ACL 2015 论文列表

Topic Model such as Latent Dirichlet Allocation(LDA) makes assumption that topic assignment of different words are conditionally independent. In this paper, we propose a new model Extended Global Topic Random Field (EGTRF) to model non-linear dependencies between words. Specifically, we parse sentences into dependency trees and represent them as a graph, and assume the topic assignment of a word is influenced by its adjacent words and distance-2 words. Word similarity information learned from large corpus is incorporated to enhance word topic assignment. Parameters are estimated efficiently by variational inference and experimental results on two datasets show EGTRF achieves lower perplexity and higher log predictive probability.

This paper proposes a new unsupervised method for decomposing a multi-author document into authorial components. We assume that we do not know anything about the document and the authors, except the number of the authors of that document. The key idea is to exploit the difference in the posterior probability of the Naive-Bayesian model to increase the precision of the clustering assignment and the accuracy of the classification process of our method. Experimental results show that the proposed method outperforms two state-of-the-art methods.

In this paper, we propose the new fixedsize ordinally-forgetting encoding (FOFE) method, which can almost uniquely encode any variable-length sequence of words into a fixed-size representation. FOFE can model the word order in a sequence using a simple ordinally-forgetting mechanism according to the positions of words. In this work, we have applied FOFE to feedforward neural network language models (FNN-LMs). Experimental results have shown that without using any recurrent feedbacks, FOFE based FNNLMs can significantly outperform not only the standard fixed-input FNN-LMs but also the popular recurrent neural network (RNN) LMs.

Biterm Topic Model (BTM) is designed to model the generative process of the word co-occurrence patterns in short texts such as tweets. However, two aspects of BTM may restrict its performance: 1) user individualities are ignored to obtain the corpus level words co-occurrence patterns; and 2) the strong assumptions that two co-occurring words will be assigned the same topic label could not distinguish background words from topical words. In this paper, we propose Twitter-BTM model to address those issues by considering user level personalization in BTM. Firstly, we use user based biterms aggregation to learn user specific topic distribution. Secondly, each user’s preference between background words and topical words is estimated by incorporating a background topic. Experiments on a large-scale real-world Twitter dataset show that Twitter-BTM outperforms several stateof-the-art baselines.

Many NLP tools for English and German are based on manually annotated articles from the Wall Street Journal and Frankfurter Rundschau. The average readers of these two newspapers are middle-aged (55 and 47 years old, respectively), and the annotated articles are more than 20 years old by now. This leads us to speculate whether tools induced from these resources (such as part-of-speech taggers) put older language users at an advantage. We show that this is actually the case in both languages, and that the cause goes beyond simple vocabulary differences. In our experiments, we control for gender and region.

It has been extensively observed that languages minimise the distance between two related words. Dependency length minimisation effects are explained as a means to reduce memory load and for effective communication. In this paper, we ask whether they hold in typically short spans, such as noun phrases, which could be thought of being less subject to efficiency pressure. We demonstrate that minimisation does occur in short spans, but also that it is a complex effect: it is not only the length of the dependency that is at stake, but also the effect of the surrounding dependencies.

In this article, we first propose to exploit a new criterion for improving distributional thesauri. Following a bootstrapping perspective, we select relations between the terms of similar nominal compounds for building in an unsupervised way the training set of a classifier performing the reranking of a thesaurus. Then, we evaluate several ways to combine thesauri reranked according to different criteria and show that exploiting the complementary information brought by these criteria leads to significant improvements.

Data-driven representation learning for words is a technique of central importance in NLP. While indisputably useful as a source of features in downstream tasks, such vectors tend to consist of uninterpretable components whose relationship to the categories of traditional lexical semantic theories is tenuous at best. We present a method for constructing interpretable word vectors from hand-crafted linguistic resources like WordNet, FrameNet etc. These vectors are binary (i.e, contain only 0 and 1) and are 99.9% sparse. We analyze their performance on state-of-the-art evaluation methods for distributional models of word vectors and find they are competitive to standard distributional approaches.

We propose two improvements on lexical association used in embedding learning: factorizing individual dependency relations and using lexicographic knowledge from monolingual dictionaries. Both proposals provide low-entropy lexical cooccurrence information, and are empirically shown to improve embedding learning by performing notably better than several popular embedding models in similarity tasks. 1 Lexical Embeddings and Relatedness Lexical embeddings are essentially real-valued distributed representations of words. As a vectorspace model, an embedding model approximates semantic relatedness with the Euclidean distance between embeddings, the result of which helps better estimate the real lexical distribution in various NLP tasks. In recent years, researchers have developed efficient and effective algorithms for learning embeddings (Mikolov et al., 2013a; Pennington et al., 2014) and extended model applications from language modelling to various areas in NLP including lexical semantics (Mikolov et al., 2013b) and parsing (Bansal et al., 2014). To approximate semantic relatedness with geometric distance, objective functions are usually chosen to correlate positively with the Euclidean similarity between the embeddings of related words. Maximizing such an objective function is then equivalent to adjusting the embeddings so that those of the related words will be geometrically closer. The definition of relatedness among words can have a profound influence on the quality of the resulting embedding models. In most existing studies, relatedness is defined by co-occurrence within a window frame sliding over texts. Although supported by the distributional hypothesis (Harris, 1954), this definition suffers from two major limitations. Firstly, the window frame size is usually rather small (for efficiency and sparsity considerations), which increases the false negative rate by missing long-distance dependencies. Secondly, a window frame can (and often does) span across different constituents in a sentence, resulting in an increased false positive rate by associating unrelated words. The problem is worsened as the size of the window increases since each false-positive n-gram will appear in two subsuming false-positive (n+1)-grams. Several existing studies have addressed these limitations of window-based contexts. Nonetheless, we hypothesize that lexical embedding learning can further benefit from (1) factorizing syntactic relations into individual relations for structured syntactic information and (2) defining relatedness using lexicographic knowledge. We will show that implementation of these ideas brings notable improvement in lexical similarity tasks.

In Tibetan, as words are written consecutively without delimiters, finding unknown word boundary is difficult. This paper presents a hybrid approach for Tibetan unknown word identification for offline corpus processing. Firstly, Tibetan named entity is preprocessed based on natural annotation. Secondly, other Tibetan unknown words are extracted from word segmentation fragments using MTC, the combination of a statistical metric and a set of context sensitive rules. In addition, the preliminary experimental results on Tibetan News Corpus are reported. Lexicon-based Tibetan word segmentation system SegT with proposed unknown word extension mechanism is indeed helpful to promote the performance of Tibetan word segmentation. It increases the F-score of Tibetan word segmentation by 4.15% on random-selected test set. Our unknown word identification scheme can find new words in any length and in any field.

We introduce Discriminative BLEU (∆BLEU), a novel metric for intrinsic evaluation of generated text in tasks that admit a diverse range of possible outputs. Reference strings are scored for quality by human raters on a scale of [−1, +1] to weight multi-reference BLEU. In tasks involving generation of conversational responses, ∆BLEU correlates reasonably with human judgments and outperforms sentence-level and IBM BLEU in terms of both Spearman’s ρ and Kendall’s τ .

We describe the first version of the Media Frames Corpus: several thousand news articles on three policy issues, annotated in terms of media framing. We motivate framing as a phenomenon of study for computational linguistics and describe our annotation process.

Identifying the type of relationship between words provides a deeper insight into the history of a language and allows a better characterization of language relatedness. In this paper, we propose a computational approach for discriminating between cognates and borrowings. We show that orthographic features have discriminative power and we analyze the underlying linguistic factors that prove relevant in the classification task. To our knowledge, this is the first attempt of this kind.

We present a new release of the Paraphrase Database. PPDB 2.0 includes a discriminatively re-ranked set of paraphrases that achieve a higher correlation with human judgments than PPDB 1.0’s heuristic rankings. Each paraphrase pair in the database now also includes finegrained entailment relations, word embedding similarities, and style annotations.

Measuring word relatedness is an important ingredient of many NLP applications. Several datasets have been developed in order to evaluate such measures. The main drawback of existing datasets is the focus on single words, although natural language contains a large proportion of multiword terms. We propose the new TR9856 dataset which focuses on multi-word terms and is significantly larger than existing datasets. The new dataset includes many real world terms such as acronyms and named entities, and further handles term ambiguity by providing topical context for all term pairs. We report baseline results for common relatedness methods over the new data, and exploit its magnitude to demonstrate that a combination of these methods outperforms each individual method.

Nowadays, there are a lot of natural language processing pipelines that are based on training data created by a few experts. This paper examines how the proliferation of the internet and its collaborative application possibilities can be practically used for NLP. For that purpose, we examine how the German version of Wiktionary can be used for a lemmatization task. We introduce IWNLP, an opensource parser for Wiktionary, that reimplements several MediaWiki markup language templates for conjugated verbs and declined adjectives. The lemmatization task is evaluated on three German corpora on which we compare our results with existing software for lemmatization. With Wiktionary as a resource, we obtain a high accuracy for the lemmatization of nouns and can even improve on the results of existing software for the lemmatization of nouns.

We increase the lexical coverage of FrameNet through automatic paraphrasing. We use crowdsourcing to manually filter out bad paraphrases in order to ensure a high-precision resource. Our expanded FrameNet contains an additional 22K lexical units, a 3-fold increase over the current FrameNet, and achieves 40% better coverage when evaluated in a practical setting on New York Times data.

Labeled data is not readily available for many natural language domains, and it typically requires expensive human effort with considerable domain knowledge to produce a set of labeled data. In this paper, we propose a simple unsupervised system that helps us create a labeled resource for categorical data (e.g., a document set) using only fifteen minutes of human input. We utilize the labeled resources to discover important insights about the data. The entire process is domain independent, and demands no prior annotation samples, or rules specific to an annotation.

In this paper, we propose an `1-norm Symmetric Nonnegative Matrix TriFactorization (`1 S-NMTF) framework to cluster multi-type relational data by utilizing their interrelatedness. Due to introducing the `1-norm distances in our new objective function, the proposed approach is robust against noise and outliers, which are inherent in multi-relational data. We also derive the solution algorithm and rigorously analyze its correctness and convergence. The promising experimental results of the algorithm applied to text clustering on IMDB dataset validate the proposed approach.