Geoparsing is a fundamental technique for analyzing geo-entity information in text, which is useful for geographic applications, e.g., tourist spot recommendation. We focus on document-level geoparsing that considers geographic relatedness among geo-entity mentions and present a Japanese travelogue dataset designed for training and evaluating document-level geoparsing systems. Our dataset comprises 200 travelogue documents with rich geo-entity information: 12,171 mentions, 6,339 coreference clusters, and 2,551 geo-entities linked to geo-database entries.

Tables in scientific papers contain crucial information, such as experimental results.Entity Linking (EL) is a promising technology that analyses tables and associates them with a knowledge base.EL for table cells requires identifying the referent concept of each cell while understanding the context relevant to each cell in the paper. However, extracting the relevant context from the paper is challenging because the relevant parts are scattered in the main text and captions.This study defines a rule-based method for extracting broad context from the main text, including table captions and sentences that mention the table.Furthermore, we propose synthetic context as a more refined context generated by large language models (LLMs).In a synthetic context, contexts from the entire paper are refined by summarizing, injecting supplemental knowledge, and clarifying the referent concept.We observe this approach improves accuracy for EL by more than 10 points on the S2abEL dataset, and our qualitative analysis suggests potential future works.

We propose a global entity disambiguation (ED) model based on BERT. To capture global contextual information for ED, our model treats not only words but also entities as input tokens, and solves the task by sequentially resolving mentions to their referent entities and using resolved entities as inputs at each step. We train the model using a large entity-annotated corpus obtained from Wikipedia. We achieve new state-of-the-art results on five standard ED datasets: AIDA-CoNLL, MSNBC, AQUAINT, ACE2004, and WNED-WIKI. The source code and model checkpoint are available at https://github.com/studio-ousia/luke.

This paper presents a novel method for nested named entity recognition. As a layered method, our method extends the prior second-best path recognition method by explicitly excluding the influence of the best path. Our method maintains a set of hidden states at each time step and selectively leverages them to build a different potential function for recognition at each level. In addition, we demonstrate that recognizing innermost entities first results in better performance than the conventional outermost entities first scheme. We provide extensive experimental results on ACE2004, ACE2005, and GENIA datasets to show the effectiveness and efficiency of our proposed method.

We propose a simple method for nominal coordination boundary identification. As the main strength of our method, it can identify the coordination boundaries without training on labeled data, and can be applied even if coordination structure annotations are not available. Our system employs pre-trained word embeddings to measure the similarities of words and detects the span of coordination, assuming that conjuncts share syntactic and semantic similarities. We demonstrate that our method yields good results in identifying coordinated noun phrases in the GENIA corpus and is comparable to a recent supervised method for the case when the coordinator conjoins simple noun phrases.

Entity representations are useful in natural language tasks involving entities. In this paper, we propose new pretrained contextualized representations of words and entities based on the bidirectional transformer. The proposed model treats words and entities in a given text as independent tokens, and outputs contextualized representations of them. Our model is trained using a new pretraining task based on the masked language model of BERT. The task involves predicting randomly masked words and entities in a large entity-annotated corpus retrieved from Wikipedia. We also propose an entity-aware self-attention mechanism that is an extension of the self-attention mechanism of the transformer, and considers the types of tokens (words or entities) when computing attention scores. The proposed model achieves impressive empirical performance on a wide range of entity-related tasks. In particular, it obtains state-of-the-art results on five well-known datasets: Open Entity (entity typing), TACRED (relation classification), CoNLL-2003 (named entity recognition), ReCoRD (cloze-style question answering), and SQuAD 1.1 (extractive question answering). Our source code and pretrained representations are available at https://github.com/studio-ousia/luke.

The embeddings of entities in a large knowledge base (e.g., Wikipedia) are highly beneficial for solving various natural language tasks that involve real world knowledge. In this paper, we present Wikipedia2Vec, a Python-based open-source tool for learning the embeddings of words and entities from Wikipedia. The proposed tool enables users to learn the embeddings efficiently by issuing a single command with a Wikipedia dump file as an argument. We also introduce a web-based demonstration of our tool that allows users to visualize and explore the learned embeddings. In our experiments, our tool achieved a state-of-the-art result on the KORE entity relatedness dataset, and competitive results on various standard benchmark datasets. Furthermore, our tool has been used as a key component in various recent studies. We publicize the source code, demonstration, and the pretrained embeddings for 12 languages at https://wikipedia2vec.github.io/.

For unsegmented languages such as Japanese and Chinese, tokenization of a sentence has a significant impact on the performance of text classification. Sentences are usually segmented with words or subwords by a morphological analyzer or byte pair encoding and then encoded with word (or subword) representations for neural networks. However, segmentation is potentially ambiguous, and it is unclear whether the segmented tokens achieve the best performance for the target task. In this paper, we propose a method to simultaneously learn tokenization and text classification to address these problems. Our model incorporates a language model for unsupervised tokenization into a text classifier and then trains both models simultaneously. To make the model robust against infrequent tokens, we sampled segmentation for each sentence stochastically during training, which resulted in improved performance of text classification. We conducted experiments on sentiment analysis as a text classification task and show that our method achieves better performance than previous methods.

This study proposes a Neural Attentive Bag-of-Entities model, which is a neural network model that performs text classification using entities in a knowledge base. Entities provide unambiguous and relevant semantic signals that are beneficial for text classification. We combine simple high-recall entity detection based on a dictionary, to detect entities in a document, with a novel neural attention mechanism that enables the model to focus on a small number of unambiguous and relevant entities. We tested the effectiveness of our model using two standard text classification datasets (i.e., the 20 Newsgroups and R8 datasets) and a popular factoid question answering dataset based on a trivia quiz game. As a result, our model achieved state-of-the-art results on all datasets. The source code of the proposed model is available online at https://github.com/wikipedia2vec/wikipedia2vec.

Recently, relation classification has gained much success by exploiting deep neural networks. In this paper, we propose a new model effectively combining Segment-level Attention-based Convolutional Neural Networks (SACNNs) and Dependency-based Recurrent Neural Networks (DepRNNs). While SACNNs allow the model to selectively focus on the important information segment from the raw sequence, DepRNNs help to handle the long-distance relations from the shortest dependency path of relation entities. Experiments on the SemEval-2010 Task 8 dataset show that our model is comparable to the state-of-the-art without using any external lexical features.

We propose a simple and accurate model for coordination boundary identification. Our model decomposes the task into three sub-tasks during training; finding a coordinator, identifying inside boundaries of a pair of conjuncts, and selecting outside boundaries of it. For inference, we make use of probabilities of coordinators and conjuncts in the CKY parsing to find the optimal combination of coordinate structures. Experimental results demonstrate that our model achieves state-of-the-art results, ensuring that the global structure of coordinations is consistent.

Understanding procedural text requires tracking entities, actions and effects as the narrative unfolds. We focus on the challenging real-world problem of action-graph extraction from materials science papers, where language is highly specialized and data annotation is expensive and scarce. We propose a novel approach, Text2Quest, where procedural text is interpreted as instructions for an interactive game. A learning agent completes the game by executing the procedure correctly in a text-based simulated lab environment. The framework can complement existing approaches and enables richer forms of learning compared to static texts. We discuss potential limitations and advantages of the approach, and release a prototype proof-of-concept, hoping to encourage research in this direction.

We present a computer-assisted learning system, Jastudy, which is particularly designed for Chinese-speaking learners of Japanese as a second language (JSL) to learn Japanese functional expressions with suggestion of appropriate example sentences. The system automatically recognizes Japanese functional expressions using a free Japanese morphological analyzer MeCab, which is retrained on a new Conditional Random Fields (CRF) model. In order to select appropriate example sentences, we apply a pairwise-based machine learning tool, Support Vector Machine for Ranking (SVMrank) to estimate the complexity of the example sentences using Japanese–Chinese homographs as an important feature. In addition, we cluster the example sentences that contain Japanese functional expressions with two or more meanings and usages, based on part-of-speech, conjugation forms of verbs and semantic attributes, using the K-means clustering algorithm in Scikit-Learn. Experimental results demonstrate the effectiveness of our approach.

In this paper, we describe TextEnt, a neural network model that learns distributed representations of entities and documents directly from a knowledge base (KB). Given a document in a KB consisting of words and entity annotations, we train our model to predict the entity that the document describes and map the document and its target entity close to each other in a continuous vector space. Our model is trained using a large number of documents extracted from Wikipedia. The performance of the proposed model is evaluated using two tasks, namely fine-grained entity typing and multiclass text classification. The results demonstrate that our model achieves state-of-the-art performance on both tasks. The code and the trained representations are made available online for further academic research.

We present tools for lexicon and corpus management that offer cooperating functionality in corpus annotation. The former, named Cradle, stores a set of words and expressions where multi-word expressions are defined with their own part-of-speech information and internal syntactic structures. The latter, named ChaKi, manages text corpora with part-of-speech (POS) and syntactic dependency structure annotations. Those two tools cooperate so that the words and multi-word expressions stored in Cradle are directly referred to by ChaKi in conducting corpus annotation, and the words and expressions annotated in ChaKi can be output as a list of lexical entities that are to be stored in Cradle.

We present a simple and accurate span-based model for semantic role labeling (SRL). Our model directly takes into account all possible argument spans and scores them for each label. At decoding time, we greedily select higher scoring labeled spans. One advantage of our model is to allow us to design and use span-level features, that are difficult to use in token-based BIO tagging approaches. Experimental results demonstrate that our ensemble model achieves the state-of-the-art results, 87.4 F1 and 87.0 F1 on the CoNLL-2005 and 2012 datasets, respectively.

Part-of-speech (POS) tagging for morphologically rich languages such as Arabic is a challenging problem because of their enormous tag sets. One reason for this is that in the tagging scheme for such languages, a complete POS tag is formed by combining tags from multiple tag sets defined for each morphosyntactic category. Previous approaches in Arabic POS tagging applied one model for each morphosyntactic tagging task, without utilizing shared information between the tasks. In this paper, we propose an approach that utilizes this information by jointly modeling multiple morphosyntactic tagging tasks with a multi-task learning framework. We also propose a method of incorporating tag dictionary information into our neural models by combining word representations with representations of the sets of possible tags. Our experiments showed that the joint model with tag dictionary information results in an accuracy of 91.38% on the Penn Arabic Treebank data set, with an absolute improvement of 2.11% over the current state-of-the-art tagger.

We propose a neural network model for coordination boundary detection. Our method relies on the two common properties - similarity and replaceability in conjuncts - in order to detect both similar pairs of conjuncts and dissimilar pairs of conjuncts. The model improves identification of clause-level coordination using bidirectional RNNs incorporating two properties as features. We show that our model outperforms the existing state-of-the-art methods on the coordination annotated Penn Treebank and Genia corpus without any syntactic information from parsers.

We present Segment-level Neural CRF, which combines neural networks with a linear chain CRF for segment-level sequence modeling tasks such as named entity recognition (NER) and syntactic chunking. Our segment-level CRF can consider higher-order label dependencies compared with conventional word-level CRF. Since it is difficult to consider all possible variable length segments, our method uses segment lattice constructed from the word-level tagging model to reduce the search space. Performing experiments on NER and chunking, we demonstrate that our method outperforms conventional word-level CRF with neural networks.

We describe a neural network model that jointly learns distributed representations of texts and knowledge base (KB) entities. Given a text in the KB, we train our proposed model to predict entities that are relevant to the text. Our model is designed to be generic with the ability to address various NLP tasks with ease. We train the model using a large corpus of texts and their entity annotations extracted from Wikipedia. We evaluated the model on three important NLP tasks (i.e., sentence textual similarity, entity linking, and factoid question answering) involving both unsupervised and supervised settings. As a result, we achieved state-of-the-art results on all three of these tasks. Our code and trained models are publicly available for further academic research.

The performance of Japanese predicate argument structure (PAS) analysis has improved in recent years thanks to the joint modeling of interactions between multiple predicates. However, this approach relies heavily on syntactic information predicted by parsers, and suffers from errorpropagation. To remedy this problem, we introduce a model that uses grid-type recurrent neural networks. The proposed model automatically induces features sensitive to multi-predicate interactions from the word sequence information of a sentence. Experiments on the NAIST Text Corpus demonstrate that without syntactic information, our model outperforms previous syntax-dependent models.

Because syntactic structures and spans of multiword expressions (MWEs) are independently annotated in many English syntactic corpora, they are generally inconsistent with respect to one another, which is harmful to the implementation of an aggregate system. In this work, we construct a corpus that ensures consistency between dependency structures and MWEs, including named entities. Further, we explore models that predict both MWE-spans and an MWE-aware dependency structure. Experimental results show that our joint model using additional MWE-span features achieves an MWE recognition improvement of 1.35 points over a pipeline model.

This paper presents our ongoing work on compilation of English multi-word expression (MWE) lexicon. We are especially interested in collecting flexible MWEs, in which some other components can intervene the expression such as “a number of” vs “a large number of” where a modifier of “number” can be placed in the expression and inherit the original meaning. We fiest collect possible candidates of flexible English MWEs from the web, and annotate all of their occurrences in the Wall Street Journal portion of Ontonotes corpus. We make use of word dependency strcuture information of the sentences converted from the phrase structure annotation. This process enables semi-automatic annotation of MWEs in the corpus and simultanaously produces the internal and external dependency representation of flexible MWEs.

Text normalization is the task of transforming lexical variants to their canonical forms. We model the problem of text normalization as a character-level sequence to sequence learning problem and present a neural encoder-decoder model for solving it. To train the encoder-decoder model, many sentences pairs are generally required. However, Japanese non-standard canonical pairs are scarce in the form of parallel corpora. To address this issue, we propose a method of data augmentation to increase data size by converting existing resources into synthesized non-standard forms using handcrafted rules. We conducted an experiment to demonstrate that the synthesized corpus contributes to stably train an encoder-decoder model and improve the performance of Japanese text normalization.

The recognition of multiword expressions (MWEs) in a sentence is important for such linguistic analyses as syntactic and semantic parsing, because it is known that combining an MWE into a single token improves accuracy for various NLP tasks, such as dependency parsing and constituency parsing. However, MWEs are not annotated in Penn Treebank. Furthermore, when converting word-based dependency to MWE-aware dependency directly, one could combine nodes in an MWE into a single node. Nevertheless, this method often leads to the following problem: A node derived from an MWE could have multiple heads and the whole dependency structure including MWE might be cyclic. Therefore we converted a phrase structure to a dependency structure after establishing an MWE as a single subtree. This approach can avoid an occurrence of multiple heads and/or cycles. In this way, we constructed an English dependency corpus taking into account compound function words, which are one type of MWEs that serve as functional expressions. In addition, we report experimental results of dependency parsing using a constructed corpus.