COLING 2016 论文列表

This paper introduces a continuous system capable of automatically producing the most adequate speaking style to synthesize a desired target text. This is done thanks to a joint modeling of the acoustic and lexical parameters of the speaker models by adapting the CVSM projection of the training texts using MR-HMM techniques. As such, we consider that as long as sufficient variety in the training data is available, we should be able to model a continuous lexical space into a continuous acoustic space. The proposed continuous automatic text to speech system was evaluated by means of a perceptual evaluation in order to compare them with traditional approaches to the task. The system proved to be capable of conveying the correct expressiveness (average adequacy of 3.6) with an expressive strength comparable to oracle traditional expressive speech synthesis (average of 3.6) although with a drop in speech quality mainly due to the semi-continuous nature of the data (average quality of 2.9). This means that the proposed system is capable of improving traditional neutral systems without requiring any additional user interaction.

In the context of natural language processing, representation learning has emerged as a newly active research subject because of its excellent performance in many applications. Learning representations of words is a pioneering study in this school of research. However, paragraph (or sentence and document) embedding learning is more suitable/reasonable for some tasks, such as sentiment classification and document summarization. Nevertheless, as far as we are aware, there is only a dearth of research focusing on launching unsupervised paragraph embedding methods. Classic paragraph embedding methods infer the representation of a given paragraph by considering all of the words occurring in the paragraph. Consequently, those stop or function words that occur frequently may mislead the embedding learning process to produce a misty paragraph representation. Motivated by these observations, our major contributions are twofold. First, we propose a novel unsupervised paragraph embedding method, named the essence vector (EV) model, which aims at not only distilling the most representative information from a paragraph but also excluding the general background information to produce a more informative low-dimensional vector representation for the paragraph. We evaluate the proposed EV model on benchmark sentiment classification and multi-document summarization tasks. The experimental results demonstrate the effectiveness and applicability of the proposed embedding method. Second, in view of the increasing importance of spoken content processing, an extension of the EV model, named the denoising essence vector (D-EV) model, is proposed. The D-EV model not only inherits the advantages of the EV model but also can infer a more robust representation for a given spoken paragraph against imperfect speech recognition. The utility of the D-EV model is evaluated on a spoken document summarization task, confirming the effectiveness of the proposed embedding method in relation to several well-practiced and state-of-the-art summarization methods.

Development of hand crafted rule for syllabifying words of a language is an expensive task. This paper proposes several data-driven methods for automatic syllabification of words written in Manipuri language. Manipuri is one of the scheduled Indian languages. First, we propose a language-independent rule-based approach formulated using entropy based phonotactic segmentation. Second, we project the syllabification problem as a sequence labeling problem and investigate its effect using various sequence labeling approaches. Third, we combine the effect of sequence labeling and rule-based method and investigate the performance of the hybrid approach. From various experimental observations, it is evident that the proposed methods outperform the baseline rule-based method. The entropy based phonotactic segmentation provides a word accuracy of 96%, CRF (sequence labeling approach) provides 97% and hybrid approach provides 98% word accuracy.

Parsing texts into universal dependencies (UD) in realistic scenarios requires infrastructure for the morphological analysis and disambiguation (MA&D) of typologically different languages as a first tier. MA&D is particularly challenging in morphologically rich languages (MRLs), where the ambiguous space-delimited tokens ought to be disambiguated with respect to their constituent morphemes, each morpheme carrying its own tag and a rich set features. Here we present a novel, language-agnostic, framework for MA&D, based on a transition system with two variants — word-based and morpheme-based — and a dedicated transition to mitigate the biases of variable-length morpheme sequences. Our experiments on a Modern Hebrew case study show state of the art results, and we show that the morpheme-based MD consistently outperforms our word-based variant. We further illustrate the utility and multilingual coverage of our framework by morphologically analyzing and disambiguating the large set of languages in the UD treebanks.

We propose a new approach to PoS tagging where in a first step, we assign a coarse-grained tag corresponding to the main syntactic category. Based on this high-precision decision, in the second step we utilize specially trained fine-grained models with heavily reduced decision complexity. By analyzing the system under oracle conditions, we show that there is a quite large potential for significantly outperforming a competitive baseline. When we take error-propagation from the coarse-grained tagging into account, our approach is on par with the state of the art. Our approach also allows tailoring the tagger towards recognizing single word classes which are of interest e.g. for researchers searching for specific phenomena in large corpora. In a case study, we significantly outperform a standard model that also makes use of the same optimizations.

Previous studies on Thai Sentence Boundary Detection (SBD) mostly assumed sentence ends at a space disambiguation problem, which classified space either as an indicator for Sentence Boundary (SB) or non-Sentence Boundary (nSB). In this paper, we propose a word labeling approach which treats space as a normal word, and detects SB between any two words. This removes the restriction for SB to be oc-curred only at space and makes our system more robust for modern Thai writing. It is because in modern Thai writing, space is not consistently used to indicate SB. As syntactic information contributes to better SBD, we further propose a joint Part-Of-Speech (POS) tagging and SBD framework based on Factorial Conditional Random Field (FCRF) model. We compare the performance of our proposed ap-proach with reported methods on ORCHID corpus. We also performed experiments of FCRF model on the TaLAPi corpus. The results show that the word labelling approach has better performance than pre-vious space-based classification approaches and FCRF joint model outperforms LCRF model in terms of SBD in all experiments.

Sequence labeling architectures use word embeddings for capturing similarity, but suffer when handling previously unseen or rare words. We investigate character-level extensions to such models and propose a novel architecture for combining alternative word representations. By using an attention mechanism, the model is able to dynamically decide how much information to use from a word- or character-level component. We evaluated different architectures on a range of sequence labeling datasets, and character-level extensions were found to improve performance on every benchmark. In addition, the proposed attention-based architecture delivered the best results even with a smaller number of trainable parameters.

In this paper, we propose a new annotation approach to Chinese word segmentation, part-of-speech (POS) tagging and dependency labelling that aims to overcome the two major issues in traditional morphology-based annotation: Inconsistency and data sparsity. We re-annotate the Penn Chinese Treebank 5.0 (CTB5) and demonstrate the advantages of this approach compared to the original CTB5 annotation through word segmentation, POS tagging and machine translation experiments.

The paper applies a deep recurrent neural network to the task of sentence boundary detection in Sanskrit, an important, yet underresourced ancient Indian language. The deep learning approach improves the F scores set by a metrical baseline and by a Conditional Random Field classifier by more than 10%.

In this paper, we study the problem of disfluency detection using the encoder-decoder framework. We treat disfluency detection as a sequence-to-sequence problem and propose a neural attention-based model which can efficiently model the long-range dependencies between words and make the resulting sentence more likely to be grammatically correct. Our model firstly encode the source sentence with a bidirectional Long Short-Term Memory (BI-LSTM) and then use the neural attention as a pointer to select an ordered sub sequence of the input as the output. Experiments show that our model achieves the state-of-the-art f-score of 86.7% on the commonly used English Switchboard test set. We also evaluate the performance of our model on the in-house annotated Chinese data and achieve a significantly higher f-score compared to the baseline of CRF-based approach.

Predictive incremental parsing produces syntactic representations of sentences as they are produced, e.g. by typing or speaking. In order to generate connected parses for such unfinished sentences, upcoming word types can be hypothesized and structurally integrated with already realized words. For example, the presence of a determiner as the last word of a sentence prefix may indicate that a noun will appear somewhere in the completion of that sentence, and the determiner can be attached to the predicted noun. We combine the forward-looking parser predictions with backward-looking N-gram histories and analyze in a set of experiments the impact on language models, i.e. stronger discriminative power but also higher data sparsity. Conditioning N-gram models, MaxEnt models or RNN-LMs on parser predictions yields perplexity reductions of about 6%. Our method (a) retains online decoding capabilities and (b) incurs relatively little computational overhead which sets it apart from previous approaches that use syntax for language modeling. Our method is particularly attractive for modular systems that make use of a syntax parser anyway, e.g. as part of an understanding pipeline where predictive parsing improves language modeling at no additional cost.

This paper presents a deep learning architecture for the semantic decoder component of a Statistical Spoken Dialogue System. In a slot-filling dialogue, the semantic decoder predicts the dialogue act and a set of slot-value pairs from a set of n-best hypotheses returned by the Automatic Speech Recognition. Most current models for spoken language understanding assume (i) word-aligned semantic annotations as in sequence taggers and (ii) delexicalisation, or a mapping of input words to domain-specific concepts using heuristics that try to capture morphological variation but that do not scale to other domains nor to language variation (e.g., morphology, synonyms, paraphrasing ). In this work the semantic decoder is trained using unaligned semantic annotations and it uses distributed semantic representation learning to overcome the limitations of explicit delexicalisation. The proposed architecture uses a convolutional neural network for the sentence representation and a long-short term memory network for the context representation. Results are presented for the publicly available DSTC2 corpus and an In-car corpus which is similar to DSTC2 but has a significantly higher word error rate (WER).

Extracting summaries via integer linear programming and submodularity are popular and successful techniques in extractive multi-document summarization. However, many interesting optimization objectives are neither submodular nor factorizable into an integer linear program. We address this issue and present a general optimization framework where any function of input documents and a system summary can be plugged in. Our framework includes two kinds of summarizers – one based on genetic algorithms, the other using a swarm intelligence approach. In our experimental evaluation, we investigate the optimization of two information-theoretic summary evaluation metrics and find that our framework yields competitive results compared to several strong summarization baselines. Our comparative analysis of the genetic and swarm summarizers reveals interesting complementary properties.

This paper studies the abstractive multi-document summarization for event-oriented news texts through event information extraction and abstract representation. Fine-grained event mentions and semantic relations between them are extracted to build a unified and connected event semantic link network, an abstract representation of source texts. A network reduction algorithm is proposed to summarize the most salient and coherent event information. New sentences with good linguistic quality are automatically generated and selected through sentences over-generation and greedy-selection processes. Experimental results on DUC 2006 and DUC 2007 datasets show that our system significantly outperforms the state-of-the-art extractive and abstractive baselines under both pyramid and ROUGE evaluation metrics.

A key component in surface realization in natural language generation is to choose concrete syntactic relationships to express a target meaning. We develop a new method for syntactic choice based on learning a stochastic tree grammar in a neural architecture. This framework can exploit state-of-the-art methods for modeling word sequences and generalizing across vocabulary. We also induce embeddings to generalize over elementary tree structures and exploit a tree recurrence over the input structure to model long-distance influences between NLG choices. We evaluate the models on the task of linearizing unannotated dependency trees, documenting the contribution of our modeling techniques to improvements in both accuracy and run time.

Summarization aims to represent source documents by a shortened passage. Existing methods focus on the extraction of key information, but often neglect coherence. Hence the generated summaries suffer from a lack of readability. To address this problem, we have developed a graph-based method by exploring the links between text to produce coherent summaries. Our approach involves finding a sequence of sentences that best represent the key information in a coherent way. In contrast to the previous methods that focus only on salience, the proposed method addresses both coherence and informativeness based on textual linkages. We conduct experiments on the DUC2004 summarization task data set. A performance comparison reveals that the summaries generated by the proposed system achieve comparable results in terms of the ROUGE metric, and show improvements in readability by human evaluation.

In this paper, we empirically explore the effects of various kinds of skip connections in stacked bidirectional LSTMs for sequential tagging. We investigate three kinds of skip connections connecting to LSTM cells: (a) skip connections to the gates, (b) skip connections to the internal states and (c) skip connections to the cell outputs. We present comprehensive experiments showing that skip connections to cell outputs outperform the remaining two. Furthermore, we observe that using gated identity functions as skip mappings works pretty well. Based on this novel skip connections, we successfully train deep stacked bidirectional LSTM models and obtain state-of-the-art results on CCG supertagging and comparable results on POS tagging.

Existing asynchronous parallel learning methods are only for the sparse feature models, and they face new challenges for the dense feature models like neural networks (e.g., LSTM, RNN). The problem for dense features is that asynchronous parallel learning brings gradient errors derived from overwrite actions. We show that gradient errors are very common and inevitable. Nevertheless, our theoretical analysis shows that the learning process with gradient errors can still be convergent towards the optimum of objective functions for many practical applications. Thus, we propose a simple method AsynGrad for asynchronous parallel learning with gradient error. Base on various dense feature models (LSTM, dense-CRF) and various NLP tasks, experiments show that AsynGrad achieves substantial improvement on training speed, and without any loss on accuracy.

Languages with rich morphology often introduce sparsity in language processing tasks. While morphological analyzers can reduce this sparsity by providing morpheme-level analyses for words, they will often introduce ambiguity by returning multiple analyses for the same surface form. The problem of disambiguating between these morphological parses is further complicated by the fact that a correct parse for a word is not only be dependent on the surface form but also on other words in its context. In this paper, we present a language-agnostic approach to morphological disambiguation. We address the problem of using context in morphological disambiguation by presenting several LSTM-based neural architectures that encode long-range surface-level and analysis-level contextual dependencies. We applied our approach to Turkish, Russian, and Arabic to compare effectiveness across languages, matching state-of-the-art results in two of the three languages. Our results also demonstrate that while context plays a role in learning how to disambiguate, the type and amount of context needed varies between languages.