ACL 2016 Papers

Today’s extraction of temporal information for events heavily depends on annotated temporal links. These so called TLINKs capture the relation between pairs of event mentions and time expressions. One problem is that the number of possible TLINKs grows quadratic with the number of event mentions, therefore most annotation studies concentrate on links for mentions in the same or in adjacent sentences. However, as our annotation study shows, this restriction results for 58% of the event mentions in a less precise information when the event took place. This paper proposes a new annotation scheme to anchor events in time. Not only is the annotation effort much lower as it scales linear with the number of events, it also gives a more precise anchoring when the events have happened as the complete document can be taken into account. Using this scheme, we annotated a subset of the TimeBank Corpus and compare our re-sults to other annotation schemes. Additionally, we present some baseline experiments to automatically anchor events in time. Our annotation scheme, the automated system and the annotated corpus are publicly available. 1

Word segmentation is a fundamental task for Chinese language processing. However, with the successive improvements, the standard metric is becoming hard to distinguish state-of-the-art word segmentation systems. In this paper, we propose a new psychometric-inspired evaluation metric for Chinese word segmentation, which addresses to balance the very skewed word distribution at different levels of difficulty 1 . The performance on a real evaluation shows that the proposed metric gives more reasonable and distinguishable scores and correlates well with human judgement. In addition, the proposed metric can be easily extended to evaluate other sequence labelling based NLP tasks.

Essay stance classification, the task of determining how much an essay’s author agrees with a given proposition, is an important yet under-investigated subtask in understanding an argumentative essay’s overall content. We introduce a new corpus of argumentative student essays annotated with stance information and pro-pose a computational model for automatically predicting essay stance. In an evaluation on 826 essays, our approach significantly outperforms four baselines, one of which relies on features previously developed specifically for stance classification in student essays, yielding relative error reductions of at least 11.3% and 5.3%, in micro and macro F-score, respectively.

We examine adjective-noun (AN) composition in the task of recognizing textual entailment (RTE). We analyze behavior of ANs in large corpora and show that, despite conventional wisdom, adjectives do not always restrict the denotation of the nouns they modify. We use natural logic to characterize the variety of entailment relations that can result from AN composition. Predicting these relations depends on context and on common-sense knowledge, making AN composition especially challenging for current RTE systems. We demonstrate the inability of current state-of-the-art systems to handle AN composition in a simplified RTE task which involves the insertion of only a single word.

In this paper we explore the correlation be-tween the sound of words and their meaning, by testing if the polarity (‘good guy’ or ‘bad guy’) of a character’s role in a work of fiction can be predicted by the name of the character in the absence of any other context. Our approach is based on phonological and other features proposed in prior theoretical studies of fic-tional names. These features are used to construct a predictive model over a manually annotated corpus of characters from motion pictures. By experimenting with different mixtures of features, we identify phonological features as being the most discriminative by comparison to social and other types of features, and we delve into a discussion of specific phonological and phonotactic indicators of a character’s role’s polarity.

Solving simple arithmetic word problems is one of the challenges in Natural Language Understanding. This paper presents a novel method to learn to use formulas to solve simple arithmetic word problems. Our system, analyzes each of the sentences to identify the variables and their attributes; and automatically maps this information into a higher level representation. It then uses that representation to recognize the presence of a formula along with its associated variables. An equation is then generated from the formal description of the formula. In the training phase, it learns to score the pair from the systematically generated higher level representation. It is able to solve 86.07% of the problems in a corpus of standard primary school test questions and beats the state-of-the-art by

Frames defined in FrameNet (FN) share highly similar structures with events in ACE event extraction program. An even-t in ACE is composed of an event trigger and a set of arguments. Analogously, a frame in FN is composed of a lexical unit and a set of frame elements, which play similar roles as triggers and arguments of ACE events respectively. Besides having similar structures, many frames in FN actually express certain types of events. The above observations motivate us to explore whether there exists a good mapping from frames to event-types and if it is possible to improve event detection by using FN. In this paper, we propose a global inference approach to detect events in FN. Further, based on the detected results, we analyze possible mappings from frames to event-types. Finally, we improve the performance of event detection and achieve a new state-of-the-art result by using the events automatically detected from FN.

Distant supervised relation extraction has been widely used to find novel relational facts from text. However, distant supervision inevitably accompanies with the wrong labelling problem, and these noisy data will substantially hurt the performance of relation extraction. To alleviate this issue, we propose a sentence-level attention-based model for relation extraction. In this model, we employ convolutional neural networks to embed the semantics of sentences. Afterwards, we build sentence-level attention over multiple instances, which is expected to dynamically reduce the weights of those noisy instances. Experimental results on real-world datasets show that, our model can make full use of all informative sentences and effectively reduce the influence of wrong labelled instances. Our model achieves significant and consistent improvements on relation extraction as compared with baselines. The source code of this paper can be obtained from https: //github.com/thunlp/NRE .

Conventional topic models are ineffective for topic extraction from microblog messages since the lack of structure and context among the posts renders poor message-level word co-occurrence patterns. In this work, we organize microblog posts as conversation trees based on reposting and replying relations, which enrich context information to alleviate data sparseness. Our model generates words according to topic dependencies derived from the conversation structures. In specific, we differentiate messages as leader messages, which initiate key aspects of previously focused topics or shift the focus to different topics, and follower messages that do not introduce any new information but simply echo topics from the messages that they repost or reply. Our model captures the different extents that leader and follower messages may contain the key topical words, thus further enhances the quality of the induced topics. The results of thorough experiments demonstrate the effectiveness of our proposed model.

The run time complexity of state-of-the-art inference algorithms in graph-based dependency parsing is super-linear in the number of input words (n). Recently, pruning algorithms for these models have shown to cut a large portion of the graph edges, with minimal damage to the resulting parse trees. Solving the inference problem in run time complexity determined solely by the number of edges (m) is hence of obvious importance. We propose such an inference algorithm for first-order models, which encodes the problem as a minimum spanning tree (MST) problem in an undirected graph. This allows us to utilize state-of-the-art undirected MST algorithms whose run time is O(m) at expectation and with a very high probability. A directed parse tree is then inferred from the undirected MST and is subsequently improved with respect to the directed parsing model through local greedy updates, both steps running in O(n) time. In experiments with 18 languages, a variant of the first-order MSTParser (McDonald et al., 2005b) that employs our algorithm performs very similarly to the original parser that runs an O(n^2) directed MST inference.

There are numerous studies suggesting that published news stories have an important effect on the direction of the stock market, its volatility, the volume of trades, and the value of individual stocks mentioned in the news. There is even some published research suggesting that automated sentiment analysis of news documents, quarterly reports, blogs and/or twitter data can be productively used as part of a trading strategy. This paper presents just such a family of trading strategies, and then uses this application to re-examine some of the tacit assumptions behind how sentiment analyzers are generally evaluated, in spite of the contexts of their application. This discrepancy comes at a cost.

To create accessible content for deaf users, we investigate automatically synthesizing animations of American Sign Language (ASL), including grammatically important facial expressions and head movements. Based on recordings of humans performing various types of syntactic face and head movements (which include idiosyncratic variation), we evaluate the efficacy of Continuous Profile Models (CPMs) at identifying an essential “latent trace” of the performance, for use in producing ASL animations. A metric-based evaluation and a study with deaf users indicated that this approach was more effective than a prior method for producing animations.

High quality source code is often paired with high level summaries of the computation it performs, for example in code documentation or in descriptions posted in online forums. Such summaries are extremely useful for applications such as code search but are expensive to manually author, hence only done for a small fraction of all code that is produced. In this paper, we present the first completely datadriven approach for generating high level summaries of source code. Our model, CODE-NN , uses Long Short Term Memory (LSTM) networks with attention to produce sentences that describe C# code snippets and SQL queries. CODE-NN is trained on a new corpus that is automatically collected from StackOverflow, which we release. Experiments demonstrate strong performance on two tasks: (1) code summarization, where we establish the first end-to-end learning results and outperform strong baselines, and (2) code retrieval, where our learned model improves the state of the art on a recently introduced C# benchmark by a large margin.

We propose a new unsupervised learning model, hidden softmax sequence model (HSSM), based on Boltzmann machine for dialogue structure analysis. The model employs three types of units in the hidden layer to discovery dialogue latent structures: softmax units which represent latent states of utterances; binary units which represent latent topics specified by dialogues; and a binary unit that represents the global general topic shared across the whole dialogue corpus. In addition, the model contains extra connections between adjacent hidden softmax units to formulate the dependency between latent states. Two different kinds of real world dialogue corpora, Twitter-Post and AirTicketBook-ing, are utilized for extensive comparing experiments, and the results illustrate that the proposed model outperforms sate-of-the-art popular approaches.

Distributional semantics creates vector-space representations that capture many forms of semantic similarity, but their relation to semantic entailment has been less clear. We propose a vector-space model which provides a formal foundation for a distributional semantics of entailment. Using a mean-field approximation, we develop approximate inference procedures and entailment operators over vectors of probabilities of features being known (versus unknown). We use this framework to reinterpret an existing distributional-semantic model (Word2Vec) as approximating an entailment-based model of the distributions of words in contexts, thereby predicting lexical entailment relations. In both unsupervised and semi-supervised experiments on hyponymy detection, we get substantial improvements over previous results.

Parsing for a wide variety of grammar formalisms can be performed by intersecting finite tree automata. However, naive implementations of parsing by intersection are very inefficient. We present techniques that speed up tree-automata-based parsing, to the point that it becomes practically feasible on realistic data when applied to context-free, TAG, and graph parsing. For graph parsing, we obtain the best runtimes in the literature.

Coarse-grained semantic categories such as supersenses have proven useful for a range of downstream tasks such as question answering or machine translation. To date, no effort has been put into integrating the supersenses into distributional word representations. We present a novel joint embedding model of words and supersenses, providing insights into the relationship between words and supersenses in the same vector space. Using these embeddings in a deep neural network model, we demonstrate that the supersense enrichment leads to a significant improvement in a range of downstream classification tasks.

We propose a joint formulation for learning task-specific cross-lingual word embeddings, along with classifiers for that task. Unlike prior work, which first learns the embeddings from parallel data and then plugs them in a supervised learning problem, our approach is oneshot: a single optimization problem combines a co-regularizer for the multilingual embeddings with a task-specific loss. We present theoretical results showing the limitation of Euclidean co-regularizers to increase the embedding dimension, a limitation which does not exist for other co-regularizers (such as the ‘1distance). Despite its simplicity, our method achieves state-of-the-art accuracies on the RCV1/RCV2 dataset when transferring from English to German, with training times below 1 minute. On the TED Corpus, we obtain the highest reported scores on 10 out of 11 languages.

We describe and evaluate a simple method to extract parallel sentences from comparable corpora. The approach, termed STACC, is based on expanded lexical sets and the Jaccard similarity coefficient. We evaluate our system against state-of-theart methods on a large range of datasets in different domains, for ten language pairs, showing that it either matches or outperforms current methods across the board and gives significantly better results on the noisiest datasets. STACC is a portable method, requiring no particular adaptation for new domains or language pairs, thus enabling the efficient mining of parallel sentences in comparable corpora.