adjust the learning paces according to a scheduled policy. Our experiments in machine translation show CTNMT gains of up to 3 BLEU score on the WMT14 English-German language pair which even surpasses the previous state-of-the-art pre-training aided NMT by 1.4 BLEU score. While for the large WMT14 English-French task with 40 millions of sentence-pairs, our base model still significantly improves upon the state-of-the-art Transformer big model by more than 1 BLEU score. The code and model can be downloaded from https://github.com/bytedance/neurst/ tree/master/examples/ctnmt.