COLING 2018 Papers

In this paper, we propose a new sentence weighting method for the domain adaptation of neural machine translation. We introduce a domain similarity metric to evaluate the relevance between a sentence and an available entire domain dataset. The similarity of each sentence to the target domain is calculated with various methods. The computed similarity is then integrated into the training objective to weight sentences. The adaptation results on both IWSLT Chinese-English TED task and a task with only synthetic training parallel data show that our sentence weighting method is able to achieve an significant improvement over strong baselines.

We present a variation of the incremental and memory-limited algorithm in (Sadeghi et al., 2017) for Bayesian cross-situational word learning and evaluate the model in terms of its functional performance and its sensitivity to input order. We show that the functional performance of our sub-optimal model on corpus data is close to that of its optimal counterpart (Frank et al., 2009), while only the sub-optimal model is capable of predicting the input order effects reported in experimental studies.

Knowing the state-of-the-art for a particular task is an essential component of any computational linguistics investigation. But can we be truly confident that the current state-of-the-art is indeed the best performing model? In this paper, we study the case of frame semantic parsing, a well-established task with multiple shared datasets. We show that in spite of all the care taken to provide a standard evaluation resource, small variations in data processing can have dramatic consequences for ranking parser performance. This leads us to propose an open-source standardized processing pipeline, which can be shared and reused for robust model comparison.

Predicting Machine Translation (MT) quality can help in many practical tasks such as MT post-editing. The performance of Quality Estimation (QE) methods has drastically improved recently with the introduction of neural approaches to the problem. However, thus far neural approaches have only been designed for word and sentence-level prediction. We present a neural framework that is able to accommodate neural QE approaches at these fine-grained levels and generalize them to the level of documents. We test the framework with two sentence-level neural QE approaches: a state of the art approach that requires extensive pre-training, and a new light-weight approach that we propose, which employs basic encoders. Our approach is significantly faster and yields performance improvements for a range of document-level quality estimation tasks. To our knowledge, this is the first neural architecture for document-level QE. In addition, for the first time we apply QE models to the output of both statistical and neural MT systems for a series of European languages and highlight the new challenges resulting from the use of neural MT.

A morphologically complex word (MCW) is a hierarchical constituent with meaning-preserving subunits, so word-based models which rely on surface forms might not be powerful enough to translate such structures. When translating from morphologically rich languages (MRLs), a source word could be mapped to several words or even a full sentence on the target side, which means an MCW should not be treated as an atomic unit. In order to provide better translations for MRLs, we boost the existing neural machine translation (NMT) architecture with a double- channel encoder and a double-attentive decoder. The main goal targeted in this research is to provide richer information on the encoder side and redesign the decoder accordingly to benefit from such information. Our experimental results demonstrate that we could achieve our goal as the proposed model outperforms existing subword- and character-based architectures and showed significant improvements on translating from German, Russian, and Turkish into English.

Methods for automated cognate detection in historical linguistics invariably build on some measure of form similarity which is designed to capture the remaining systematic similarities between cognate word forms after thousands of years of divergence. A wide range of clustering and classification algorithms has been explored for the purpose, whereas possible improvements on the level of pairwise form similarity measures have not been the main focus of research. The approach presented in this paper improves on this core component of cognate detection systems by a novel combination of information weighting, a technique for putting less weight on reoccurring morphological material, with sound correspondence modeling by means of pointwise mutual information. In evaluations on expert cognacy judgments over a subset of the IPA-encoded NorthEuraLex database, the combination of both techniques is shown to lead to considerable improvements in average precision for binary cognate detection, and modest improvements for distance-based cognate clustering.

Multilingual machine translation addresses the task of translating between multiple source and target languages. We propose task-specific attention models, a simple but effective technique for improving the quality of sequence-to-sequence neural multilingual translation. Our approach seeks to retain as much of the parameter sharing generalization of NMT models as possible, while still allowing for language-specific specialization of the attention model to a particular language-pair or task. Our experiments on four languages of the Europarl corpus show that using a target-specific model of attention provides consistent gains in translation quality for all possible translation directions, compared to a model in which all parameters are shared. We observe improved translation quality even in the (extreme) low-resource zero-shot translation directions for which the model never saw explicitly paired parallel data.

In this paper, we describe an attempt towards the development of parallel corpora for English and Ethiopian Languages, such as Amharic, Tigrigna, Afan-Oromo, Wolaytta and Ge’ez. The corpora are used for conducting a bi-directional statistical machine translation experiments. The BLEU scores of the bi-directional Statistical Machine Translation (SMT) systems show a promising result. The morphological richness of the Ethiopian languages has a great impact on the performance of SMT specially when the targets are Ethiopian languages. Now we are working towards an optimal alignment for a bi-directional English-Ethiopian languages SMT.

Noise-Contrastive Estimation (NCE) is a learning criterion that is regularly used to train neural language models in place of Maximum Likelihood Estimation, since it avoids the computational bottleneck caused by the output softmax. In this paper, we analyse and explain some of the weaknesses of this objective function, linked to the mechanism of self-normalization, by closely monitoring comparative experiments. We then explore several remedies and modifications to propose tractable and efficient NCE training strategies. In particular, we propose to make the scaling factor a trainable parameter of the model, and to use the noise distribution to initialize the output bias. These solutions, yet simple, yield stable and competitive performances in either small and large scale language modelling tasks.

Syntactic parsing plays a crucial role in improving the quality of natural language processing tasks. Although there have been several research projects on syntactic parsing in Vietnamese, the parsing quality has been far inferior than those reported in major languages, such as English and Chinese. In this work, we evaluated representative constituency parsing models on a Vietnamese Treebank to look for the most suitable parsing method for Vietnamese. We then combined the advantages of automatic and manual analysis to investigate errors produced by the experimented parsers and find the reasons for them. Our analysis focused on three possible sources of parsing errors, namely limited training data, part-of-speech (POS) tagging errors, and ambiguous constructions. As a result, we found that the last two sources, which frequently appear in Vietnamese text, significantly attributed to the poor performance of Vietnamese parsing.

Encoder-decoder based Sequence to Sequence learning (S2S) has made remarkable progress in recent years. Different network architectures have been used in the encoder/decoder. Among them, Convolutional Neural Networks (CNN) and Self Attention Networks (SAN) are the prominent ones. The two architectures achieve similar performances but use very different ways to encode and decode context: CNN use convolutional layers to focus on the local connectivity of the sequence, while SAN uses self-attention layers to focus on global semantics. In this work we propose Double Path Networks for Sequence to Sequence learning (DPN-S2S), which leverage the advantages of both models by using double path information fusion. During the encoding step, we develop a double path architecture to maintain the information coming from different paths with convolutional layers and self-attention layers separately. To effectively use the encoded context, we develop a gated attention fusion module and use it to automatically pick up the information needed during the decoding step, which is also a double path network. By deeply integrating the two paths, both types of information are combined and well exploited. Experiments show that our proposed method can significantly improve the performance of sequence to sequence learning over state-of-the-art systems.

We formulate a generalization of Petrov et al. (2006)’s split/merge algorithm for interpreted regular tree grammars (Koller and Kuhlmann, 2011), which capture a large class of grammar formalisms. We evaluate its effectiveness empirically on the task of discontinuous constituent parsing with two mildly context-sensitive grammar formalisms: linear context-free rewriting systems (Vijay-Shanker et al., 1987) as well as hybrid grammars (Nederhof and Vogler, 2014).

In the popular sequence to sequence (seq2seq) neural machine translation (NMT), there exist many weighted sum models (WSMs), each of which takes a set of input and generates one output. However, the weights in a WSM are independent of each other and fixed for all inputs, suggesting that by ignoring different needs of inputs, the WSM lacks effective control on the influence of each input. In this paper, we propose adaptive weighting for WSMs to control the contribution of each input. Specifically, we apply adaptive weighting for both GRU and the output state in NMT. Experimentation on Chinese-to-English translation and English-to-German translation demonstrates that the proposed adaptive weighting is able to much improve translation accuracy by achieving significant improvement of 1.49 and 0.92 BLEU points for the two translation tasks. Moreover, we discuss in-depth on what type of information is encoded in the encoder and how information influences the generation of target words in the decoder.

To enrich vocabulary of low resource settings, we proposed a novel method which identify loanwords in monolingual corpora. More specifically, we first use cross-lingual word embeddings as the core feature to generate semantically related candidates based on comparable corpora and a small bilingual lexicon; then, a log-linear model which combines several shallow features such as pronunciation similarity and hybrid language model features to predict the final results. In this paper, we use Uyghur as the receipt language and try to detect loanwords in four donor languages: Arabic, Chinese, Persian and Russian. We conduct two groups of experiments to evaluate the effectiveness of our proposed approach: loanword identification and OOV translation in four language pairs and eight translation directions (Uyghur-Arabic, Arabic-Uyghur, Uyghur-Chinese, Chinese-Uyghur, Uyghur-Persian, Persian-Uyghur, Uyghur-Russian, and Russian-Uyghur). Experimental results on loanword identification show that our method outperforms other baseline models significantly. Neural machine translation models integrating results of loanword identification experiments achieve the best results on OOV translation(with 0.5-0.9 BLEU improvements)

Neural machine translation systems require a number of stacked layers for deep models. But the prediction depends on the sentence representation of the top-most layer with no access to low-level representations. This makes it more difficult to train the model and poses a risk of information loss to prediction. In this paper, we propose a multi-layer representation fusion (MLRF) approach to fusing stacked layers. In particular, we design three fusion functions to learn a better representation from the stack. Experimental results show that our approach yields improvements of 0.92 and 0.56 BLEU points over the strong Transformer baseline on IWSLT German-English and NIST Chinese-English MT tasks respectively. The result is new state-of-the-art in German-English translation.

This paper describes the augmentation of an existing corpus of child-directed speech. The resulting corpus is a gold-standard labeled corpus for supervised learning of semantic role labels in adult-child dialogues. Semantic role labeling (SRL) models assign semantic roles to sentence constituents, thus indicating who has done what to whom (and in what way). The current corpus is derived from the Adam files in the Brown corpus (Brown 1973) of the CHILDES corpora, and augments the partial annotation described in Connor et al. (2010). It provides labels for both semantic arguments of verbs and semantic arguments of prepositions. The semantic role labels and senses of verbs follow Propbank guidelines Kingsbury and Palmer, 2002; Gildea and Palmer 2002; Palmer et al., 2005) and those for prepositions follow Srikumar and Roth (2011). The corpus was annotated by two annotators. Inter-annotator agreement is given separately for prepositions and verbs, and for adult speech and child speech. Overall, across child and adult samples, including verbs and prepositions, the kappa score for sense is 72.6, for the number of semantic-role-bearing arguments, the kappa score is 77.4, for identical semantic role labels on a given argument, the kappa score is 91.1, for the span of semantic role labels, and the kappa for agreement is 93.9. The sense and number of arguments was often open to multiple interpretations in child speech, due to the rapidly changing discourse and omission of constituents in production. Annotators used a discourse context window of ten sentences before and ten sentences after the target utterance to determine the annotation labels. The derived corpus is available for use in CHAT (MacWhinney, 2000) and XML format.

Incrementality is ubiquitous in human-human interaction and beneficial for human-computer interaction. It has been a topic of research in different parts of the NLP community, mostly with focus on the specific topic at hand even though incremental systems have to deal with similar challenges regardless of domain. In this survey, I consolidate and categorize the approaches, identifying similarities and differences in the computation and data, and show trade-offs that have to be considered. A focus lies on evaluating incremental systems because the standard metrics often fail to capture the incremental properties of a system and coming up with a suitable evaluation scheme is non-trivial.

K-fold cross validation (CV) is a popular method for estimating the true performance of machine learning models, allowing model selection and parameter tuning. However, the very process of CV requires random partitioning of the data and so our performance estimates are in fact stochastic, with variability that can be substantial for natural language processing tasks. We demonstrate that these unstable estimates cannot be relied upon for effective parameter tuning. The resulting tuned parameters are highly sensitive to how our data is partitioned, meaning that we often select sub-optimal parameter choices and have serious reproducibility issues. Instead, we propose to use the less variable J-K-fold CV, in which J independent K-fold cross validations are used to assess performance. Our main contributions are extending J-K-fold CV from performance estimation to parameter tuning and investigating how to choose J and K. We argue that variability is more important than bias for effective tuning and so advocate lower choices of K than are typically seen in the NLP literature and instead use the saved computation to increase J. To demonstrate the generality of our recommendations we investigate a wide range of case-studies: sentiment classification (both general and target-specific), part-of-speech tagging and document classification.

We study three general multi-task learning (MTL) approaches on 11 sequence tagging tasks. Our extensive empirical results show that in about 50% of the cases, jointly learning all 11 tasks improves upon either independent or pairwise learning of the tasks. We also show that pairwise MTL can inform us what tasks can benefit others or what tasks can be benefited if they are learned jointly. In particular, we identify tasks that can always benefit others as well as tasks that can always be harmed by others. Interestingly, one of our MTL approaches yields embeddings of the tasks that reveal the natural clustering of semantic and syntactic tasks. Our inquiries have opened the doors to further utilization of MTL in NLP.