ACL 2016 论文列表

This paper proposes an unsupervised approach for segmenting a multi-author document into authorial components. The key novelty is that we utilize the sequential patterns hidden among document elements when determining their authorships. For this purpose, we adopt Hidden Markov Model (HMM) and construct a sequential probabilistic model to capture the dependencies of sequential sentences and their authorships. An unsupervised learning method is developed to initialize the HMM parameters. Experimental results on benchmark datasets have demonstrated the significant ben-efit of our idea and our approach has outperformed the state-of-the-arts on all tests. As an example of its applications, the proposed approach is applied for attributing authorship of a document and has also shown promising re-sults.

Sentiment Analysis (SA) is an active research area nowadays due to the tremendous interest in aggregating and evaluating opinions being disseminated by users on the Web. SA of English has been thoroughly researched; however research on SA of Arabic has just flourished. Twitter is considered a powerful tool for disseminating information and a rich resource for opinionated text containing views on many different topics. In this paper we attempt to bridge a gap in Arabic SA of Twitter which is the lack of sentiment lexi-cons that are tailored for the informal language of Twitter. We generate two lexicons extracted from a large dataset of tweets using two approaches and evaluate their use in a simple lexicon based method. The evaluation is performed on internal and external datasets. The performance of these automatically generated lexicons was very promising, albeit the simple method used for classification. The best F-score obtained was 89.58% on the internal dataset and 63.1-64.7% on the exter-nal datasets.

Document collections often have links between documents—citations, hyperlinks, or revisions—and which links are added is often based on topical similarity. To model these intuitions, we introduce a new topic model for documents situated within a network structure, integrating latent blocks of documents with a max-margin learning criterion for link prediction using topicand word-level features. Experiments on a scientific paper dataset and collection of webpages show that, by more robustly exploiting the rich link structure within a document network, our model improves link prediction, topic quality, and block distributions.

News reader comments found in many on-line news websites are typically massive in amount. We investigate the task of Cultural-common Topic Detection (CTD), which is aimed at discovering common discussion topics from news reader comments written in different languages. We propose a new probabilistic graphical model called MCTA which can cope with the language gap and capture the common semantics in different languages. We also develop a partially collapsed Gibbs sampler which effectively incorporates the term translation relationship into the detection of cultural-common topics for model parameter learning. Experimental results show improvements over the state-of-the-art model.

Word embedding maps words into a low-dimensional continuous embedding space by exploiting the local word collocation patterns in a small context window. On the other hand, topic modeling maps documents onto a low-dimensional topic space, by utilizing the global word collocation patterns in the same document. These two types of patterns are complementary. In this paper, we propose a generative topic embedding model to combine the two types of patterns. In our model, topics are represented by embedding vectors, and are shared across documents. The probability of each word is influenced by both its local context and its topic. A variational inference method yields the topic embeddings as well as the topic mixing proportions for each document. Jointly they represent the document in a low-dimensional continuous space. In two document classification tasks, our method performs better than eight existing methods, with fewer features. In addition, we illustrate with an example that our method can generate coherent topics even based on only one document.

Properties of corpora, such as the diversity of vocabulary and how tightly related texts cluster together, impact the best way to cluster short texts. We examine several such properties in a variety of corpora and track their effects on various combinations of similarity metrics and clustering algorithms. We show that semantic similarity metrics outperform traditional n-gram and dependency similarity metrics for kmeans clustering of a linguistically creative dataset, but do not help with less creative texts. Yet the choice of similarity metric interacts with the choice of clustering method. We find that graphbased clustering methods perform well on tightly clustered data but poorly on loosely clustered data. Semantic similarity metrics generate loosely clustered output even when applied to a tightly clustered dataset. Thus, the best performing clustering systems could not use semantic metrics.

A long-standing challenge in coreference resolution has been the incorporation of entity-level information - features defined over clusters of mentions instead of mention pairs. We present a neural network based coreference system that produces high-dimensional vector representations for pairs of coreference clusters. Using these representations, our system learns when combining clusters is desirable. We train the system with a learning-to-search algorithm that teaches it which local decisions (cluster merges) will lead to a high-scoring final coreference partition. The system substantially outperforms the current state-of-the-art on the English and Chinese portions of the CoNLL 2012 Shared Task dataset despite using few hand-engineered features.

Interpretability and discriminative power are the two most basic requirements for an evaluation metric. In this paper, we report the mention identification effect in the B 3 , CEAF , and BLANC coreference evaluation metrics that makes it impossible to interpret their results properly. The only metric which is insensitive to this flaw is MUC , which, however, is known to be the least discriminative metric. It is a known fact that none of the current metrics are reliable. The common practice for ranking coreference resolvers is to use the average of three different metrics. However, one cannot expect to obtain a reliable score by averaging three unreliable metrics. We propose LEA, a Link-based Entity-Aware evaluation metric that is designed to overcome the shortcomings of the current evaluation metrics. LEA is available as branch LEA-scorer in the reference implementation of the official CoNLL scorer.

Entity resolution is the task of linking each mention of an entity in text to the corresponding record in a knowledge base (KB). Coherence models for entity resolution encourage all referring expressions in a document to resolve to entities that are related in the KB. We explore attention-like mechanisms for coherence, where the evidence for each candidate is based on a small set of strong relations, rather than relations to all other entities in the document. The rationale is that document-wide support may simply not exist for non-salient entities, or entities not densely connected in the KB. Our proposed sys-tem outperforms state-of-the-art systems on the CoNLL 2003, TAC KBP 2010, 2011 and 2012 tasks.

Research on generating referring expressions has so far mostly focussed on “one-shot reference”, where the aim is to generate a single, discriminating expression. In interactive settings, however, it is not uncommon for reference to be established in “installments”, where referring information is offered piecewise until success has been confirmed. We show that this strategy can also be advantageous in technical systems that only have uncertain access to object attributes and categories. We train a recently introduced model of grounded word meaning on a data set of REs for objects in images and learn to predict semantically appropriate expressions. In a human evaluation, we observe that users are sensitive to inadequate object names - which unfortunately are not unlikely to be generated from low-level visual input. We propose a solution inspired from human task-oriented interaction and implement strategies for avoiding and repairing semantically inaccurate words. We enhance a word-based REG with context-aware, referential installments and find that they substantially improve the referential success of the system.

Many language generation tasks require the production of text conditioned on both structured and unstructured inputs. We present a novel neural network architecture which generates an output sequence conditioned on an arbitrary number of input functions. Crucially, our approach allows both the choice of conditioning context and the granularity of generation, for example characters or tokens, to be marginalised, thus permitting scalable and effective training. Using this framework, we address the problem of generating programming code from a mixed natural language and structured specification. We create two new data sets for this paradigm derived from the collectible trading card games Magic the Gathering and Hearthstone. On these, and a third preexisting corpus, we demonstrate that marginalising multiple predictors allows our model to outperform strong benchmarks.

Over the past decade, large-scale supervised learning corpora have enabled machine learning researchers to make substantial advances. However, to this date, there are no large-scale question-answer corpora available. In this paper we present the 30M Factoid Question-Answer Corpus, an enormous question answer pair corpus produced by applying a novel neural network architecture on the knowledge base Freebase to transduce facts into natural language questions. The produced question answer pairs are evaluated both by human evaluators and using automatic evaluation metrics, including well-established machine translation and sentence similarity metrics. Across all evaluation criteria the question-generation model outperforms the competing template-based baseline. Furthermore, when presented to human evaluators, the generated questions appear comparable in quality to real human-generated questions.

How much is 131 million US dollars? To help readers put such numbers in context, we propose a new task of automatically generating short descriptions known as perspectives, e.g. "$131 million is about the cost to employ everyone in Texas over a lunch period". First, we collect a dataset of numeric mentions in news articles, where each mention is labeled with a set of rated perspectives. We then propose a system to generate these descriptions consisting of two steps: formula construction and description generation. In construction, we compose formulae from numeric facts in a knowledge base and rank the resulting formulas based on familiarity, numeric proximity and semantic compatibility. In generation, we convert a formula into natural language using a sequence-to-sequence recurrent neural network. Our system obtains a 15.2% F1 improvement over a non-compositional baseline at formula construction and a 12.5 BLEU point improvement over a baseline description generation.

In this study, we introduce a nondeterministic method for referring expression generation. We describe two models that account for individual variation in the choice of referential form in automatically generated text: a Naive Bayes model and a Recurrent Neural Network. Both are evaluated using the VaREG corpus. Then we select the best performing model to generate referential forms in texts from the GREC-2.0 corpus and conduct an evaluation experiment in which humans judge the coherence and comprehensibility of the generated texts, comparing them both with the original references and those produced by a random baseline model.

Modeling interactions between two sentences is crucial for a number of natu-ral language processing tasks including Answer Selection, Dialogue Act Analysis, etc. While deep learning methods like Recurrent Neural Network or Convolutional Neural Network have been proved to be powerful for sentence modeling, prior studies paid less attention on interactions between sentences. In this work, we propose a Sentence Interaction Network (SIN) for modeling the complex interactions between two sentences. By introducing “interaction states” for word and phrase pairs, SIN is powerful and flexi-ble in capturing sentence interactions for different tasks. We obtain significant improvements on Answer Selection and Dialogue Act Analysis without any feature engineering.

To establish sophisticated dialogue systems, text planning needs to cope with congruent as well as incongruent interlocutor interests as given in everyday dialogues. Little attention has been given to this topic in text planning in contrast to dialogues that are fully aligned with anticipated user interests. When considering dialogues with congruent and incongruent interlocutor interests, dialogue partners are facing the constant challenge of finding a balance between cooperation and competition. We introduce the concept of fairness that operationalize an equal and adequate, i.e. equitable satisfaction of all interlocutors’ interests. Focusing on Question-Answering (QA) settings, we describe an answer planning approach that support fair dialogues under congruent and incongruent interlocutor interests. Due to the fact that fairness is subjective per se, we present promising results from an empirical study (N=107) in which human subjects interacted with a QA system implementing the proposed approach.

The applicability of entropy rate constancy to dialogue is examined on two spoken dialogue corpora. The principle is found to hold; however, new entropy change patterns within the topic episodes of dialogue are described, which are different from written text. Speaker’s dynamic roles as topic initiators and topic responders are associated with decreasing and increasing entropy, respectively, which results in lo-cal convergence between these speakers in each topic episode. This implies that the sentence entropy in dialogue is conditioned on different contexts determined by the speaker’s roles. Explanations from the perspectives of grounding theory and interactive alignment are discussed, resulting in a novel, unified information-theoretic approach of dialogue.

In conversation, speakers tend to “accommodate” or “align” to their partners, changing the style and substance of their communications to be more similar to their partners’ utterances. We focus here on “linguistic alignment,” changes in word choice based on others’ choices. Although linguistic alignment is observed across many different contexts and its degree correlates with important social factors such as power and likability, its sources are still uncertain. We build on a recent probabilistic model of alignment, using it to separate out alignment attributable to words versus word categories. We model alignment in two contexts: telephone conversations and microblog replies. Our results show evidence of alignment, but it is primarily lexical rather than categorical. Furthermore, we find that discourse acts modulate alignment substantially. This evidence supports the view that alignment is shaped by strategic communicative processes related to the ongoing discourse.

Most current chatbot engines are designed to reply to user utterances based on existing utterance-response (or Q-R) 1 pairs. In this paper, we present DocChat, a novel information retrieval approach for chat-bot engines that can leverage unstructured documents, instead of Q-R pairs, to respond to utterances. A learning to rank model with features designed at different levels of granularity is proposed to measure the relevance between utterances and responses directly. We evaluate our proposed approach in both English and Chinese: (i) For English, we evaluate Doc-Chat on WikiQA and QASent, two answer sentence selection tasks, and compare it with state-of-the-art methods. Reasonable improvements and good adaptability are observed. (ii) For Chinese, we compare DocChat with XiaoIce 2 , a famous chitchat engine in China, and side-by-side evaluation shows that DocChat is a perfect complement for chatbot engines using Q-R pairs as main source of responses.