COLING 2018 Papers

This paper proposes a simple CNN model for creating general-purpose sentence embeddings that can transfer easily across domains and can also act as effective initialization for downstream tasks. Recently, averaging the embeddings of words in a sentence has proven to be a surprisingly successful and efficient way of obtaining sentence embeddings. However, these models represent a sentence, only in terms of features of words or uni-grams in it. In contrast, our model (CSE) utilizes both features of words and n-grams to encode sentences, which is actually a generalization of these bag-of-words models. The extensive experiments demonstrate that CSE performs better than average models in transfer learning setting and exceeds the state of the art in supervised learning setting by initializing the parameters with the pre-trained sentence embeddings.

This paper describes CzEngClass, a bilingual lexical resource being built to investigate verbal synonymy in bilingual context and to relate semantic roles common to one synonym class to verb arguments (verb valency). In addition, the resource is linked to existing resources with the same of a similar aim: English and Czech WordNet, FrameNet, PropBank, VerbNet (SemLink), and valency lexicons for Czech and English (PDT-Vallex, Vallex, and EngVallex). There are several goals of this work and resource: (a) to provide gold standard data for automatic experiments in the future (such as automatic discovery of synonym classes, word sense disambiguation, assignment of classes to occurrences of verbs in text, coreferential linking of verb and event arguments in text, etc.), (b) to build a core (bilingual) lexicon linked to existing resources, for comparative studies and possibly for training automatic tools, and (c) to enrich the annotation of a parallel treebank, the Prague Czech English Dependency Treebank, which so far contained valency annotation but has not linked synonymous senses of verbs together. The method used for extracting the synonym classes is a semi-automatic process with a substantial amount of manual work during filtering, role assignment to classes and individual Class members’ arguments, and linking to the external lexical resources. We present the first version with 200 classes (about 1800 verbs) and evaluate interannotator agreement using several metrics.

In this paper, we first utilize the word embedding that focuses on sub-word units to the Mongolian Phrase Break (PB) prediction task by using Long-Short-Term-Memory (LSTM) model. Mongolian is an agglutinative language. Each root can be followed by several suffixes to form probably millions of words, but the existing Mongolian corpus is not enough to build a robust entire word embedding, thus it suffers a serious data sparse problem and brings a great difficulty for Mongolian PB prediction. To solve this problem, we look at sub-word units in Mongolian word, and encode their information to a meaningful representation, then fed it to LSTM to decode the best corresponding PB label. Experimental results show that the proposed model significantly outperforms traditional CRF model using manually features and obtains 7.49% F-Measure gain.

Despite the success of existing works on single-turn conversation generation, taking the coherence in consideration, human conversing is actually a context-sensitive process. Inspired by the existing studies, this paper proposed the static and dynamic attention based approaches for context-sensitive generation of open-domain conversational responses. Experimental results on two public datasets show that the proposed static attention based approach outperforms all the baselines on automatic and human evaluation.

Knowledge graphs are a versatile framework to encode richly structured data relationships, but it can be challenging to combine these graphs with unstructured data. Methods for retrofitting pre-trained entity representations to the structure of a knowledge graph typically assume that entities are embedded in a connected space and that relations imply similarity. However, useful knowledge graphs often contain diverse entities and relations (with potentially disjoint underlying corpora) which do not accord with these assumptions. To overcome these limitations, we present Functional Retrofitting, a framework that generalizes current retrofitting methods by explicitly modeling pairwise relations. Our framework can directly incorporate a variety of pairwise penalty functions previously developed for knowledge graph completion. Further, it allows users to encode, learn, and extract information about relation semantics. We present both linear and neural instantiations of the framework. Functional Retrofitting significantly outperforms existing retrofitting methods on complex knowledge graphs and loses no accuracy on simpler graphs (in which relations do imply similarity). Finally, we demonstrate the utility of the framework by predicting new drug–disease treatment pairs in a large, complex health knowledge graph.

With more and more people around the world learning Chinese as a second language, the need of Chinese error correction tools is increasing. In the HSK dynamic composition corpus, word usage error (WUE) is the most common error type. In this paper, we build a neural network model that considers both target erroneous token and context to generate a correction vector and compare it against a candidate vocabulary to propose suitable corrections. To deal with potential alternative corrections, the top five proposed candidates are judged by native Chinese speakers. For more than 91% of the cases, our system can propose at least one acceptable correction within a list of five candidates. To the best of our knowledge, this is the first research addressing general-type Chinese WUE correction. Our system can help non-native Chinese learners revise their sentences by themselves.

Stance detection aims to assign a stance label (for or against) to a post toward a specific target. Recently, there is a growing interest in using neural models to detect stance of documents. Most of these works model the sequence of words to learn document representation. However, much linguistic information, such as polarity and arguments of the document, is correlated with the stance of the document, and can inspire us to explore the stance. Hence, we present a neural model to fully employ various linguistic information to construct the document representation. In addition, since the influences of different linguistic information are different, we propose a hierarchical attention network to weigh the importance of various linguistic information, and learn the mutual attention between the document and the linguistic information. The experimental results on two datasets demonstrate the effectiveness of the proposed hierarchical attention neural model.

This paper explores the influence of spelling errors on lexical variation measures. Lexical richness measures such as Type-Token Ration (TTR) and Yule’s K are often used for learner English analysis and assessment. When applied to learner English, however, they can be unreliable because of the spelling errors appearing in it. Namely, they are, directly or indirectly, based on the counts of distinct word types, and spelling errors undesirably increase the number of distinct words. This paper introduces and examines the hypothesis that lexical richness measures become unstable in learner English because of spelling errors. Specifically, it tests the hypothesis on English learner corpora of three groups (middle school, high school, and college students). To be precise, it estimates the difference in TTR and Yule’s K caused by spelling errors, by calculating their values before and after spelling errors are manually corrected. Furthermore, it examines the results theoretically and empirically to deepen the understanding of the influence of spelling errors on them.

We examine the utility of linguistic content and vocal characteristics for multimodal deep learning in human spoken language understanding. We present a deep multimodal network with both feature attention and modality attention to classify utterance-level speech data. The proposed hybrid attention architecture helps the system focus on learning informative representations for both modality-specific feature extraction and model fusion. The experimental results show that our system achieves state-of-the-art or competitive results on three published multimodal datasets. We also demonstrated the effectiveness and generalization of our system on a medical speech dataset from an actual trauma scenario. Furthermore, we provided a detailed comparison and analysis of traditional approaches and deep learning methods on both feature extraction and fusion.

Large scale knowledge graphs (KGs) such as Freebase are generally incomplete. Reasoning over multi-hop (mh) KG paths is thus an important capability that is needed for question answering or other NLP tasks that require knowledge about the world. mh-KG reasoning includes diverse scenarios, e.g., given a head entity and a relation path, predict the tail entity; or given two entities connected by some relation paths, predict the unknown relation between them. We present ROPs, recurrent one-hop predictors, that predict entities at each step of mh-KB paths by using recurrent neural networks and vector representations of entities and relations, with two benefits: (i) modeling mh-paths of arbitrary lengths while updating the entity and relation representations by the training signal at each step; (ii) handling different types of mh-KG reasoning in a unified framework. Our models show state-of-the-art for two important multi-hop KG reasoning tasks: Knowledge Base Completion and Path Query Answering.

Capturing semantic relations between sentences, such as entailment, is a long-standing challenge for computational semantics. Logic-based models analyse entailment in terms of possible worlds (interpretations, or situations) where a premise P entails a hypothesis H iff in all worlds where P is true, H is also true. Statistical models view this relationship probabilistically, addressing it in terms of whether a human would likely infer H from P. In this paper, we wish to bridge these two perspectives, by arguing for a visually-grounded version of the Textual Entailment task. Specifically, we ask whether models can perform better if, in addition to P and H, there is also an image (corresponding to the relevant “world” or “situation”). We use a multimodal version of the SNLI dataset (Bowman et al., 2015) and we compare “blind” and visually-augmented models of textual entailment. We show that visual information is beneficial, but we also conduct an in-depth error analysis that reveals that current multimodal models are not performing “grounding” in an optimal fashion.

Natural language inference (NLI) is the task of determining if a natural language hypothesis can be inferred from a given premise in a justifiable manner. NLI was proposed as a benchmark task for natural language understanding. Existing models perform well at standard datasets for NLI, achieving impressive results across different genres of text. However, the extent to which these models understand the semantic content of sentences is unclear. In this work, we propose an evaluation methodology consisting of automatically constructed “stress tests” that allow us to examine whether systems have the ability to make real inferential decisions. Our evaluation of six sentence-encoder models on these stress tests reveals strengths and weaknesses of these models with respect to challenging linguistic phenomena, and suggests important directions for future work in this area.

This survey discusses how recent developments in multimodal processing facilitate conceptual grounding of language. We categorize the information flow in multimodal processing with respect to cognitive models of human information processing and analyze different methods for combining multimodal representations. Based on this methodological inventory, we discuss the benefit of multimodal grounding for a variety of language processing tasks and the challenges that arise. We particularly focus on multimodal grounding of verbs which play a crucial role for the compositional power of language.

Several semi-supervised representation learning methods have been proposed recently that mitigate the drawbacks of traditional bootstrapping: they reduce the amount of semantic drift introduced by iterative approaches through one-shot learning; others address the sparsity of data through the learning of custom, dense representation for the information modeled. In this work, we are the first to adapt three of these methods, most of which have been originally proposed for image processing, to an information extraction task, specifically, named entity classification. Further, we perform a rigorous comparative analysis on two distinct datasets. Our analysis yields several important observations. First, all representation learning methods outperform state-of-the-art semi-supervised methods that do not rely on representation learning. To the best of our knowledge, we report the latest state-of-the-art results on the semi-supervised named entity classification task. Second, one-shot learning methods clearly outperform iterative representation learning approaches. Lastly, one of the best performers relies on the mean teacher framework (Tarvainen and Valpola, 2017), a simple teacher/student approach that is independent of the underlying task-specific model.

We present an Open Information Extraction (IE) approach that uses a two-layered transformation stage consisting of a clausal disembedding layer and a phrasal disembedding layer, together with rhetorical relation identification. In that way, we convert sentences that present a complex linguistic structure into simplified, syntactically sound sentences, from which we can extract propositions that are represented in a two-layered hierarchy in the form of core relational tuples and accompanying contextual information which are semantically linked via rhetorical relations. In a comparative evaluation, we demonstrate that our reference implementation Graphene outperforms state-of-the-art Open IE systems in the construction of correct n-ary predicate-argument structures. Moreover, we show that existing Open IE approaches can benefit from the transformation process of our framework.

We develop CALM, a coordination analyzer that improves upon the conjuncts identified from dependency parses. It uses a language model based scoring and several linguistic constraints to search over hierarchical conjunct boundaries (for nested coordination). By splitting a conjunctive sentence around these conjuncts, CALM outputs several simple sentences. We demonstrate the value of our coordination analyzer in the end task of Open Information Extraction (Open IE). State-of-the-art Open IE systems lose substantial yield due to ineffective processing of conjunctive sentences. Our Open IE system, CALMIE, performs extraction over the simple sentences identified by CALM to obtain up to 1.8x yield with a moderate increase in precision compared to extractions from original sentences.

Finding names of people killed by police has become increasingly important as police shootings get more and more public attention (police killing detection). Unfortunately, there has been not much work in the literature addressing this problem. The early work in this field (Keith etal., 2017) proposed a distant supervision framework based on Expectation Maximization (EM) to deal with the multiple appearances of the names in documents. However, such EM-based framework cannot take full advantages of deep learning models, necessitating the use of handdesigned features to improve the detection performance. In this work, we present a novel deep learning method to solve the problem of police killing recognition. The proposed method relies on hierarchical LSTMs to model the multiple sentences that contain the person names of interests, and introduce supervised attention mechanisms based on semantical word lists and dependency trees to upweight the important contextual words. Our experiments demonstrate the benefits of the proposed model and yield the state-of-the-art performance for police killing detection.

The state-of-the-art methods for relation classification are primarily based on deep neural net- works. This kind of supervised learning method suffers from not only limited training data, but also the large number of low-frequency relations in specific domains. In this paper, we propose the task of exploratory relation classification for domain knowledge harvesting. The goal is to learn a classifier on pre-defined relations and discover new relations expressed in texts. A dynamically structured neural network is introduced to classify entity pairs to a continuously expanded relation set. We further propose the similarity sensitive Chinese restaurant process to discover new relations. Experiments conducted on a large corpus show the effectiveness of our neural network, while new relations are discovered with high precision and recall.

This paper reports on a reimplementation of a system on detecting implicit positive meaning from negated statements. In the original regression experiment, different positive interpretations per negation are scored according to their likelihood. We convert the scores to classes and report our results on both the regression and classification tasks. We show that a baseline taking the mean score or most frequent class is hard to beat because of class imbalance in the dataset. Our error analysis indicates that an approach that takes the information structure into account (i.e. which information is new or contrastive) may be promising, which requires looking beyond the syntactic and semantic characteristics of negated statements.